From 671394ec2c49a4d317ad7a930875ec3cd71b33bf Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Fri, 23 Apr 2021 20:33:43 -0600
Subject: [PATCH 01/17] [Benchmark] Add microbenchmarks for mac with different
 sizes for DSP testing

---
 .../benchmarks/micro_benchmark/mac_2/mac_2.v  | 22 +++++++++++++++++++
 .../benchmarks/micro_benchmark/mac_4/mac_4.v  | 22 +++++++++++++++++++
 .../benchmarks/micro_benchmark/mac_6/mac_6.v  | 22 +++++++++++++++++++
 3 files changed, 66 insertions(+)
 create mode 100644 openfpga_flow/benchmarks/micro_benchmark/mac_2/mac_2.v
 create mode 100644 openfpga_flow/benchmarks/micro_benchmark/mac_4/mac_4.v
 create mode 100644 openfpga_flow/benchmarks/micro_benchmark/mac_6/mac_6.v

diff --git a/openfpga_flow/benchmarks/micro_benchmark/mac_2/mac_2.v b/openfpga_flow/benchmarks/micro_benchmark/mac_2/mac_2.v
new file mode 100644
index 000000000..457a94deb
--- /dev/null
+++ b/openfpga_flow/benchmarks/micro_benchmark/mac_2/mac_2.v
@@ -0,0 +1,22 @@
+//-------------------------------------------------------
+//  Functionality: A 2-bit multiply-acculumate circuit
+//  Author:        Xifan Tang
+//-------------------------------------------------------
+
+module mac_2(a, b, c, out);
+parameter DATA_WIDTH = 2;  /* declare a parameter. default required */
+input [DATA_WIDTH - 1 : 0] a, b, c;
+output [DATA_WIDTH - 1 : 0] out;
+
+assign out = a * b + c;
+
+endmodule
+
+
+
+
+
+
+
+
+
diff --git a/openfpga_flow/benchmarks/micro_benchmark/mac_4/mac_4.v b/openfpga_flow/benchmarks/micro_benchmark/mac_4/mac_4.v
new file mode 100644
index 000000000..c07495eb5
--- /dev/null
+++ b/openfpga_flow/benchmarks/micro_benchmark/mac_4/mac_4.v
@@ -0,0 +1,22 @@
+//-------------------------------------------------------
+//  Functionality: A 4-bit multiply-acculumate circuit
+//  Author:        Xifan Tang
+//-------------------------------------------------------
+
+module mac_4(a, b, c, out);
+parameter DATA_WIDTH = 4;  /* declare a parameter. default required */
+input [DATA_WIDTH - 1 : 0] a, b, c;
+output [DATA_WIDTH - 1 : 0] out;
+
+assign out = a * b + c;
+
+endmodule
+
+
+
+
+
+
+
+
+
diff --git a/openfpga_flow/benchmarks/micro_benchmark/mac_6/mac_6.v b/openfpga_flow/benchmarks/micro_benchmark/mac_6/mac_6.v
new file mode 100644
index 000000000..1909c2d30
--- /dev/null
+++ b/openfpga_flow/benchmarks/micro_benchmark/mac_6/mac_6.v
@@ -0,0 +1,22 @@
+//-------------------------------------------------------
+//  Functionality: A 6-bit multiply-acculumate circuit
+//  Author:        Xifan Tang
+//-------------------------------------------------------
+
+module mac_6(a, b, c, out);
+parameter DATA_WIDTH = 6;  /* declare a parameter. default required */
+input [DATA_WIDTH - 1 : 0] a, b, c;
+output [DATA_WIDTH - 1 : 0] out;
+
+assign out = a * b + c;
+
+endmodule
+
+
+
+
+
+
+
+
+

From 200b6d39a6f17fda3e95c836960a4b7ba123f401 Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Fri, 23 Apr 2021 20:36:28 -0600
Subject: [PATCH 02/17] [Benchmark] Add more micro benchmarks for mac ranging
 from 8 bit to 32 bit

---
 .../micro_benchmark/mac/mac_12/mac_12.v       | 22 +++++++++++++++++++
 .../micro_benchmark/mac/mac_16/mac_16.v       | 22 +++++++++++++++++++
 .../micro_benchmark/{ => mac}/mac_2/mac_2.v   |  0
 .../micro_benchmark/mac/mac_32/mac_32.v       | 22 +++++++++++++++++++
 .../micro_benchmark/{ => mac}/mac_4/mac_4.v   |  0
 .../micro_benchmark/{ => mac}/mac_6/mac_6.v   |  0
 .../micro_benchmark/{ => mac}/mac_8/mac_8.v   |  0
 7 files changed, 66 insertions(+)
 create mode 100644 openfpga_flow/benchmarks/micro_benchmark/mac/mac_12/mac_12.v
 create mode 100644 openfpga_flow/benchmarks/micro_benchmark/mac/mac_16/mac_16.v
 rename openfpga_flow/benchmarks/micro_benchmark/{ => mac}/mac_2/mac_2.v (100%)
 create mode 100644 openfpga_flow/benchmarks/micro_benchmark/mac/mac_32/mac_32.v
 rename openfpga_flow/benchmarks/micro_benchmark/{ => mac}/mac_4/mac_4.v (100%)
 rename openfpga_flow/benchmarks/micro_benchmark/{ => mac}/mac_6/mac_6.v (100%)
 rename openfpga_flow/benchmarks/micro_benchmark/{ => mac}/mac_8/mac_8.v (100%)

diff --git a/openfpga_flow/benchmarks/micro_benchmark/mac/mac_12/mac_12.v b/openfpga_flow/benchmarks/micro_benchmark/mac/mac_12/mac_12.v
new file mode 100644
index 000000000..ccbade6c0
--- /dev/null
+++ b/openfpga_flow/benchmarks/micro_benchmark/mac/mac_12/mac_12.v
@@ -0,0 +1,22 @@
+//-------------------------------------------------------
+//  Functionality: A 12-bit multiply-acculumate circuit
+//  Author:        Xifan Tang
+//-------------------------------------------------------
+
+module mac_12(a, b, c, out);
+parameter DATA_WIDTH = 12;  /* declare a parameter. default required */
+input [DATA_WIDTH - 1 : 0] a, b, c;
+output [DATA_WIDTH - 1 : 0] out;
+
+assign out = a * b + c;
+
+endmodule
+
+
+
+
+
+
+
+
+
diff --git a/openfpga_flow/benchmarks/micro_benchmark/mac/mac_16/mac_16.v b/openfpga_flow/benchmarks/micro_benchmark/mac/mac_16/mac_16.v
new file mode 100644
index 000000000..3fc2270e5
--- /dev/null
+++ b/openfpga_flow/benchmarks/micro_benchmark/mac/mac_16/mac_16.v
@@ -0,0 +1,22 @@
+//-------------------------------------------------------
+//  Functionality: A 16-bit multiply-acculumate circuit
+//  Author:        Xifan Tang
+//-------------------------------------------------------
+
+module mac_16(a, b, c, out);
+parameter DATA_WIDTH = 16;  /* declare a parameter. default required */
+input [DATA_WIDTH - 1 : 0] a, b, c;
+output [DATA_WIDTH - 1 : 0] out;
+
+assign out = a * b + c;
+
+endmodule
+
+
+
+
+
+
+
+
+
diff --git a/openfpga_flow/benchmarks/micro_benchmark/mac_2/mac_2.v b/openfpga_flow/benchmarks/micro_benchmark/mac/mac_2/mac_2.v
similarity index 100%
rename from openfpga_flow/benchmarks/micro_benchmark/mac_2/mac_2.v
rename to openfpga_flow/benchmarks/micro_benchmark/mac/mac_2/mac_2.v
diff --git a/openfpga_flow/benchmarks/micro_benchmark/mac/mac_32/mac_32.v b/openfpga_flow/benchmarks/micro_benchmark/mac/mac_32/mac_32.v
new file mode 100644
index 000000000..e22fd596c
--- /dev/null
+++ b/openfpga_flow/benchmarks/micro_benchmark/mac/mac_32/mac_32.v
@@ -0,0 +1,22 @@
+//-------------------------------------------------------
+//  Functionality: A 32-bit multiply-acculumate circuit
+//  Author:        Xifan Tang
+//-------------------------------------------------------
+
+module mac_32(a, b, c, out);
+parameter DATA_WIDTH = 32;  /* declare a parameter. default required */
+input [DATA_WIDTH - 1 : 0] a, b, c;
+output [DATA_WIDTH - 1 : 0] out;
+
+assign out = a * b + c;
+
+endmodule
+
+
+
+
+
+
+
+
+
diff --git a/openfpga_flow/benchmarks/micro_benchmark/mac_4/mac_4.v b/openfpga_flow/benchmarks/micro_benchmark/mac/mac_4/mac_4.v
similarity index 100%
rename from openfpga_flow/benchmarks/micro_benchmark/mac_4/mac_4.v
rename to openfpga_flow/benchmarks/micro_benchmark/mac/mac_4/mac_4.v
diff --git a/openfpga_flow/benchmarks/micro_benchmark/mac_6/mac_6.v b/openfpga_flow/benchmarks/micro_benchmark/mac/mac_6/mac_6.v
similarity index 100%
rename from openfpga_flow/benchmarks/micro_benchmark/mac_6/mac_6.v
rename to openfpga_flow/benchmarks/micro_benchmark/mac/mac_6/mac_6.v
diff --git a/openfpga_flow/benchmarks/micro_benchmark/mac_8/mac_8.v b/openfpga_flow/benchmarks/micro_benchmark/mac/mac_8/mac_8.v
similarity index 100%
rename from openfpga_flow/benchmarks/micro_benchmark/mac_8/mac_8.v
rename to openfpga_flow/benchmarks/micro_benchmark/mac/mac_8/mac_8.v

From 189c94ff196474c2aee3215bc52bb51f7da49e73 Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Fri, 23 Apr 2021 20:44:14 -0600
Subject: [PATCH 03/17] [Test] Deploy new mac benchmarks to tests

---
 .../dsp/single_mode_mult_8x8/config/task.conf          | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/openfpga_flow/tasks/fpga_verilog/dsp/single_mode_mult_8x8/config/task.conf b/openfpga_flow/tasks/fpga_verilog/dsp/single_mode_mult_8x8/config/task.conf
index 1c84a02f7..a03071351 100644
--- a/openfpga_flow/tasks/fpga_verilog/dsp/single_mode_mult_8x8/config/task.conf
+++ b/openfpga_flow/tasks/fpga_verilog/dsp/single_mode_mult_8x8/config/task.conf
@@ -30,13 +30,19 @@ openfpga_vpr_device_layout=3x2
 arch0=${PATH:OPENFPGA_PATH}/openfpga_flow/vpr_arch/k4_frac_N8_tileable_reset_softadder_register_scan_chain_dsp8_nonLR_caravel_io_skywater130nm.xml
 
 [BENCHMARKS]
-bench0=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac_8/mac_8.v
+bench0=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_2/mac_2.v
+bench1=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_4/mac_4.v
+bench2=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_6/mac_6.v
+bench3=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_8/mac_8.v
 
 [SYNTHESIS_PARAM]
 bench_yosys_common=${PATH:OPENFPGA_PATH}/openfpga_flow/misc/ys_tmpl_yosys_vpr_dsp_flow.ys
 bench_yosys_rewrite_common=${PATH:OPENFPGA_PATH}/openfpga_flow/misc/ys_tmpl_yosys_vpr_flow_with_rewrite.ys;${PATH:OPENFPGA_PATH}/openfpga_flow/misc/ys_tmpl_rewrite_flow.ys
 
-bench0_top = mac_8
+bench0_top = mac_2
+bench1_top = mac_4
+bench2_top = mac_6
+bench3_top = mac_8
 
 [SCRIPT_PARAM_MIN_ROUTE_CHAN_WIDTH]
 end_flow_with_test=

From 09cc7f0007775b996ce8b2d78de99ce745a7de15 Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Fri, 23 Apr 2021 20:44:36 -0600
Subject: [PATCH 04/17] [Script] Enable constant net routing for heterogeneous
 FPGAs

---
 .../fix_heterogeneous_device_example_script.openfpga            | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/openfpga_flow/openfpga_shell_scripts/fix_heterogeneous_device_example_script.openfpga b/openfpga_flow/openfpga_shell_scripts/fix_heterogeneous_device_example_script.openfpga
index 10ed5f17a..1d6cca82f 100644
--- a/openfpga_flow/openfpga_shell_scripts/fix_heterogeneous_device_example_script.openfpga
+++ b/openfpga_flow/openfpga_shell_scripts/fix_heterogeneous_device_example_script.openfpga
@@ -1,6 +1,6 @@
 # Run VPR for the 'and' design
 #--write_rr_graph example_rr_graph.xml
-vpr ${VPR_ARCH_FILE} ${VPR_TESTBENCH_BLIF} --clock_modeling route --device ${OPENFPGA_VPR_DEVICE_LAYOUT}
+vpr ${VPR_ARCH_FILE} ${VPR_TESTBENCH_BLIF} --device ${OPENFPGA_VPR_DEVICE_LAYOUT} --constant_net_method route
 
 # Read OpenFPGA architecture definition
 read_openfpga_arch -f ${OPENFPGA_ARCH_FILE}

From c44688739d4833337bf33e38935b7646ec593b60 Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Fri, 23 Apr 2021 22:12:26 -0600
Subject: [PATCH 05/17] [HDL] Add verilog netlist for the fracturable 16-bit
 multiplier blocks

---
 .../verilog/frac_mult_16x16.v                 | 29 +++++++++++++++++++
 1 file changed, 29 insertions(+)
 create mode 100644 openfpga_flow/openfpga_cell_library/verilog/frac_mult_16x16.v

diff --git a/openfpga_flow/openfpga_cell_library/verilog/frac_mult_16x16.v b/openfpga_flow/openfpga_cell_library/verilog/frac_mult_16x16.v
new file mode 100644
index 000000000..7fedc711c
--- /dev/null
+++ b/openfpga_flow/openfpga_cell_library/verilog/frac_mult_16x16.v
@@ -0,0 +1,29 @@
+//-----------------------------------------------------
+// Design Name : frac_mult_16x16
+// File Name   : frac_mult_16x16.v
+// Function    : A 16-bit multiplier which can operate in fracturable modes:
+//               1. two 8-bit multipliers
+//               2. one 16-bit multipliers
+// Coder       : Xifan Tang
+//-----------------------------------------------------
+
+module frac_mult_16x16 (
+	input [0:15] a,
+	input [0:15] b,
+	output [0:31] out,
+    input [0:0] mode);
+
+    reg [0:63] out_reg;
+
+    always @(mode, a, b) begin 
+       if (1'b1 == mode) begin
+         out_reg[0:15] <= a[0:7] * b[0:7];
+         out_reg[16:31] <= a[8:15] * b[8:15];
+       end else begin
+         out_reg <= a * b;
+       end
+    end
+
+    assign out = out_reg;
+
+endmodule

From 0709e5bb81eb38e3c137a113da75b64b502c4ecb Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Sat, 24 Apr 2021 13:27:44 -0600
Subject: [PATCH 06/17] [Tool] Fixed a bug in the routing trace finder for
 direct connections inside repacker

---
 openfpga/src/repack/repack.cpp | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/openfpga/src/repack/repack.cpp b/openfpga/src/repack/repack.cpp
index 5e25179a5..d6cb669bc 100644
--- a/openfpga/src/repack/repack.cpp
+++ b/openfpga/src/repack/repack.cpp
@@ -28,6 +28,8 @@ namespace openfpga {
  * - sink is an input of a primitive pb_type
  * 
  * Note: 
+ *  - This function is applicable ONLY to single-mode pb_types!!! Because their routing traces
+ *    are deterministic: there is only 1 valid path from a source pin to a sink pin!!!
  *  - If there is a fan-out of the current source pb graph pin is not a direct interconnection
  *    the direct search should stop. 
  *  - This function is designed for pb graph without local routing
@@ -58,6 +60,11 @@ bool rec_direct_search_sink_pb_graph_pins(const t_pb_graph_pin* source_pb_pin,
 
   std::vector<t_pb_graph_pin*> sink_pb_pins_to_search;
 
+  /* Only support single-mode pb_type!!! */
+  if (1 != source_pb_pin->parent_node->pb_type->num_modes) {
+    return false;
+  }
+
   for (int iedge = 0; iedge < source_pb_pin->num_output_edges; ++iedge) {
     if (DIRECT_INTERC != source_pb_pin->output_edges[iedge]->interconnect->type) {
       return false;

From 1c6b9a23d78cc1498582b588f122d262b9bae21e Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Sat, 24 Apr 2021 13:29:29 -0600
Subject: [PATCH 07/17] [Test] Add new test for multi-mode 16-bit DSP blocks

---
 .../multi_mode_mult_16x16/config/task.conf    | 47 +++++++++++++++++++
 1 file changed, 47 insertions(+)
 create mode 100644 openfpga_flow/tasks/fpga_verilog/dsp/multi_mode_mult_16x16/config/task.conf

diff --git a/openfpga_flow/tasks/fpga_verilog/dsp/multi_mode_mult_16x16/config/task.conf b/openfpga_flow/tasks/fpga_verilog/dsp/multi_mode_mult_16x16/config/task.conf
new file mode 100644
index 000000000..0db63ef8b
--- /dev/null
+++ b/openfpga_flow/tasks/fpga_verilog/dsp/multi_mode_mult_16x16/config/task.conf
@@ -0,0 +1,47 @@
+# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
+# Configuration file for running experiments
+# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
+# timeout_each_job : FPGA Task script splits fpga flow into multiple jobs
+# Each job execute fpga_flow script on combination of architecture & benchmark
+# timeout_each_job is timeout for each job
+# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
+
+[GENERAL]
+run_engine=openfpga_shell
+power_tech_file = ${PATH:OPENFPGA_PATH}/openfpga_flow/tech/PTM_45nm/45nm.xml
+power_analysis = false
+spice_output=false
+verilog_output=true
+timeout_each_job = 20*60
+fpga_flow=yosys_vpr
+
+[OpenFPGA_SHELL]
+openfpga_shell_template=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_shell_scripts/fix_heterogeneous_device_example_script.openfpga
+openfpga_arch_file=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_arch/k4_frac_N8_reset_softadder_register_scan_chain_frac_dsp16_caravel_io_skywater130nm_fdhd_cc_openfpga.xml
+openfpga_sim_setting_file=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_simulation_settings/fixed_sim_openfpga.xml
+# Yosys script parameters
+yosys_cell_sim_verilog=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_yosys_techlib/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm_cell_sim.v
+yosys_dsp_map_verilog=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_yosys_techlib/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm_dsp_map.v
+yosys_dsp_map_parameters=-D DSP_A_MAXWIDTH=8 -D DSP_B_MAXWIDTH=8 -D DSP_A_MINWIDTH=2 -D DSP_B_MINWIDTH=2 -D DSP_NAME=mult_8x8
+# VPR parameter
+openfpga_vpr_device_layout=3x2
+
+[ARCHITECTURES]
+arch0=${PATH:OPENFPGA_PATH}/openfpga_flow/vpr_arch/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm.xml
+
+[BENCHMARKS]
+bench0=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_4/mac_4.v
+bench1=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_8/mac_8.v
+bench2=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_16/mac_16.v
+
+[SYNTHESIS_PARAM]
+bench_yosys_common=${PATH:OPENFPGA_PATH}/openfpga_flow/misc/ys_tmpl_yosys_vpr_dsp_flow.ys
+bench_yosys_rewrite_common=${PATH:OPENFPGA_PATH}/openfpga_flow/misc/ys_tmpl_yosys_vpr_flow_with_rewrite.ys;${PATH:OPENFPGA_PATH}/openfpga_flow/misc/ys_tmpl_rewrite_flow.ys
+
+bench0_top = mac_4
+bench1_top = mac_8
+bench2_top = mac_16
+
+[SCRIPT_PARAM_MIN_ROUTE_CHAN_WIDTH]
+end_flow_with_test=
+vpr_fpga_verilog_formal_verification_top_netlist=

From ddcdb35b280c1bda4cd4fd9cbc4a17f9418cc7da Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Sat, 24 Apr 2021 13:30:03 -0600
Subject: [PATCH 08/17] [Arch] Bug fix in single-mode 8-bit DSP architectures

---
 ...scan_chain_dsp8_caravel_io_skywater130nm_fdhd_cc_openfpga.xml | 1 -
 1 file changed, 1 deletion(-)

diff --git a/openfpga_flow/openfpga_arch/k4_frac_N8_reset_softadder_register_scan_chain_dsp8_caravel_io_skywater130nm_fdhd_cc_openfpga.xml b/openfpga_flow/openfpga_arch/k4_frac_N8_reset_softadder_register_scan_chain_dsp8_caravel_io_skywater130nm_fdhd_cc_openfpga.xml
index ebef951da..1114d0071 100644
--- a/openfpga_flow/openfpga_arch/k4_frac_N8_reset_softadder_register_scan_chain_dsp8_caravel_io_skywater130nm_fdhd_cc_openfpga.xml
+++ b/openfpga_flow/openfpga_arch/k4_frac_N8_reset_softadder_register_scan_chain_dsp8_caravel_io_skywater130nm_fdhd_cc_openfpga.xml
@@ -262,7 +262,6 @@
     <!-- Binding operating pb_type to physical pb_type -->
     <!-- Binding operating pb_types in mode 'arithmetic' -->
     <pb_type name="clb.fle[arithmetic].soft_adder.adder_lut4" physical_pb_type_name="clb.fle[physical].fabric.frac_logic.frac_lut4" mode_bits="1"/>
-    <pb_type name="clb.fle[arithmetic].soft_adder.ff" physical_pb_type_name="clb.fle[physical].fabric.ff"/>
     <pb_type name="clb.fle[arithmetic].soft_adder.carry_follower" physical_pb_type_name="clb.fle[physical].fabric.frac_logic.carry_follower"/>
     <!-- Binding operating pb_types in mode 'n2_lut3' -->
     <pb_type name="clb.fle[n2_lut3].lut3inter.ble3.lut3" physical_pb_type_name="clb.fle[physical].fabric.frac_logic.frac_lut4" mode_bits="1" physical_pb_type_index_factor="0.5">

From 272d1fffb70fb6a02c4eb4d27f28b9b6b2b4a338 Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Sat, 24 Apr 2021 13:30:46 -0600
Subject: [PATCH 09/17] [HDL] Add tech library for architecture using
 multi-mode 16-bit DSP blocks

---
 ..._nonLR_caravel_io_skywater130nm_cell_sim.v | 25 +++++++++++
 ...6_nonLR_caravel_io_skywater130nm_dsp_map.v | 41 +++++++++++++++++++
 2 files changed, 66 insertions(+)
 create mode 100644 openfpga_flow/openfpga_yosys_techlib/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm_cell_sim.v
 create mode 100644 openfpga_flow/openfpga_yosys_techlib/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm_dsp_map.v

diff --git a/openfpga_flow/openfpga_yosys_techlib/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm_cell_sim.v b/openfpga_flow/openfpga_yosys_techlib/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm_cell_sim.v
new file mode 100644
index 000000000..d455c79d0
--- /dev/null
+++ b/openfpga_flow/openfpga_yosys_techlib/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm_cell_sim.v
@@ -0,0 +1,25 @@
+//-----------------------------
+// 8-bit multiplier
+//-----------------------------
+module mult_8(
+  input [0:7] A,
+  input [0:7] B,
+  output [0:15] Y
+);
+
+assign Y = A * B;
+
+endmodule
+
+//-----------------------------
+// 16-bit multiplier
+//-----------------------------
+module mult_16(
+  input [0:15] A,
+  input [0:15] B,
+  output [0:31] Y
+);
+
+assign Y = A * B;
+
+endmodule
diff --git a/openfpga_flow/openfpga_yosys_techlib/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm_dsp_map.v b/openfpga_flow/openfpga_yosys_techlib/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm_dsp_map.v
new file mode 100644
index 000000000..035e6f7eb
--- /dev/null
+++ b/openfpga_flow/openfpga_yosys_techlib/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm_dsp_map.v
@@ -0,0 +1,41 @@
+//-----------------------------
+// 8-bit multiplier
+//-----------------------------
+module mult_8x8 (
+  input [0:7] A,
+  input [0:7] B,
+  output [0:15] Y
+);
+  parameter A_SIGNED = 0;
+  parameter B_SIGNED = 0;
+  parameter A_WIDTH = 0;
+  parameter B_WIDTH = 0;
+  parameter Y_WIDTH = 0;
+
+  mult_8 #() _TECHMAP_REPLACE_ (
+    .A    (A),
+    .B    (B),
+    .Y    (Y) );
+
+endmodule
+
+//-----------------------------
+// 16-bit multiplier
+//-----------------------------
+module mult_16x16 (
+  input [0:15] A,
+  input [0:15] B,
+  output [0:31] Y
+);
+  parameter A_SIGNED = 0;
+  parameter B_SIGNED = 0;
+  parameter A_WIDTH = 0;
+  parameter B_WIDTH = 0;
+  parameter Y_WIDTH = 0;
+
+  mult_16 #() _TECHMAP_REPLACE_ (
+    .A    (A),
+    .B    (B),
+    .Y    (Y) );
+
+endmodule

From 4f454abfde2ad6693be9dc422dee48a40fe9ed6a Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Sat, 24 Apr 2021 14:01:42 -0600
Subject: [PATCH 10/17] [Arch] Add a new architecture using fracturable 16-bit
 DSP blocks

---
 ...avel_io_skywater130nm_fdhd_cc_openfpga.xml | 295 ++++++
 ...c_dsp16_nonLR_caravel_io_skywater130nm.xml | 962 ++++++++++++++++++
 2 files changed, 1257 insertions(+)
 create mode 100644 openfpga_flow/openfpga_arch/k4_frac_N8_reset_softadder_register_scan_chain_frac_dsp16_caravel_io_skywater130nm_fdhd_cc_openfpga.xml
 create mode 100644 openfpga_flow/vpr_arch/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm.xml

diff --git a/openfpga_flow/openfpga_arch/k4_frac_N8_reset_softadder_register_scan_chain_frac_dsp16_caravel_io_skywater130nm_fdhd_cc_openfpga.xml b/openfpga_flow/openfpga_arch/k4_frac_N8_reset_softadder_register_scan_chain_frac_dsp16_caravel_io_skywater130nm_fdhd_cc_openfpga.xml
new file mode 100644
index 000000000..7ce8c5c28
--- /dev/null
+++ b/openfpga_flow/openfpga_arch/k4_frac_N8_reset_softadder_register_scan_chain_frac_dsp16_caravel_io_skywater130nm_fdhd_cc_openfpga.xml
@@ -0,0 +1,295 @@
+<!-- Architecture annotation for OpenFPGA framework
+     This annotation supports the k4_frac_cc_sky130nm.xml
+     - General purpose logic block
+       - K = 6, N = 10, I = 40
+       - Single mode
+     - Routing architecture
+       - L = 4, fc_in = 0.15, fc_out = 0.1
+     - Skywater 130nm PDK
+       - circuit models are binded to the opensource skywater
+         foundry middle-speed (ms) standard cell library
+  -->
+<openfpga_architecture>
+  <technology_library>
+    <device_library>
+      <device_model name="logic" type="transistor">
+        <lib type="industry" corner="TOP_TT" ref="M" path="${OPENFPGA_PATH}/openfpga_flow/tech/PTM_45nm/45nm.pm"/>
+        <design vdd="0.9" pn_ratio="2"/>
+        <pmos name="pch" chan_length="40e-9" min_width="140e-9" variation="logic_transistor_var"/>
+        <nmos name="nch" chan_length="40e-9" min_width="140e-9" variation="logic_transistor_var"/>
+      </device_model>
+      <device_model name="io" type="transistor">
+        <lib type="academia" ref="M" path="${OPENFPGA_PATH}/openfpga_flow/tech/PTM_45nm/45nm.pm"/>
+        <design vdd="2.5" pn_ratio="3"/>
+        <pmos name="pch_25" chan_length="270e-9" min_width="320e-9" variation="io_transistor_var"/>
+        <nmos name="nch_25" chan_length="270e-9" min_width="320e-9" variation="io_transistor_var"/>
+      </device_model>
+    </device_library>
+    <variation_library>
+      <variation name="logic_transistor_var" abs_deviation="0.1" num_sigma="3"/>
+      <variation name="io_transistor_var" abs_deviation="0.1" num_sigma="3"/>
+    </variation_library>
+  </technology_library>
+  <circuit_library>
+    <circuit_model type="inv_buf" name="sky130_fd_sc_hd__inv_1" prefix="sky130_fd_sc_hd__inv_1" is_default="true">
+      <design_technology type="cmos" topology="inverter" size="1"/>
+      <device_technology device_model_name="logic"/>
+      <port type="input" prefix="in" lib_name="A" size="1"/>
+      <port type="output" prefix="out" lib_name="Y" size="1"/>
+      <delay_matrix type="rise" in_port="in" out_port="out">
+        10e-12
+      </delay_matrix>
+      <delay_matrix type="fall" in_port="in" out_port="out">
+        10e-12
+      </delay_matrix>
+    </circuit_model>
+    <circuit_model type="inv_buf" name="sky130_fd_sc_hd__buf_2" prefix="sky130_fd_sc_hd__buf_2" is_default="false">
+      <design_technology type="cmos" topology="buffer" size="1" num_level="2" f_per_stage="2"/>
+      <device_technology device_model_name="logic"/>
+      <port type="input" prefix="in" lib_name="A" size="1"/>
+      <port type="output" prefix="out" lib_name="X" size="1"/>
+      <delay_matrix type="rise" in_port="in" out_port="out">
+        10e-12
+      </delay_matrix>
+      <delay_matrix type="fall" in_port="in" out_port="out">
+        10e-12
+      </delay_matrix>
+    </circuit_model>
+    <circuit_model type="inv_buf" name="sky130_fd_sc_hd__buf_4" prefix="sky130_fd_sc_hd__buf_4" is_default="false">
+      <design_technology type="cmos" topology="buffer" size="1" num_level="2" f_per_stage="4"/>
+      <device_technology device_model_name="logic"/>
+      <port type="input" prefix="in" lib_name="A" size="1"/>
+      <port type="output" prefix="out" lib_name="X" size="1"/>
+      <delay_matrix type="rise" in_port="in" out_port="out">
+        10e-12
+      </delay_matrix>
+      <delay_matrix type="fall" in_port="in" out_port="out">
+        10e-12
+      </delay_matrix>
+    </circuit_model>
+    <circuit_model type="inv_buf" name="sky130_fd_sc_hd__inv_2" prefix="sky130_fd_sc_hd__inv_2" is_default="false">
+      <design_technology type="cmos" topology="buffer" size="1"/>
+      <device_technology device_model_name="logic"/>
+      <port type="input" prefix="in" lib_name="A" size="1"/>
+      <port type="output" prefix="out" lib_name="Y" size="1"/>
+      <delay_matrix type="rise" in_port="in" out_port="out">
+        10e-12
+      </delay_matrix>
+      <delay_matrix type="fall" in_port="in" out_port="out">
+        10e-12
+      </delay_matrix>
+    </circuit_model>
+    <circuit_model type="gate" name="sky130_fd_sc_hd__or2_1" prefix="sky130_fd_sc_hd__or2_1" is_default="true">
+      <design_technology type="cmos" topology="OR"/>
+      <device_technology device_model_name="logic"/>
+      <input_buffer exist="false"/>
+      <output_buffer exist="false"/>
+      <port type="input" prefix="a" lib_name="A" size="1"/>
+      <port type="input" prefix="b" lib_name="B" size="1"/>
+      <port type="output" prefix="out" lib_name="X" size="1"/>
+      <delay_matrix type="rise" in_port="a b" out_port="out">
+        10e-12 5e-12
+      </delay_matrix>
+      <delay_matrix type="fall" in_port="a b" out_port="out">
+        10e-12 5e-12
+      </delay_matrix>
+    </circuit_model>
+    <!-- Define a circuit model for the standard cell MUX2
+         OpenFPGA requires the following truth table for the MUX2
+         When the select signal sel is enabled, the first input, i.e., in0
+         will be propagated to the output, i.e., out
+         If your standard cell provider does not offer the exact truth table,
+         you can simply swap the inputs as shown in the example below
+      -->
+    <circuit_model type="gate" name="sky130_fd_sc_hd__mux2_1" prefix="sky130_fd_sc_hd__mux2_1">
+      <design_technology type="cmos" topology="MUX2"/>
+      <device_technology device_model_name="logic"/>
+      <input_buffer exist="false"/>
+      <output_buffer exist="false"/>
+      <port type="input" prefix="in0" lib_name="A1" size="1"/>
+      <port type="input" prefix="in1" lib_name="A0" size="1"/>
+      <port type="input" prefix="sel" lib_name="S" size="1"/>
+      <port type="output" prefix="out" lib_name="X" size="1"/>
+    </circuit_model>
+    <circuit_model type="chan_wire" name="chan_segment" prefix="track_seg" is_default="true">
+      <design_technology type="cmos"/>
+      <input_buffer exist="false"/>
+      <output_buffer exist="false"/>
+      <port type="input" prefix="in" size="1"/>
+      <port type="output" prefix="out" size="1"/>
+      <wire_param model_type="pi" R="101" C="22.5e-15" num_level="1"/>
+      <!-- model_type could be T, res_val and cap_val DON'T CARE -->
+    </circuit_model>
+    <circuit_model type="wire" name="direct_interc" prefix="direct_interc" is_default="true">
+      <design_technology type="cmos"/>
+      <input_buffer exist="false"/>
+      <output_buffer exist="false"/>
+      <port type="input" prefix="in" size="1"/>
+      <port type="output" prefix="out" size="1"/>
+      <wire_param model_type="pi" R="0" C="0" num_level="1"/>
+      <!-- model_type could be T, res_val cap_val should be defined -->
+    </circuit_model>
+    <circuit_model type="mux" name="mux_tree" prefix="mux_tree" is_default="true" dump_structural_verilog="true">
+      <design_technology type="cmos" structure="tree" add_const_input="true" const_input_val="1"/>
+      <input_buffer exist="false"/>
+      <output_buffer exist="false"/>
+      <pass_gate_logic circuit_model_name="sky130_fd_sc_hd__mux2_1"/>
+      <port type="input" prefix="in" size="1"/>
+      <port type="output" prefix="out" size="1"/>
+      <port type="sram" prefix="sram" size="1"/>
+    </circuit_model>
+    <circuit_model type="mux" name="mux_tree_tapbuf" prefix="mux_tree_tapbuf" dump_structural_verilog="true">
+      <design_technology type="cmos" structure="tree" add_const_input="true" const_input_val="1"/>
+      <input_buffer exist="false"/>
+      <output_buffer exist="true" circuit_model_name="sky130_fd_sc_hd__buf_4"/>
+      <pass_gate_logic circuit_model_name="sky130_fd_sc_hd__mux2_1"/>
+      <port type="input" prefix="in" size="1"/>
+      <port type="output" prefix="out" size="1"/>
+      <port type="sram" prefix="sram" size="1"/>
+    </circuit_model>
+    <!--DFF subckt ports should be defined as <D> <Q> <CLK> <RESET> <SET>  -->
+    <circuit_model type="ff" name="SDFFRQ" prefix="SDFFRQ" verilog_netlist="${OPENFPGA_PATH}/openfpga_flow/openfpga_cell_library/verilog/dff.v">
+      <design_technology type="cmos"/>
+      <input_buffer exist="true" circuit_model_name="sky130_fd_sc_hd__inv_1"/>
+      <output_buffer exist="true" circuit_model_name="sky130_fd_sc_hd__inv_1"/>
+      <port type="input" prefix="D" size="1"/>
+      <port type="input" prefix="DI" lib_name="SI" size="1"/>
+      <port type="input" prefix="Test_en" lib_name="SE" size="1" is_global="true" default_val="0"/>
+      <port type="input" prefix="reset" lib_name="RST" size="1" default_val="0"/> 
+      <port type="output" prefix="Q" size="1"/>
+      <port type="clock" prefix="clk" lib_name="CK" size="1" default_val="0" />
+    </circuit_model>
+    <circuit_model type="lut" name="frac_lut4" prefix="frac_lut4" dump_structural_verilog="true">
+      <design_technology type="cmos" fracturable_lut="true"/>
+      <input_buffer exist="false"/>
+      <output_buffer exist="true" circuit_model_name="sky130_fd_sc_hd__buf_2"/>
+      <lut_input_inverter exist="true" circuit_model_name="sky130_fd_sc_hd__inv_1"/>
+      <lut_input_buffer exist="true" circuit_model_name="sky130_fd_sc_hd__buf_2"/>
+      <lut_intermediate_buffer exist="true" circuit_model_name="sky130_fd_sc_hd__buf_2" location_map="-1-"/>
+      <pass_gate_logic circuit_model_name="sky130_fd_sc_hd__mux2_1"/>
+      <port type="input" prefix="in" size="4" tri_state_map="---1" circuit_model_name="sky130_fd_sc_hd__or2_1"/>
+      <port type="output" prefix="lut2_out" size="2" lut_frac_level="2" lut_output_mask="2,3"/>
+      <port type="output" prefix="lut3_out" size="2" lut_frac_level="3" lut_output_mask="0,1"/>
+      <port type="output" prefix="lut4_out" size="1" lut_output_mask="0"/>
+      <port type="sram" prefix="sram" size="16"/>
+      <port type="sram" prefix="mode" size="1" mode_select="true" circuit_model_name="DFFRQ" default_val="1"/>
+    </circuit_model>
+    <!--Scan-chain DFF subckt ports should be defined as <D> <Q> <Qb> <CLK> <RESET> <SET>  -->
+    <circuit_model type="ccff" name="DFFRQ" prefix="DFFRQ" verilog_netlist="${OPENFPGA_PATH}/openfpga_flow/openfpga_cell_library/verilog/dff.v">
+      <design_technology type="cmos"/>
+      <input_buffer exist="true" circuit_model_name="sky130_fd_sc_hd__inv_1"/>
+      <output_buffer exist="true" circuit_model_name="sky130_fd_sc_hd__inv_1"/>
+      <port type="input" prefix="D" size="1"/>
+      <port type="output" prefix="Q" size="1"/>
+      <port type="clock" prefix="prog_clk" lib_name="CK" size="1" is_global="true" default_val="0" is_prog="true"/>
+      <port type="input" prefix="pReset" lib_name="RST" size="1" is_global="true" default_val="0" is_prog="true" is_reset="true"/> 
+    </circuit_model>
+    <circuit_model type="iopad" name="EMBEDDED_IO_ISOLN" prefix="EMBEDDED_IO_ISOLN" is_default="true" verilog_netlist="${OPENFPGA_PATH}/openfpga_flow/openfpga_cell_library/verilog/gpio.v">
+      <design_technology type="cmos"/>
+      <input_buffer exist="true" circuit_model_name="sky130_fd_sc_hd__inv_1"/>
+      <output_buffer exist="true" circuit_model_name="sky130_fd_sc_hd__inv_1"/>
+      <port type="input" prefix="SOC_IN" lib_name="SOC_IN" size="1" is_global="true" is_io="true" is_data_io="true"/>
+      <port type="output" prefix="SOC_OUT" lib_name="SOC_OUT" size="1" is_global="true" is_io="true" is_data_io="true"/>
+      <port type="output" prefix="SOC_DIR" lib_name="SOC_DIR" size="1" is_global="true" is_io="true"/>
+      <port type="input" prefix="IO_ISOL_N" lib_name="IO_ISOL_N" size="1" is_global="true" default_val="1"/>
+      <port type="output" prefix="inpad" lib_name="FPGA_IN" size="1"/>
+      <port type="input" prefix="outpad" lib_name="FPGA_OUT" size="1"/>
+      <port type="sram" prefix="en" lib_name="FPGA_DIR" size="1" mode_select="true" circuit_model_name="DFFRQ" default_val="1"/>
+    </circuit_model>
+    <circuit_model type="hard_logic" name="CARRY_MUX2" prefix="CARRY_MUX2" verilog_netlist="${OPENFPGA_PATH}/openfpga_flow/openfpga_cell_library/verilog/mux2.v">
+      <design_technology type="cmos"/>
+      <device_technology device_model_name="logic"/>
+      <input_buffer exist="false"/>
+      <output_buffer exist="false"/>
+      <port type="input" prefix="a" lib_name="A0" size="1"/>
+      <port type="input" prefix="b" lib_name="A1" size="1"/>
+      <port type="input" prefix="cin" lib_name="S" size="1"/>
+      <port type="output" prefix="cout" lib_name="Y" size="1"/>
+    </circuit_model>
+    <circuit_model type="hard_logic" name="frac_mult_16x16" prefix="frac_mult_16x16" is_default="true" spice_netlist="${OPENFPGA_PATH}/openfpga_flow/openfpga_cell_library/spice/frac_mult_16x16.sp" verilog_netlist="${OPENFPGA_PATH}/openfpga_flow/openfpga_cell_library/verilog/frac_mult_16x16.v">
+      <design_technology type="cmos"/>
+      <input_buffer exist="true" circuit_model_name="INVTX1"/>
+      <output_buffer exist="true" circuit_model_name="INVTX1"/>
+      <port type="input" prefix="A" lib_name="A" size="16"/>
+      <port type="input" prefix="B" lib_name="B" size="16"/>
+      <port type="output" prefix="Y" lib_name="Y" size="32"/>
+      <port type="sram" prefix="mode" lib_name="mode" size="1" mode_select="true" circuit_model_name="DFFRQ" default_val="0"/>
+    </circuit_model>
+  </circuit_library>
+  <configuration_protocol>
+    <organization type="scan_chain" circuit_model_name="DFFRQ" num_regions="1"/>
+  </configuration_protocol>
+  <connection_block>
+    <switch name="ipin_cblock" circuit_model_name="mux_tree_tapbuf"/>
+  </connection_block>
+  <switch_block>
+    <switch name="L1_mux" circuit_model_name="mux_tree_tapbuf"/>
+    <switch name="L2_mux" circuit_model_name="mux_tree_tapbuf"/>
+    <switch name="L4_mux" circuit_model_name="mux_tree_tapbuf"/>
+  </switch_block>
+  <routing_segment>
+    <segment name="L1" circuit_model_name="chan_segment"/>
+    <segment name="L2" circuit_model_name="chan_segment"/>
+    <segment name="L4" circuit_model_name="chan_segment"/>
+  </routing_segment>
+  <direct_connection>
+    <direct name="carry_chain" circuit_model_name="direct_interc"/>
+    <direct name="shift_register" circuit_model_name="direct_interc"/>
+    <direct name="scan_chain" circuit_model_name="direct_interc" type="column" x_dir="positive" y_dir="positive"/>
+  </direct_connection>
+  <tile_annotations>
+      <global_port name="clk" is_clock="true" default_val="0">
+          <tile name="clb" port="clk" x="-1" y="-1"/>
+      </global_port>
+      <global_port name="Reset" is_reset="true" default_val="1">
+          <tile name="clb" port="reset" x="-1" y="-1"/>
+      </global_port>
+  </tile_annotations>
+  <pb_type_annotations>
+    <!-- physical pb_type binding in complex block IO -->
+    <pb_type name="io" physical_mode_name="physical" idle_mode_name="inpad"/>
+    <!-- IMPORTANT: must set unused I/Os to operating in INPUT mode !!! -->
+    <pb_type name="io[physical].iopad" circuit_model_name="EMBEDDED_IO_ISOLN" mode_bits="1"/> 
+    <pb_type name="io[inpad].inpad" physical_pb_type_name="io[physical].iopad" mode_bits="1"/> 
+    <pb_type name="io[outpad].outpad" physical_pb_type_name="io[physical].iopad" mode_bits="0"/> 
+    <!-- End physical pb_type binding in complex block IO -->
+
+    <!-- physical pb_type binding in complex block CLB -->
+    <!-- physical mode will be the default mode if not specified -->
+    <pb_type name="clb.fle" physical_mode_name="physical"/>
+    <pb_type name="clb.fle[physical].fabric.frac_logic.frac_lut4" circuit_model_name="frac_lut4" mode_bits="0"/>
+    <pb_type name="clb.fle[physical].fabric.frac_logic.carry_follower" circuit_model_name="CARRY_MUX2"/>
+    <pb_type name="clb.fle[physical].fabric.ff" circuit_model_name="SDFFRQ"/>
+    <!-- Binding operating pb_type to physical pb_type -->
+    <!-- Binding operating pb_types in mode 'arithmetic' -->
+    <pb_type name="clb.fle[arithmetic].soft_adder.adder_lut4" physical_pb_type_name="clb.fle[physical].fabric.frac_logic.frac_lut4" mode_bits="1"/>
+    <pb_type name="clb.fle[arithmetic].soft_adder.carry_follower" physical_pb_type_name="clb.fle[physical].fabric.frac_logic.carry_follower"/>
+    <!-- Binding operating pb_types in mode 'n2_lut3' -->
+    <pb_type name="clb.fle[n2_lut3].lut3inter.ble3.lut3" physical_pb_type_name="clb.fle[physical].fabric.frac_logic.frac_lut4" mode_bits="1" physical_pb_type_index_factor="0.5">
+      <!-- Binding the lut3 to the first 3 inputs of fracturable lut4 -->
+      <port name="in" physical_mode_port="in[0:2]"/>
+      <port name="out" physical_mode_port="lut3_out[0:0]" physical_mode_pin_rotate_offset="1"/>
+    </pb_type>
+    <pb_type name="clb.fle[n2_lut3].lut3inter.ble3.ff" physical_pb_type_name="clb.fle[physical].fabric.ff"/>
+    <!-- Binding operating pb_types in mode 'ble4' -->
+    <pb_type name="clb.fle[n1_lut4].ble4.lut4" physical_pb_type_name="clb.fle[physical].fabric.frac_logic.frac_lut4" mode_bits="0">
+      <!-- Binding the lut4 to the first 4 inputs of fracturable lut4 -->
+      <port name="in" physical_mode_port="in[0:3]"/>
+      <port name="out" physical_mode_port="lut4_out"/>
+    </pb_type>
+    <pb_type name="clb.fle[n1_lut4].ble4.ff" physical_pb_type_name="clb.fle[physical].fabric.ff" physical_pb_type_index_factor="2" physical_pb_type_index_offset="0"/>
+    <!-- Binding operating pb_types in mode 'shift_register' -->
+    <pb_type name="clb.fle[shift_register].shift_reg.ff" physical_pb_type_name="clb.fle[physical].fabric.ff"/>
+    <!-- End physical pb_type binding in complex block IO -->
+
+    <!-- physical pb_type binding in complex block dsp -->
+    <pb_type name="mult_16" physical_mode_name="mult_16x16"/>
+    <!-- Bind the primitive pb_type in the physical mode to a circuit model --> 
+    <pb_type name="mult_16[mult_16x16].mult_16x16_slice.mult_16x16" circuit_model_name="frac_mult_16x16" mode_bits="0"/>
+    <pb_type name="mult_16[mult_8x8].mult_8x8_slice.mult_8x8" physical_pb_type_name="mult_16[mult_16x16].mult_16x16_slice.mult_16x16" mode_bits="1" physical_pb_type_index_factor="0">
+      <port name="A" physical_mode_port="A[0:7]" physical_mode_port_rotate_offset="8"/>
+      <port name="B" physical_mode_port="B[0:7]" physical_mode_port_rotate_offset="8"/>
+      <port name="Y" physical_mode_port="Y[0:15]" physical_mode_port_rotate_offset="16"/>
+    </pb_type>
+  </pb_type_annotations>
+</openfpga_architecture>
diff --git a/openfpga_flow/vpr_arch/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm.xml b/openfpga_flow/vpr_arch/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm.xml
new file mode 100644
index 000000000..98bae4fa2
--- /dev/null
+++ b/openfpga_flow/vpr_arch/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm.xml
@@ -0,0 +1,962 @@
+<!-- 
+  Low-cost homogeneous FPGA Architecture.
+
+  - Skywater 130 nm technology
+  - General purpose logic block: 
+    K = 4, N = 8, fracturable 4 LUTs (can operate as one 4-LUT or two 3-LUTs with all 3 inputs shared)
+    with optionally registered outputs
+  - Heterogeneous block
+    8-bit multiplier
+  - Routing architecture:
+      - 10% L = 1, fc_in = 0.15, Fc_out = 0.10
+      - 10% L = 2, fc_in = 0.15, Fc_out = 0.10
+      - 80% L = 4, fc_in = 0.15, Fc_out = 0.10
+      - 100 routing tracks per channel
+
+  Authors: Xifan Tang
+-->
+<architecture>
+  <!-- 
+       ODIN II specific config begins 
+       Describes the types of user-specified netlist blocks (in blif, this corresponds to 
+       ".model [type_of_block]") that this architecture supports.
+
+       Note: Basic LUTs, I/Os, and flip-flops are not included here as there are 
+       already special structures in blif (.names, .input, .output, and .latch) 
+       that describe them.
+  -->
+  <models>
+    <model name="mult_8">
+      <input_ports>
+        <port name="A" combinational_sink_ports="Y"/>
+        <port name="B" combinational_sink_ports="Y"/>
+      </input_ports>
+      <output_ports>
+        <port name="Y"/>
+      </output_ports>
+    </model>
+    <model name="mult_16">
+      <input_ports>
+        <port name="A" combinational_sink_ports="Y"/>
+        <port name="B" combinational_sink_ports="Y"/>
+      </input_ports>
+      <output_ports>
+        <port name="Y"/>
+      </output_ports>
+    </model>
+    <!-- A virtual model for I/O to be used in the physical mode of io block -->
+    <model name="io">
+      <input_ports>
+        <port name="outpad"/>
+      </input_ports>
+      <output_ports>
+        <port name="inpad"/>
+      </output_ports>
+    </model>
+    <model name="adder_lut4">
+      <input_ports>
+        <port name="in" combinational_sink_ports="lut2_out lut4_out"/>
+      </input_ports>
+      <output_ports>
+        <port name="lut2_out"/>
+        <port name="lut4_out"/>
+      </output_ports>
+    </model>
+    <model name="carry_follower">
+      <input_ports>
+        <port name="a" combinational_sink_ports="cout"/>
+        <port name="b" combinational_sink_ports="cout"/>
+        <port name="cin" combinational_sink_ports="cout"/>
+      </input_ports>
+      <output_ports>
+        <port name="cout"/>
+      </output_ports>
+    </model>
+    <model name="frac_lut4">
+      <input_ports>
+        <port name="in"/>
+      </input_ports>
+      <output_ports>
+        <port name="lut2_out"/>
+        <port name="lut3_out"/>
+        <port name="lut4_out"/>
+      </output_ports>
+    </model>
+    <model name="carry_follower_physical">
+      <input_ports>
+        <port name="a" combinational_sink_ports="cout"/>
+        <port name="b" combinational_sink_ports="cout"/>
+        <port name="cin" combinational_sink_ports="cout"/>
+      </input_ports>
+      <output_ports>
+        <port name="cout"/>
+      </output_ports>
+    </model>
+    <!-- A virtual model for scan-chain flip-flop to be used in the physical mode of FF -->
+    <model name="scff">
+      <input_ports>
+        <port name="D" clock="clk"/>
+        <port name="DI" clock="clk"/>
+        <port name="reset" clock="clk"/>
+        <port name="clk" is_clock="1"/>
+      </input_ports>
+      <output_ports>
+        <port name="Q" clock="clk"/>
+      </output_ports>
+    </model>
+  </models>
+  <tiles>
+    <!-- Do NOT add clock pins to I/O here!!! VPR does not build clock network in the way that OpenFPGA can support
+         If you need to register the I/O, define clocks in the circuit models
+         These clocks can be handled in back-end
+     -->
+    <!-- Top-side has 1 I/O per tile -->
+    <tile name="io_top" capacity="1" area="0">
+      <equivalent_sites>
+        <site pb_type="io"/>
+      </equivalent_sites>
+      <input name="outpad" num_pins="1"/>
+      <output name="inpad" num_pins="1"/>
+      <fc in_type="frac" in_val="0.15" out_type="frac" out_val="0.10"/>
+      <pinlocations pattern="custom">
+        <loc side="bottom">io_top.outpad io_top.inpad</loc>
+      </pinlocations>
+    </tile>
+    <!-- Right-side has 1 I/O per tile -->
+    <tile name="io_right" capacity="1" area="0">
+      <equivalent_sites>
+        <site pb_type="io"/>
+      </equivalent_sites>
+      <input name="outpad" num_pins="1"/>
+      <output name="inpad" num_pins="1"/>
+      <fc in_type="frac" in_val="0.15" out_type="frac" out_val="0.10"/>
+      <pinlocations pattern="custom">
+        <loc side="left">io_right.outpad io_right.inpad</loc>
+      </pinlocations>
+    </tile>
+    <!-- Bottom-side has 9 I/O per tile -->
+    <tile name="io_bottom" capacity="9" area="0">
+      <equivalent_sites>
+        <site pb_type="io"/>
+      </equivalent_sites>
+      <input name="outpad" num_pins="1"/>
+      <output name="inpad" num_pins="1"/>
+      <fc in_type="frac" in_val="0.15" out_type="frac" out_val="0.10"/>
+      <pinlocations pattern="custom">
+        <loc side="top">io_bottom.outpad io_bottom.inpad</loc>
+      </pinlocations>
+    </tile>
+    <!-- Left-side has 1 I/O per tile -->
+    <tile name="io_left" capacity="1" area="0">
+      <equivalent_sites>
+        <site pb_type="io"/>
+      </equivalent_sites>
+      <input name="outpad" num_pins="1"/>
+      <output name="inpad" num_pins="1"/>
+      <fc in_type="frac" in_val="0.15" out_type="frac" out_val="0.10"/>
+      <pinlocations pattern="custom">
+        <loc side="right">io_left.outpad io_left.inpad</loc>
+      </pinlocations>
+    </tile>
+    <!-- CLB has most pins on the top and right sides -->
+    <tile name="clb" area="53894">
+      <equivalent_sites>
+        <site pb_type="clb"/>
+      </equivalent_sites>
+      <input name="I0" num_pins="2" equivalent="full"/>
+      <input name="I0i" num_pins="2" equivalent="none"/>
+      <input name="I1" num_pins="2" equivalent="full"/>
+      <input name="I1i" num_pins="2" equivalent="none"/>
+      <input name="I2" num_pins="2" equivalent="full"/>
+      <input name="I2i" num_pins="2" equivalent="none"/>
+      <input name="I3" num_pins="2" equivalent="full"/>
+      <input name="I3i" num_pins="2" equivalent="none"/>
+      <input name="I4" num_pins="2" equivalent="full"/>
+      <input name="I4i" num_pins="2" equivalent="none"/>
+      <input name="I5" num_pins="2" equivalent="full"/>
+      <input name="I5i" num_pins="2" equivalent="none"/>
+      <input name="I6" num_pins="2" equivalent="full"/>
+      <input name="I6i" num_pins="2" equivalent="none"/>
+      <input name="I7" num_pins="2" equivalent="full"/>
+      <input name="I7i" num_pins="2" equivalent="none"/>
+      <input name="reg_in" num_pins="1"/>
+      <input name="sc_in" num_pins="1"/>
+      <input name="cin" num_pins="1"/>
+      <input name="reset" num_pins="1" is_non_clock_global="true"/>
+      <output name="O" num_pins="16" equivalent="none"/>
+      <output name="reg_out" num_pins="1"/>
+      <output name="sc_out" num_pins="1"/>
+      <output name="cout" num_pins="1"/>
+      <clock name="clk" num_pins="1"/>
+      <fc in_type="frac" in_val="0.15" out_type="frac" out_val="0.10">
+        <fc_override port_name="reg_in" fc_type="frac" fc_val="0"/>
+        <fc_override port_name="reg_out" fc_type="frac" fc_val="0"/>
+        <fc_override port_name="sc_in" fc_type="frac" fc_val="0"/>
+        <fc_override port_name="sc_out" fc_type="frac" fc_val="0"/>
+        <fc_override port_name="cin" fc_type="frac" fc_val="0"/>
+        <fc_override port_name="cout" fc_type="frac" fc_val="0"/>
+        <fc_override port_name="clk" fc_type="frac" fc_val="0"/>
+        <fc_override port_name="reset" fc_type="frac" fc_val="0"/>
+      </fc>
+      <!--pinlocations pattern="spread"/-->
+      <pinlocations pattern="custom">
+        <loc side="left">clb.clk clb.reset</loc>
+        <loc side="top">clb.reg_in clb.sc_in clb.cin clb.O[7:0] clb.I0 clb.I0i clb.I1 clb.I1i clb.I2 clb.I2i clb.I3 clb.I3i</loc>
+        <loc side="right">clb.O[15:8] clb.I4 clb.I4i clb.I5 clb.I5i clb.I6 clb.I6i clb.I7 clb.I7i</loc>
+        <loc side="bottom">clb.reg_out clb.sc_out clb.cout</loc>
+      </pinlocations>
+    </tile>
+    <tile name="mult_16" height="2" area="396000">
+      <equivalent_sites>
+        <site pb_type="mult_16" pin_mapping="direct"/>
+      </equivalent_sites>
+      <input name="a" num_pins="16"/>
+      <input name="b" num_pins="16"/>
+      <output name="out" num_pins="32"/>
+      <fc in_type="frac" in_val="0.15" out_type="frac" out_val="0.10"/>
+      <!--  Highly recommand to customize pin location when direct connection is used!!! -->
+      <!--pinlocations pattern="spread"/-->
+      <pinlocations pattern="custom">
+        <loc side="left">mult_16.a[0:5] mult_16.b[0:5] mult_16.out[0:10]</loc>
+        <loc side="top">mult_16.a[6:7] mult_16.b[6:7] mult_16.out[11:15]</loc>
+        <loc side="right">mult_16.a[8:13] mult_16.b[8:13] mult_16.out[16:26]</loc>
+        <loc side="bottom">mult_16.a[14:15] mult_16.b[14:15] mult_16.out[27:31]</loc>
+      </pinlocations>
+    </tile>
+  </tiles>
+  <!-- ODIN II specific config ends -->
+  <!-- Physical descriptions begin -->
+  <layout tileable="true">
+    <auto_layout aspect_ratio="1.0">
+      <!--Perimeter of 'io' blocks with 'EMPTY' blocks at corners-->
+      <row type="io_top" starty="H-1" priority="100"/>
+      <row type="io_bottom" starty="0" priority="100"/>
+      <col type="io_left" startx="0" priority="100"/>
+      <col type="io_right" startx="W-1" priority="100"/>
+      <corners type="EMPTY" priority="101"/>
+      <!--Fill with 'clb'-->
+      <fill type="clb" priority="10"/>
+      <!--Column of 'mult_8' with 'EMPTY' blocks wherever a 'mult_8' does not fit. Vertical offset by 1 for perimeter.-->
+      <col type="mult_16" startx="2" starty="1" repeatx="8" priority="20"/>
+    </auto_layout>
+    <fixed_layout name="3x2" width="5" height="4">
+      <!--Perimeter of 'io' blocks with 'EMPTY' blocks at corners-->
+      <row type="io_top" starty="H-1" priority="100"/>
+      <row type="io_bottom" starty="0" priority="100"/>
+      <col type="io_left" startx="0" priority="100"/>
+      <col type="io_right" startx="W-1" priority="100"/>
+      <corners type="EMPTY" priority="101"/>
+      <!--Fill with 'clb'-->
+      <fill type="clb" priority="10"/>
+      <!--Column of 'mult_8' with 'EMPTY' blocks wherever a 'mult_8' does not fit. Vertical offset by 1 for perimeter.-->
+      <col type="mult_16" startx="2" starty="1" repeatx="8" priority="20"/>
+    </fixed_layout>
+    <fixed_layout name="12x12" width="14" height="14">
+      <!--Perimeter of 'io' blocks with 'EMPTY' blocks at corners-->
+      <row type="io_top" starty="H-1" priority="100"/>
+      <row type="io_bottom" starty="0" priority="100"/>
+      <col type="io_left" startx="0" priority="100"/>
+      <col type="io_right" startx="W-1" priority="100"/>
+      <corners type="EMPTY" priority="101"/>
+      <!--Fill with 'clb'-->
+      <fill type="clb" priority="10"/>
+      <!--Column of 'mult_8' with 'EMPTY' blocks wherever a 'mult_8' does not fit. Vertical offset by 1 for perimeter.-->
+      <col type="mult_16" startx="2" starty="1" repeatx="8" priority="20"/>
+    </fixed_layout>
+  </layout>
+  <device>
+    <!-- VB & JL: Using Ian Kuon's transistor sizing and drive strength data for routing, at 40 nm. Ian used BPTM 
+			     models. We are modifying the delay values however, to include metal C and R, which allows more architecture
+			     experimentation. We are also modifying the relative resistance of PMOS to be 1.8x that of NMOS
+			     (vs. Ian's 3x) as 1.8x lines up with Jeff G's data from a 45 nm process (and is more typical of 
+			     45 nm in general). I'm upping the Rmin_nmos from Ian's just over 6k to nearly 9k, and dropping 
+			     RminW_pmos from 18k to 16k to hit this 1.8x ratio, while keeping the delays of buffers approximately
+			     lined up with Stratix IV. 
+			     We are using Jeff G.'s capacitance data for 45 nm (in tech/ptm_45nm).
+			     Jeff's tables list C in for transistors with widths in multiples of the minimum feature size (45 nm).
+			     The minimum contactable transistor is 2.5 * 45 nm, so I need to multiply drive strength sizes in this file
+	                     by 2.5x when looking up in Jeff's tables.
+			     The delay values are lined up with Stratix IV, which has an architecture similar to this
+			     proposed FPGA, and which is also 40 nm 
+			     C_ipin_cblock: input capacitance of a track buffer, which VPR assumes is a single-stage
+			     4x minimum drive strength buffer. -->
+    <sizing R_minW_nmos="8926" R_minW_pmos="16067"/>
+    <!-- The grid_logic_tile_area below will be used for all blocks that do not explicitly set their own (non-routing)
+     	  area; set to 0 since we explicitly set the area of all blocks currently in this architecture file.
+	  -->
+    <area grid_logic_tile_area="0"/>
+    <chan_width_distr>
+      <x distr="uniform" peak="1.000000"/>
+      <y distr="uniform" peak="1.000000"/>
+    </chan_width_distr>
+    <switch_block type="wilton" fs="3" sub_type="subset" sub_fs="3"/>
+    <connection_block input_switch_name="ipin_cblock"/>
+  </device>
+  <switchlist>
+    <!-- VB: the mux_trans_size and buf_size data below is in minimum width transistor *areas*, assuming the purple
+	       book area formula. This means the mux transistors are about 5x minimum drive strength.
+	       We assume the first stage of the buffer is 3x min drive strength to be reasonable given the large 
+	       mux transistors, and this gives a reasonable stage ratio of a bit over 5x to the second stage. We assume
+	       the n and p transistors in the first stage are equal-sized to lower the buffer trip point, since it's fed
+	       by a pass transistor mux. We can then reverse engineer the buffer second stage to hit the specified 
+	       buf_size (really buffer area) - 16.2x minimum drive nmos and 1.8*16.2 = 29.2x minimum drive.
+	       I then took the data from Jeff G.'s PTM modeling of 45 nm to get the Cin (gate of first stage) and Cout 
+	       (diff of second stage) listed below.  Jeff's models are in tech/ptm_45nm, and are in min feature multiples.
+	       The minimum contactable transistor is 2.5 * 45 nm, so I need to multiply the drive strength sizes above by 
+	       2.5x when looking up in Jeff's tables.
+	       Finally, we choose a switch delay (58 ps) that leads to length 4 wires having a delay equal to that of SIV of 126 ps.
+	       This also leads to the switch being 46% of the total wire delay, which is reasonable. -->
+    <switch type="mux" name="L1_mux" R="551" Cin=".77e-15" Cout="4e-15" Tdel="58e-12" mux_trans_size="2.630740" buf_size="27.645901"/>
+    <switch type="mux" name="L2_mux" R="551" Cin=".77e-15" Cout="4e-15" Tdel="58e-12" mux_trans_size="2.630740" buf_size="27.645901"/>
+    <switch type="mux" name="L4_mux" R="551" Cin=".77e-15" Cout="4e-15" Tdel="58e-12" mux_trans_size="2.630740" buf_size="27.645901"/>
+    <!--switch ipin_cblock resistance set to yeild for 4x minimum drive strength buffer-->
+    <switch type="mux" name="ipin_cblock" R="2231.5" Cout="0." Cin="1.47e-15" Tdel="7.247000e-11" mux_trans_size="1.222260" buf_size="auto"/>
+  </switchlist>
+  <segmentlist>
+    <!--- VB & JL: using ITRS metal stack data, 96 nm half pitch wires, which are intermediate metal width/space.  
+			     With the 96 nm half pitch, such wires would take 60 um of height, vs. a 90 nm high (approximated as square) Stratix IV tile so this seems
+			     reasonable. Using a tile length of 90 nm, corresponding to the length of a Stratix IV tile if it were square. -->
+    <!-- GIVE a specific name for the segment! OpenFPGA appreciate that! -->
+    <segment name="L1" freq="0.10" length="1" type="unidir" Rmetal="101" Cmetal="22.5e-15">
+      <mux name="L1_mux"/>
+      <sb type="pattern">1 1</sb>
+      <cb type="pattern">1</cb>
+    </segment>
+    <segment name="L2" freq="0.10" length="2" type="unidir" Rmetal="101" Cmetal="22.5e-15">
+      <mux name="L2_mux"/>
+      <sb type="pattern">1 1 1</sb>
+      <cb type="pattern">1 1</cb>
+    </segment>
+    <segment name="L4" freq="0.80" length="4" type="unidir" Rmetal="101" Cmetal="22.5e-15">
+      <mux name="L4_mux"/>
+      <sb type="pattern">1 1 1 1 1</sb>
+      <cb type="pattern">1 1 1 1</cb>
+    </segment>
+  </segmentlist>
+  <directlist>
+    <direct name="carry_chain" from_pin="clb.cout" to_pin="clb.cin" x_offset="0" y_offset="-1" z_offset="0"/>
+    <direct name="shift_register" from_pin="clb.reg_out" to_pin="clb.reg_in" x_offset="0" y_offset="-1" z_offset="0"/>
+    <direct name="scan_chain" from_pin="clb.sc_out" to_pin="clb.sc_in" x_offset="0" y_offset="-1" z_offset="0"/>
+  </directlist>
+  <complexblocklist>
+    <!-- Define input pads begin -->
+    <pb_type name="io">
+      <input name="outpad" num_pins="1"/>
+      <output name="inpad" num_pins="1"/>
+      <!-- Do NOT add clock pins to I/O here!!! VPR does not build clock network in the way that OpenFPGA can support
+           If you need to register the I/O, define clocks in the circuit models
+           These clocks can be handled in back-end
+       -->
+      <!-- A mode denotes the physical implementation of an I/O 
+           This mode will be not packable but is mainly used for fabric verilog generation   
+        -->
+      <mode name="physical" disable_packing="true">
+        <pb_type name="iopad" blif_model=".subckt io" num_pb="1">
+          <input name="outpad" num_pins="1"/>
+          <output name="inpad" num_pins="1"/>
+        </pb_type>
+        <interconnect>
+          <direct name="outpad" input="io.outpad" output="iopad.outpad">
+            <delay_constant max="1.394e-11" in_port="io.outpad" out_port="iopad.outpad"/>
+          </direct>
+          <direct name="inpad" input="iopad.inpad" output="io.inpad">
+            <delay_constant max="4.243e-11" in_port="iopad.inpad" out_port="io.inpad"/>
+          </direct>
+        </interconnect>
+      </mode>
+
+      <!-- IOs can operate as either inputs or outputs.
+	     Delays below come from Ian Kuon. They are small, so they should be interpreted as
+	     the delays to and from registers in the I/O (and generally I/Os are registered 
+	     today and that is when you timing analyze them.
+	     -->
+      <mode name="inpad">
+        <pb_type name="inpad" blif_model=".input" num_pb="1">
+          <output name="inpad" num_pins="1"/>
+        </pb_type>
+        <interconnect>
+          <direct name="inpad" input="inpad.inpad" output="io.inpad">
+            <delay_constant max="4.243e-11" in_port="inpad.inpad" out_port="io.inpad"/>
+          </direct>
+        </interconnect>
+      </mode>
+      <mode name="outpad">
+        <pb_type name="outpad" blif_model=".output" num_pb="1">
+          <input name="outpad" num_pins="1"/>
+        </pb_type>
+        <interconnect>
+          <direct name="outpad" input="io.outpad" output="outpad.outpad">
+            <delay_constant max="1.394e-11" in_port="io.outpad" out_port="outpad.outpad"/>
+          </direct>
+        </interconnect>
+      </mode>
+      <power method="ignore"/>
+    </pb_type>
+    <!-- Define I/O pads ends -->
+    <!-- Define general purpose logic block (CLB) begin -->
+    <!-- -Due to the absence of local routing, 
+         the 4 inputs of fracturable LUT4 are no longer equivalent, 
+         because the 4th input can not be switched when the dual-LUT3 modes are used.
+         So pin equivalence should be applied to the first 3 inputs only
+	  -->
+    <pb_type name="clb">
+      <input name="I0" num_pins="2" equivalent="full"/>
+      <input name="I0i" num_pins="2" equivalent="none"/>
+      <input name="I1" num_pins="2" equivalent="full"/>
+      <input name="I1i" num_pins="2" equivalent="none"/>
+      <input name="I2" num_pins="2" equivalent="full"/>
+      <input name="I2i" num_pins="2" equivalent="none"/>
+      <input name="I3" num_pins="2" equivalent="full"/>
+      <input name="I3i" num_pins="2" equivalent="none"/>
+      <input name="I4" num_pins="2" equivalent="full"/>
+      <input name="I4i" num_pins="2" equivalent="none"/>
+      <input name="I5" num_pins="2" equivalent="full"/>
+      <input name="I5i" num_pins="2" equivalent="none"/>
+      <input name="I6" num_pins="2" equivalent="full"/>
+      <input name="I6i" num_pins="2" equivalent="none"/>
+      <input name="I7" num_pins="2" equivalent="full"/>
+      <input name="I7i" num_pins="2" equivalent="none"/>
+      <input name="reg_in" num_pins="1"/>
+      <input name="sc_in" num_pins="1"/>
+      <input name="cin" num_pins="1"/>
+      <input name="reset" num_pins="1" is_non_clock_global="true"/>
+      <output name="O" num_pins="16" equivalent="none"/>
+      <output name="reg_out" num_pins="1"/>
+      <output name="sc_out" num_pins="1"/>
+      <output name="cout" num_pins="1"/>
+      <clock name="clk" num_pins="1"/>
+      <!-- Describe fracturable logic element.  
+             Each fracturable logic element has a 6-LUT that can alternatively operate as two 5-LUTs with shared inputs. 
+             The outputs of the fracturable logic element can be optionally registered
+        -->
+      <pb_type name="fle" num_pb="8">
+        <input name="in" num_pins="4"/>
+        <input name="reg_in" num_pins="1"/>
+        <input name="sc_in" num_pins="1"/>
+        <input name="cin" num_pins="1"/>
+        <input name="reset" num_pins="1"/>
+        <output name="out" num_pins="2"/>
+        <output name="reg_out" num_pins="1"/>
+        <output name="sc_out" num_pins="1"/>
+        <output name="cout" num_pins="1"/>
+        <clock name="clk" num_pins="1"/>
+        <!-- Physical mode definition begin (physical implementation of the fle) -->
+        <mode name="physical" disable_packing="true">
+          <pb_type name="fabric" num_pb="1">
+            <input name="in" num_pins="4"/>
+            <input name="reg_in" num_pins="1"/>
+            <input name="sc_in" num_pins="1"/>
+            <input name="cin" num_pins="1"/>
+            <input name="reset" num_pins="1"/>
+            <output name="out" num_pins="2"/>
+            <output name="reg_out" num_pins="1"/>
+            <output name="sc_out" num_pins="1"/>
+            <output name="cout" num_pins="1"/>
+            <clock name="clk" num_pins="1"/>
+            <pb_type name="frac_logic" num_pb="1">
+              <input name="in" num_pins="4"/>
+              <input name="cin" num_pins="1"/>
+              <output name="out" num_pins="2"/>
+              <output name="cout" num_pins="1"/>
+              <!-- Define LUT -->
+              <pb_type name="frac_lut4" blif_model=".subckt frac_lut4" num_pb="1">
+                <input name="in" num_pins="4"/>
+                <output name="lut2_out" num_pins="2"/>
+                <output name="lut3_out" num_pins="2"/>
+                <output name="lut4_out" num_pins="1"/>
+              </pb_type>
+              <pb_type name="carry_follower" blif_model=".subckt carry_follower_physical" num_pb="1">
+                <input name="a" num_pins="1"/>
+                <input name="b" num_pins="1"/>
+                <input name="cin" num_pins="1"/>
+                <output name="cout" num_pins="1"/>
+                <delay_constant max="0.3e-9" in_port="carry_follower.a" out_port="carry_follower.cout"/>
+                <delay_constant max="0.3e-9" in_port="carry_follower.b" out_port="carry_follower.cout"/>
+                <delay_constant max="0.3e-9" in_port="carry_follower.cin" out_port="carry_follower.cout"/>
+              </pb_type>
+              <interconnect>
+                <direct name="direct1" input="frac_logic.in[0:1]" output="frac_lut4.in[0:1]"/>
+                <direct name="direct2" input="frac_logic.in[3:3]" output="frac_lut4.in[3:3]"/>
+                <direct name="direct3" input="frac_logic.cin" output="carry_follower.b"/>
+                <direct name="direct4" input="frac_lut4.lut2_out[1:1]" output="carry_follower.a"/>
+                <direct name="direct5" input="frac_lut4.lut2_out[0:0]" output="carry_follower.cin"/>
+                <direct name="direct6" input="carry_follower.cout" output="frac_logic.cout"/>
+                <direct name="direct7" input="frac_lut4.lut3_out[1]" output="frac_logic.out[1]"/>
+                <!-- Xifan Tang: I use out[0] because the output of lut6 in lut6 mode is wired to the out[0] -->
+                <mux name="mux1" input="frac_lut4.lut4_out frac_lut4.lut3_out[0]" output="frac_logic.out[0]"/>
+                <mux name="mux2" input="frac_logic.cin frac_logic.in[2:2]" output="frac_lut4.in[2:2]"/>
+              </interconnect>
+            </pb_type>
+            <!-- Define flip-flop with scan-chain capability, DI is the scan-chain data input -->
+            <pb_type name="ff" blif_model=".subckt scff" num_pb="2">
+              <input name="D" num_pins="1"/>
+              <input name="DI" num_pins="1"/>
+              <input name="reset" num_pins="1"/>
+              <output name="Q" num_pins="1"/>
+              <clock name="clk" num_pins="1"/>
+              <T_setup value="66e-12" port="ff.D" clock="clk"/>
+              <T_setup value="66e-12" port="ff.DI" clock="clk"/>
+              <T_setup value="66e-12" port="ff.reset" clock="clk"/>
+              <T_clock_to_Q max="124e-12" port="ff.Q" clock="clk"/>
+            </pb_type>         
+            <interconnect>
+              <direct name="direct1" input="fabric.in" output="frac_logic.in"/>
+              <direct name="direct2" input="fabric.cin" output="frac_logic.cin"/>
+              <direct name="direct3" input="fabric.sc_in" output="ff[0].DI"/>
+              <direct name="direct4" input="ff[0].Q" output="ff[1].DI"/>
+              <direct name="direct5" input="ff[1].Q" output="fabric.sc_out"/>
+              <direct name="direct6" input="ff[1].Q" output="fabric.reg_out"/>
+              <direct name="direct7" input="frac_logic.cout" output="fabric.cout"/>
+              <complete name="complete1" input="fabric.clk" output="ff[1:0].clk"/>
+              <complete name="complete2" input="fabric.reset" output="ff[1:0].reset"/>
+              <mux name="mux1" input="frac_logic.out[0:0] fabric.reg_in" output="ff[0:0].D">
+                <delay_constant max="25e-12" in_port="frac_logic.out[0:0]" out_port="ff[0:0].D"/>
+                <delay_constant max="45e-12" in_port="fabric.reg_in" out_port="ff[0:0].D"/>
+              </mux>
+              <mux name="mux2" input="frac_logic.out[1:1] ff[0:0].Q" output="ff[1:1].D">
+                <delay_constant max="25e-12" in_port="frac_logic.out[1:1]" out_port="ff[1:1].D"/>
+                <delay_constant max="45e-12" in_port="ff[0:0].Q" out_port="ff[1:1].D"/>
+              </mux>
+              <mux name="mux3" input="ff[0].Q frac_logic.out[0]" output="fabric.out[0]">
+                <!-- LUT to output is faster than FF to output on a Stratix IV -->
+                <delay_constant max="25e-12" in_port="frac_logic.out[0]" out_port="fabric.out[0]"/>
+                <delay_constant max="45e-12" in_port="ff[0].Q" out_port="fabric.out[0]"/>
+              </mux>
+              <mux name="mux4" input="ff[1].Q frac_logic.out[1]" output="fabric.out[1]">
+                <!-- LUT to output is faster than FF to output on a Stratix IV -->
+                <delay_constant max="25e-12" in_port="frac_logic.out[1]" out_port="fabric.out[1]"/>
+                <delay_constant max="45e-12" in_port="ff[1].Q" out_port="fabric.out[1]"/>
+              </mux>
+            </interconnect>
+          </pb_type>
+          <interconnect>
+            <direct name="direct1" input="fle.in" output="fabric.in"/>
+            <direct name="direct2" input="fle.reg_in" output="fabric.reg_in"/>
+            <direct name="direct3" input="fle.sc_in" output="fabric.sc_in"/>
+            <direct name="direct4" input="fle.cin" output="fabric.cin"/>
+            <direct name="direct5" input="fabric.out" output="fle.out"/>
+            <direct name="direct6" input="fabric.reg_out" output="fle.reg_out"/>
+            <direct name="direct7" input="fabric.sc_out" output="fle.sc_out"/>
+            <direct name="direct8" input="fabric.cout" output="fle.cout"/>
+            <direct name="direct9" input="fle.clk" output="fabric.clk"/>
+            <direct name="direct10" input="fle.reset" output="fabric.reset"/>
+          </interconnect>
+        </mode>
+        <!-- Physical mode definition end (physical implementation of the fle) -->
+        <!-- Arithmetic mode definition begin -->
+        <mode name="arithmetic">
+          <pb_type name="soft_adder" num_pb="1">
+            <input name="in" num_pins="4"/>
+            <input name="cin" num_pins="1"/>
+            <output name="sumout" num_pins="1"/>
+            <output name="cout" num_pins="1"/>
+            <!-- Define special LUT marco to be used as adder -->
+            <pb_type name="adder_lut4" blif_model=".subckt adder_lut4" num_pb="1">
+              <input name="in" num_pins="4"/>
+              <output name="lut2_out" num_pins="2"/>
+              <output name="lut4_out" num_pins="1"/>
+              <delay_constant max="0.3e-9" in_port="adder_lut4.in" out_port="adder_lut4.lut2_out"/>
+              <delay_constant max="0.3e-9" in_port="adder_lut4.in" out_port="adder_lut4.lut4_out"/>
+            </pb_type>
+            <pb_type name="carry_follower" blif_model=".subckt carry_follower" num_pb="1">
+              <input name="a" num_pins="1"/>
+              <input name="b" num_pins="1"/>
+              <input name="cin" num_pins="1"/>
+              <output name="cout" num_pins="1"/>
+              <delay_constant max="0.3e-9" in_port="carry_follower.a" out_port="carry_follower.cout"/>
+              <delay_constant max="0.3e-9" in_port="carry_follower.b" out_port="carry_follower.cout"/>
+              <delay_constant max="0.3e-9" in_port="carry_follower.cin" out_port="carry_follower.cout"/>
+            </pb_type>
+            <interconnect>
+              <direct name="direct1" input="soft_adder.in[0:1]" output="adder_lut4.in[0:1]"/>
+              <direct name="direct2" input="soft_adder.in[3:3]" output="adder_lut4.in[3:3]"/>
+              <direct name="direct3" input="soft_adder.cin" output="carry_follower.b">
+                <!-- Pack pattern to build an adder chain connection considered by packer -->
+                <pack_pattern name="chain" in_port="soft_adder.cin" out_port="carry_follower.b"/>
+              </direct>
+              <direct name="direct4" input="adder_lut4.lut2_out[1:1]" output="carry_follower.a">
+                <!-- Pack pattern to pair adder_lut4 and carry_follower into a molecule 
+                     considered by packer -->
+                <pack_pattern name="lut_follower" in_port="adder_lut4.lut2_out[1:1]" out_port="carry_follower.a"/>
+              </direct>
+              <direct name="direct5" input="adder_lut4.lut2_out[0:0]" output="carry_follower.cin">
+              </direct>
+              <direct name="direct6" input="carry_follower.cout" output="soft_adder.cout">
+                <!-- Pack pattern to build an adder chain connection considered by packer -->
+                <pack_pattern name="chain" in_port="carry_follower.cout" out_port="soft_adder.cout"/>
+              </direct>
+              <direct name="direct7" input="adder_lut4.lut4_out" output="soft_adder.sumout[0:0]">
+              </direct>
+              <mux name="mux1" input="soft_adder.cin soft_adder.in[2:2]" output="adder_lut4.in[2:2]">
+              </mux>
+            </interconnect>
+          </pb_type>
+          <interconnect>
+            <direct name="direct1" input="fle.in" output="soft_adder.in"/>
+            <direct name="direct2" input="fle.cin" output="soft_adder.cin">
+              <!-- Pack pattern to build an adder chain connection considered by packer -->
+              <pack_pattern name="chain" in_port="fle.cin" out_port="soft_adder.cin"/>
+            </direct>
+            <direct name="direct3" input="soft_adder.sumout" output="fle.out[0:0]"/>
+            <direct name="direct4" input="soft_adder.cout" output="fle.cout">
+              <!-- Pack pattern to build an adder chain connection considered by packer -->
+              <pack_pattern name="chain" in_port="soft_adder.cout" out_port="fle.cout"/>
+            </direct>
+          </interconnect>
+        </mode>
+        <!-- Arithmetic mode definition end -->
+        <!-- Dual 3-LUT mode definition begin -->
+        <mode name="n2_lut3">
+          <pb_type name="lut3inter" num_pb="1">
+            <input name="in" num_pins="3"/>
+            <output name="out" num_pins="2"/>
+            <clock name="clk" num_pins="1"/>
+            <pb_type name="ble3" num_pb="2">
+              <input name="in" num_pins="3"/>
+              <output name="out" num_pins="1"/>
+              <clock name="clk" num_pins="1"/>
+              <!-- Define the LUT -->
+              <pb_type name="lut3" blif_model=".names" num_pb="1" class="lut">
+                <input name="in" num_pins="3" port_class="lut_in"/>
+                <output name="out" num_pins="1" port_class="lut_out"/>
+                <!-- LUT timing using delay matrix -->
+                <!-- These are the physical delay inputs on a Stratix IV LUT but because VPR cannot do LUT rebalancing,
+                           we instead take the average of these numbers to get more stable results
+                      82e-12
+                      173e-12
+                      261e-12
+                      263e-12
+                      398e-12
+                      -->
+                <delay_matrix type="max" in_port="lut3.in" out_port="lut3.out">
+                  235e-12
+                  235e-12
+                  235e-12
+                </delay_matrix>
+              </pb_type>
+              <!-- Define the flip-flop -->
+              <pb_type name="ff" blif_model=".latch" num_pb="1" class="flipflop">
+                <input name="D" num_pins="1" port_class="D"/>
+                <output name="Q" num_pins="1" port_class="Q"/>
+                <clock name="clk" num_pins="1" port_class="clock"/>
+                <T_setup value="66e-12" port="ff.D" clock="clk"/>
+                <T_clock_to_Q max="124e-12" port="ff.Q" clock="clk"/>
+              </pb_type>
+              <interconnect>
+                <direct name="direct1" input="ble3.in[2:0]" output="lut3[0:0].in[2:0]"/>
+                <direct name="direct2" input="lut3[0:0].out" output="ff[0:0].D">
+                  <!-- Advanced user option that tells CAD tool to find LUT+FF pairs in netlist -->
+                  <pack_pattern name="ble3" in_port="lut3[0:0].out" out_port="ff[0:0].D"/>
+                </direct>
+                <direct name="direct3" input="ble3.clk" output="ff[0:0].clk"/>
+                <mux name="mux1" input="ff[0:0].Q lut3.out[0:0]" output="ble3.out[0:0]">
+                  <!-- LUT to output is faster than FF to output on a Stratix IV -->
+                  <delay_constant max="25e-12" in_port="lut3.out[0:0]" out_port="ble3.out[0:0]"/>
+                  <delay_constant max="45e-12" in_port="ff[0:0].Q" out_port="ble3.out[0:0]"/>
+                </mux>
+              </interconnect>
+            </pb_type>
+            <interconnect>
+              <direct name="direct1" input="lut3inter.in" output="ble3[0:0].in"/>
+              <direct name="direct2" input="lut3inter.in" output="ble3[1:1].in"/>
+              <direct name="direct3" input="ble3[1:0].out" output="lut3inter.out"/>
+              <complete name="complete1" input="lut3inter.clk" output="ble3[1:0].clk"/>
+            </interconnect>
+          </pb_type>
+          <interconnect>
+            <direct name="direct1" input="fle.in[2:0]" output="lut3inter.in"/>
+            <direct name="direct2" input="lut3inter.out" output="fle.out"/>
+            <direct name="direct3" input="fle.clk" output="lut3inter.clk"/>
+          </interconnect>
+        </mode>
+        <!-- Dual 3-LUT mode definition end -->
+        <!-- 4-LUT mode definition begin -->
+        <mode name="n1_lut4">
+          <!-- Define 4-LUT mode -->
+          <pb_type name="ble4" num_pb="1">
+            <input name="in" num_pins="4"/>
+            <output name="out" num_pins="1"/>
+            <clock name="clk" num_pins="1"/>
+            <!-- Define LUT -->
+            <pb_type name="lut4" blif_model=".names" num_pb="1" class="lut">
+              <input name="in" num_pins="4" port_class="lut_in"/>
+              <output name="out" num_pins="1" port_class="lut_out"/>
+              <!-- LUT timing using delay matrix -->
+              <!-- These are the physical delay inputs on a Stratix IV LUT but because VPR cannot do LUT rebalancing,
+                       we instead take the average of these numbers to get more stable results
+                  82e-12
+                  173e-12
+                  261e-12
+                  263e-12
+                  398e-12
+                  397e-12
+                  -->
+              <delay_matrix type="max" in_port="lut4.in" out_port="lut4.out">
+                261e-12
+                261e-12
+                261e-12
+                261e-12
+              </delay_matrix>
+            </pb_type>
+            <!-- Define flip-flop -->
+            <pb_type name="ff" blif_model=".latch" num_pb="1" class="flipflop">
+              <input name="D" num_pins="1" port_class="D"/>
+              <output name="Q" num_pins="1" port_class="Q"/>
+              <clock name="clk" num_pins="1" port_class="clock"/>
+              <T_setup value="66e-12" port="ff.D" clock="clk"/>
+              <T_clock_to_Q max="124e-12" port="ff.Q" clock="clk"/>
+            </pb_type>
+            <interconnect>
+              <direct name="direct1" input="ble4.in" output="lut4[0:0].in"/>
+              <direct name="direct2" input="lut4.out" output="ff.D">
+                <!-- Advanced user option that tells CAD tool to find LUT+FF pairs in netlist -->
+                <pack_pattern name="ble4" in_port="lut4.out" out_port="ff.D"/>
+              </direct>
+              <direct name="direct3" input="ble4.clk" output="ff.clk"/>
+              <mux name="mux1" input="ff.Q lut4.out" output="ble4.out">
+                <!-- LUT to output is faster than FF to output on a Stratix IV -->
+                <delay_constant max="25e-12" in_port="lut4.out" out_port="ble4.out"/>
+                <delay_constant max="45e-12" in_port="ff.Q" out_port="ble4.out"/>
+              </mux>
+            </interconnect>
+          </pb_type>
+          <interconnect>
+            <direct name="direct1" input="fle.in" output="ble4.in"/>
+            <direct name="direct2" input="ble4.out" output="fle.out[0:0]"/>
+            <direct name="direct3" input="fle.clk" output="ble4.clk"/>
+          </interconnect>
+        </mode>
+        <!-- 4-LUT mode definition end -->
+        <!-- Define shift register begin -->
+        <mode name="shift_register">
+          <pb_type name="shift_reg" num_pb="1">
+            <input name="reg_in" num_pins="1"/>
+            <output name="ff_out" num_pins="2"/>
+            <output name="reg_out" num_pins="1"/>
+            <clock name="clk" num_pins="1"/>
+            <pb_type name="ff" blif_model=".latch" num_pb="2" class="flipflop">
+              <input name="D" num_pins="1" port_class="D"/>
+              <output name="Q" num_pins="1" port_class="Q"/>
+              <clock name="clk" num_pins="1" port_class="clock"/>
+              <T_setup value="66e-12" port="ff.D" clock="clk"/>
+              <T_clock_to_Q max="124e-12" port="ff.Q" clock="clk"/>
+            </pb_type>
+            <interconnect>
+              <direct name="direct1" input="shift_reg.reg_in" output="ff[0].D"/>
+              <direct name="direct2" input="ff[0].Q" output="ff[1].D"/>
+              <direct name="direct3" input="ff[1].Q" output="shift_reg.reg_out"/>
+              <direct name="direct4" input="ff[0].Q" output="shift_reg.ff_out[0:0]"/>
+              <direct name="direct5" input="ff[1].Q" output="shift_reg.ff_out[1:1]"/>
+              <complete name="complete1" input="shift_reg.clk" output="ff.clk"/>
+            </interconnect>
+          </pb_type>
+          <interconnect>
+            <direct name="direct1" input="fle.reg_in" output="shift_reg.reg_in"/>
+            <direct name="direct2" input="shift_reg.reg_out" output="fle.reg_out"/>
+            <direct name="direct3" input="shift_reg.ff_out" output="fle.out"/>
+            <direct name="direct4" input="fle.clk" output="shift_reg.clk"/>
+          </interconnect>
+        </mode>
+        <!-- Define shift register end --> 
+      </pb_type>
+      <interconnect>
+        <!-- We use direct connections to reduce the area to the most
+             The global local routing is going to compensate the loss in routability
+          -->
+        <!-- FIXME: The implicit port definition results in I0[0] connected to
+                    in[2]. Such twisted connection is not expected.
+                    I[0] should be connected to in[0]
+          -->
+        <direct name="direct_fle0" input="clb.I0[0:1]" output="fle[0:0].in[0:1]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <direct name="direct_fle0i" input="clb.I0i[0:1]" output="fle[0:0].in[2:3]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <direct name="direct_fle1" input="clb.I1[0:1]" output="fle[1:1].in[0:1]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <direct name="direct_fle1i" input="clb.I1i[0:1]" output="fle[1:1].in[2:3]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <direct name="direct_fle2" input="clb.I2[0:1]" output="fle[2:2].in[0:1]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <direct name="direct_fle2i" input="clb.I2i[0:1]" output="fle[2:2].in[2:3]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <direct name="direct_fle3" input="clb.I3[0:1]" output="fle[3:3].in[0:1]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <direct name="direct_fle3i" input="clb.I3i[0:1]" output="fle[3:3].in[2:3]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <direct name="direct_fle4" input="clb.I4[0:1]" output="fle[4:4].in[0:1]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <direct name="direct_fle4i" input="clb.I4i[0:1]" output="fle[4:4].in[2:3]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <direct name="direct_fle5" input="clb.I5[0:1]" output="fle[5:5].in[0:1]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <direct name="direct_fle5i" input="clb.I5i[0:1]" output="fle[5:5].in[2:3]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <direct name="direct_fle6" input="clb.I6[0:1]" output="fle[6:6].in[0:1]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <direct name="direct_fle6i" input="clb.I6i[0:1]" output="fle[6:6].in[2:3]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <direct name="direct_fle7" input="clb.I7[0:1]" output="fle[7:7].in[0:1]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <direct name="direct_fle7i" input="clb.I7i[0:1]" output="fle[7:7].in[2:3]">
+          <!-- TODO: Timing should be backannotated from post-PnR results -->
+        </direct>
+        <complete name="clks" input="clb.clk" output="fle[7:0].clk">
+        </complete>
+        <complete name="resets" input="clb.reset" output="fle[7:0].reset">
+        </complete>
+        <!-- This way of specifying direct connection to clb outputs is important because this architecture uses automatic spreading of opins.  
+               By grouping to output pins in this fashion, if a logic block is completely filled by 6-LUTs, 
+               then the outputs those 6-LUTs take get evenly distributed across all four sides of the CLB instead of clumped on two sides (which is what happens with a more
+               naive specification).
+          -->
+        <direct name="clbouts1" input="fle[3:0].out[0:1]" output="clb.O[7:0]"/>
+        <direct name="clbouts2" input="fle[7:4].out[0:1]" output="clb.O[15:8]"/>
+        <!-- Shift register chain links -->
+        <direct name="shift_register_in" input="clb.reg_in" output="fle[0:0].reg_in">
+          <!-- Put all inter-block carry chain delay on this one edge -->
+          <delay_constant max="0.16e-9" in_port="clb.reg_in" out_port="fle[0:0].reg_in"/>
+          <!--pack_pattern name="chain" in_port="clb.reg_in" out_port="fle[0:0].reg_in"/-->
+        </direct>
+        <direct name="shift_register_out" input="fle[7:7].reg_out" output="clb.reg_out">
+          <!--pack_pattern name="chain" in_port="fle[7:7].reg_out" out_port="clb.reg_out"/-->
+        </direct>
+        <direct name="shift_register_link" input="fle[6:0].reg_out" output="fle[7:1].reg_in">
+          <!--pack_pattern name="chain" in_port="fle[6:0].reg_out" out_port="fle[7:1].reg_in"/-->
+        </direct>
+        <!-- Scan chain links -->
+        <direct name="scan_chain_in" input="clb.sc_in" output="fle[0:0].sc_in">
+          <!-- Put all inter-block carry chain delay on this one edge -->
+          <delay_constant max="0.16e-9" in_port="clb.sc_in" out_port="fle[0:0].sc_in"/>
+        </direct>
+        <direct name="scan_chain_out" input="fle[7:7].sc_out" output="clb.sc_out">
+        </direct>
+        <direct name="scan_chain_link" input="fle[6:0].sc_out" output="fle[7:1].sc_in">
+        </direct>
+        <!-- Carry chain links -->
+        <direct name="carry_chain_in" input="clb.cin" output="fle[0:0].cin">
+          <!-- Put all inter-block carry chain delay on this one edge -->
+          <pack_pattern name="chain" in_port="clb.cin" out_port="fle[0:0].cin"/>
+          <delay_constant max="0.16e-9" in_port="clb.cin" out_port="fle[0:0].cin"/>
+        </direct>
+        <direct name="carry_chain_out" input="fle[7:7].cout" output="clb.cout">
+          <pack_pattern name="chain" in_port="fle[7:7].cout" out_port="clb.cout"/>
+        </direct>
+        <direct name="carry_chain_link" input="fle[6:0].cout" output="fle[7:1].cin">
+          <pack_pattern name="chain" in_port="fle[6:0].cout" out_port="fle[7:1].cin"/>
+        </direct>
+      </interconnect>
+      <!-- Every input pin is driven by 15% of the tracks in a channel, every output pin is driven by 10% of the tracks in a channel -->
+      <!-- Place this general purpose logic block in any unspecified column -->
+    </pb_type>
+    <!-- Define general purpose logic block (CLB) ends -->
+    <!-- Define fracturable multiplier begin -->
+    <pb_type name="mult_16">
+      <input name="a" num_pins="16"/>
+      <input name="b" num_pins="16"/>
+      <output name="out" num_pins="32"/>
+      <mode name="mult_8x8">
+        <pb_type name="mult_8x8_slice" num_pb="2">
+          <input name="A_cfg" num_pins="8"/>
+          <input name="B_cfg" num_pins="8"/>
+          <output name="OUT_cfg" num_pins="16"/>
+          <pb_type name="mult_8x8" blif_model=".subckt mult_8" num_pb="1">
+            <input name="A" num_pins="8"/>
+            <input name="B" num_pins="8"/>
+            <output name="Y" num_pins="16"/>
+            <delay_constant max="1.523e-9" min="0.776e-9" in_port="mult_8x8.A" out_port="mult_8x8.Y"/>
+            <delay_constant max="1.523e-9" min="0.776e-9" in_port="mult_8x8.B" out_port="mult_8x8.Y"/>
+          </pb_type>
+          <interconnect>
+            <direct name="a2a" input="mult_8x8_slice.A_cfg" output="mult_8x8.A">
+            </direct>
+            <direct name="b2b" input="mult_8x8_slice.B_cfg" output="mult_8x8.B">
+            </direct>
+            <direct name="out2out" input="mult_8x8.Y" output="mult_8x8_slice.OUT_cfg">
+            </direct>
+          </interconnect>
+          <power method="pin-toggle">
+            <port name="A_cfg" energy_per_toggle="2.13e-12"/>
+            <port name="B_cfg" energy_per_toggle="2.13e-12"/>
+            <static_power power_per_instance="0.0"/>
+          </power>
+        </pb_type>
+        <interconnect>
+          <direct name="a2a_0" input="mult_16.a[0:7]" output="mult_8x8_slice[0].A_cfg[0:7]">
+            <delay_constant max="134e-12" min="74e-12" in_port="mult_16.a[0:7]" out_port="mult_8x8_slice[0].A_cfg[0:7]"/>
+          </direct>
+          <direct name="a2a_1" input="mult_16.a[8:15]" output="mult_8x8_slice[1].A_cfg[0:7]">
+            <delay_constant max="134e-12" min="74e-12" in_port="mult_16.a[8:15]" out_port="mult_8x8_slice[1].A_cfg[0:7]"/>
+          </direct>
+          <direct name="b2b_0" input="mult_16.b[0:7]" output="mult_8x8_slice[0].B_cfg[0:7]">
+            <delay_constant max="134e-12" min="74e-12" in_port="mult_16.b[0:7]" out_port="mult_8x8_slice[0].B_cfg[0:7]"/>
+          </direct>
+          <direct name="b2b_1" input="mult_16.b[8:15]" output="mult_8x8_slice[1].B_cfg[0:7]">
+            <delay_constant max="134e-12" min="74e-12" in_port="mult_16.b[8:15]" out_port="mult_8x8_slice[1].B_cfg[0:7]"/>
+          </direct>
+          <direct name="out2out_0" input="mult_8x8_slice[0].OUT_cfg[0:15]" output="mult_16.out[0:15]">
+            <delay_constant max="1.93e-9" min="74e-12" in_port="mult_8x8_slice[0].OUT_cfg[0:15]" out_port="mult_16.out[0:15]"/>
+          </direct>
+          <direct name="out2out_1" input="mult_8x8_slice[1].OUT_cfg[0:15]" output="mult_16.out[16:31]">
+            <delay_constant max="1.93e-9" min="74e-12" in_port="mult_8x8_slice[1].OUT_cfg[0:15]" out_port="mult_16.out[16:31]"/>
+          </direct>
+        </interconnect>
+      </mode>
+      <mode name="mult_16x16">
+        <pb_type name="mult_16x16_slice" num_pb="1">
+          <input name="A_cfg" num_pins="16"/>
+          <input name="B_cfg" num_pins="16"/>
+          <output name="OUT_cfg" num_pins="32"/>
+          <pb_type name="mult_16x16" blif_model=".subckt mult_16" num_pb="1">
+            <input name="A" num_pins="16"/>
+            <input name="B" num_pins="16"/>
+            <output name="Y" num_pins="32"/>
+            <delay_constant max="1.523e-9" min="0.776e-9" in_port="mult_16x16.A" out_port="mult_16x16.Y"/>
+            <delay_constant max="1.523e-9" min="0.776e-9" in_port="mult_16x16.B" out_port="mult_16x16.Y"/>
+          </pb_type>
+          <interconnect>
+            <direct name="a2a" input="mult_16x16_slice.A_cfg" output="mult_16x16.A">
+            </direct>
+            <direct name="b2b" input="mult_16x16_slice.B_cfg" output="mult_16x16.B">
+            </direct>
+            <direct name="out2out" input="mult_16x16.Y" output="mult_16x16_slice.OUT_cfg">
+            </direct>
+          </interconnect>
+          <power method="pin-toggle">
+            <port name="A_cfg" energy_per_toggle="2.13e-12"/>
+            <port name="B_cfg" energy_per_toggle="2.13e-12"/>
+            <static_power power_per_instance="0.0"/>
+          </power>
+        </pb_type>
+        <interconnect>
+          <direct name="a2a" input="mult_16.a" output="mult_16x16_slice.A_cfg">
+            <delay_constant max="134e-12" min="74e-12" in_port="mult_16.a" out_port="mult_16x16_slice.A_cfg"/>
+          </direct>
+          <direct name="b2b" input="mult_16.b" output="mult_16x16_slice.B_cfg">
+            <delay_constant max="134e-12" min="74e-12" in_port="mult_16.b" out_port="mult_16x16_slice.B_cfg"/>
+          </direct>
+          <direct name="out2out" input="mult_16x16_slice.OUT_cfg" output="mult_16.out">
+            <delay_constant max="1.93e-9" min="74e-12" in_port="mult_16x16_slice.OUT_cfg" out_port="mult_16.out"/>
+          </direct>
+        </interconnect>
+      </mode>
+      <!-- Place this multiplier block every 8 columns from (and including) the sixth column -->
+      <power method="sum-of-children"/>
+    </pb_type>
+    <!-- Define fracturable multiplier end -->
+  </complexblocklist>
+</architecture>

From 148da80869b020aba0c73ccd6dbbc3ab982be15c Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Sat, 24 Apr 2021 14:53:29 -0600
Subject: [PATCH 11/17] [Tool] Add new syntax about
 physical_pb_port_rotate_offset to support fracturable heterogeneous block
 mapping between operating modes and physical modes

---
 .../src/pb_type_annotation.cpp                | 40 +++++++++++---
 .../libarchopenfpga/src/pb_type_annotation.h  | 24 ++++++---
 .../src/read_xml_pb_type_annotation.cpp       | 25 +++++++++
 .../src/write_xml_pb_type_annotation.cpp      |  9 +++-
 openfpga/src/annotation/annotate_pb_graph.cpp | 26 +++++++++-
 openfpga/src/annotation/annotate_pb_types.cpp |  3 +-
 .../src/annotation/vpr_device_annotation.cpp  | 52 +++++++++++++++++++
 .../src/annotation/vpr_device_annotation.h    | 15 +++++-
 openfpga/src/repack/repack.cpp                |  6 +--
 9 files changed, 178 insertions(+), 22 deletions(-)

diff --git a/libopenfpga/libarchopenfpga/src/pb_type_annotation.cpp b/libopenfpga/libarchopenfpga/src/pb_type_annotation.cpp
index 9a369e742..1a1a7ea31 100644
--- a/libopenfpga/libarchopenfpga/src/pb_type_annotation.cpp
+++ b/libopenfpga/libarchopenfpga/src/pb_type_annotation.cpp
@@ -86,11 +86,11 @@ std::vector<std::string> PbTypeAnnotation::port_names() const {
   return keys;
 }
 
-std::map<BasicPort, std::array<int, 2>> PbTypeAnnotation::physical_pb_type_port(const std::string& port_name) const {
-  std::map<std::string, std::map<BasicPort, std::array<int, 2>>>::const_iterator it = operating_pb_type_ports_.find(port_name);
+std::map<BasicPort, std::array<int, 3>> PbTypeAnnotation::physical_pb_type_port(const std::string& port_name) const {
+  std::map<std::string, std::map<BasicPort, std::array<int, 3>>>::const_iterator it = operating_pb_type_ports_.find(port_name);
   if (it == operating_pb_type_ports_.end()) {
     /* Return an empty port */
-    return std::map<BasicPort, std::array<int, 2>>();
+    return std::map<BasicPort, std::array<int, 3>>();
   }
   return operating_pb_type_ports_.at(port_name);
 }
@@ -169,25 +169,25 @@ void PbTypeAnnotation::set_physical_pb_type_index_offset(const int& value) {
 void PbTypeAnnotation::add_pb_type_port_pair(const std::string& operating_pb_port_name,
                                              const BasicPort& physical_pb_port) {
   /* Give a warning if the operating_pb_port_name already exist */
-  std::map<std::string, std::map<BasicPort, std::array<int, 2>>>::const_iterator it = operating_pb_type_ports_.find(operating_pb_port_name);
+  std::map<std::string, std::map<BasicPort, std::array<int, 3>>>::const_iterator it = operating_pb_type_ports_.find(operating_pb_port_name);
 
   /* If not exist, initialize and set a default value */
   if (it == operating_pb_type_ports_.end()) {
-    operating_pb_type_ports_[operating_pb_port_name][physical_pb_port] = {0, 0};
+    operating_pb_type_ports_[operating_pb_port_name][physical_pb_port] = {0, 0, 0};
     /* We can return early */
     return;
   }
 
   /* If the physical port is not in the list, we create one and set a default value */
   if (0 == operating_pb_type_ports_[operating_pb_port_name].count(physical_pb_port)) {
-    operating_pb_type_ports_[operating_pb_port_name][physical_pb_port] = {0, 0};
+    operating_pb_type_ports_[operating_pb_port_name][physical_pb_port] = {0, 0, 0};
   }
 }
 
 void PbTypeAnnotation::set_physical_pin_initial_offset(const std::string& operating_pb_port_name,
                                                        const BasicPort& physical_pb_port,
                                                        const int& physical_pin_initial_offset) {
-  std::map<std::string, std::map<BasicPort, std::array<int, 2>>>::const_iterator it = operating_pb_type_ports_.find(operating_pb_port_name);
+  std::map<std::string, std::map<BasicPort, std::array<int, 3>>>::const_iterator it = operating_pb_type_ports_.find(operating_pb_port_name);
 
   if (it == operating_pb_type_ports_.end()) {
     VTR_LOG_ERROR("The operating pb_type port '%s' is not valid!\n",
@@ -210,7 +210,7 @@ void PbTypeAnnotation::set_physical_pin_initial_offset(const std::string& operat
 void PbTypeAnnotation::set_physical_pin_rotate_offset(const std::string& operating_pb_port_name,
                                                       const BasicPort& physical_pb_port,
                                                       const int& physical_pin_rotate_offset) {
-  std::map<std::string, std::map<BasicPort, std::array<int, 2>>>::const_iterator it = operating_pb_type_ports_.find(operating_pb_port_name);
+  std::map<std::string, std::map<BasicPort, std::array<int, 3>>>::const_iterator it = operating_pb_type_ports_.find(operating_pb_port_name);
 
   if (it == operating_pb_type_ports_.end()) {
     VTR_LOG_ERROR("The operating pb_type port '%s' is not valid!\n",
@@ -230,6 +230,30 @@ void PbTypeAnnotation::set_physical_pin_rotate_offset(const std::string& operati
   operating_pb_type_ports_[operating_pb_port_name][physical_pb_port][1] = physical_pin_rotate_offset;
 }
 
+void PbTypeAnnotation::set_physical_port_rotate_offset(const std::string& operating_pb_port_name,
+                                                       const BasicPort& physical_pb_port,
+                                                       const int& physical_port_rotate_offset) {
+  std::map<std::string, std::map<BasicPort, std::array<int, 3>>>::const_iterator it = operating_pb_type_ports_.find(operating_pb_port_name);
+
+  if (it == operating_pb_type_ports_.end()) {
+    VTR_LOG_ERROR("The operating pb_type port '%s' is not valid!\n",
+                  operating_pb_port_name.c_str());
+    exit(1);
+  }
+
+  if (operating_pb_type_ports_[operating_pb_port_name].end() == operating_pb_type_ports_[operating_pb_port_name].find(physical_pb_port)) {
+    VTR_LOG_ERROR("The physical pb_type port '%s[%lu:%lu]' definition for operating pb_type port '%s' is not valid!\n",
+                  physical_pb_port.get_name().c_str(),
+                  physical_pb_port.get_lsb(),
+                  physical_pb_port.get_msb(),
+                  operating_pb_port_name.c_str());
+    exit(1);
+  }
+
+  operating_pb_type_ports_[operating_pb_port_name][physical_pb_port][2] = physical_port_rotate_offset;
+}
+
+
 void PbTypeAnnotation::add_interconnect_circuit_model_pair(const std::string& interc_name,
                                                            const std::string& circuit_model_name) {
   std::map<std::string, std::string>::const_iterator it = interconnect_circuit_model_names_.find(interc_name);
diff --git a/libopenfpga/libarchopenfpga/src/pb_type_annotation.h b/libopenfpga/libarchopenfpga/src/pb_type_annotation.h
index 96383679d..2a0b8804e 100644
--- a/libopenfpga/libarchopenfpga/src/pb_type_annotation.h
+++ b/libopenfpga/libarchopenfpga/src/pb_type_annotation.h
@@ -49,7 +49,7 @@ class PbTypeAnnotation {
     float physical_pb_type_index_factor() const;
     int physical_pb_type_index_offset() const;
     std::vector<std::string> port_names() const;
-    std::map<BasicPort, std::array<int, 2>> physical_pb_type_port(const std::string& port_name) const;
+    std::map<BasicPort, std::array<int, 3>> physical_pb_type_port(const std::string& port_name) const;
     std::vector<std::string> interconnect_names() const;
     std::string interconnect_circuit_model_name(const std::string& interc_name) const;
   public: /* Public mutators */
@@ -73,6 +73,9 @@ class PbTypeAnnotation {
     void set_physical_pin_rotate_offset(const std::string& operating_pb_port_name,
                                         const BasicPort& physical_pb_port,
                                         const int& physical_pin_rotate_offset);
+    void set_physical_port_rotate_offset(const std::string& operating_pb_port_name,
+                                         const BasicPort& physical_pb_port,
+                                         const int& physical_port_rotate_offset);
     void add_interconnect_circuit_model_pair(const std::string& interc_name,
                                              const std::string& circuit_model_name);
   private: /* Internal data */
@@ -138,10 +141,10 @@ class PbTypeAnnotation {
     int physical_pb_type_index_offset_;
 
     /* Link from the pins under an operating pb_type to pairs of
-     * its physical pb_type and its pin initial & rotating offset
-     *
-     * Note that initial offset is the first element of the std::array
-     * Note that rotating offset is the second element of the std::array
+     * its physical pb_type and 
+     * - its pin initial offset: the first element of the std::array
+     * - pin-level rotating offset: the second element of the std::array
+     * - port-level rotating offset: the third element of the std::array
      *
      * The offsets aim to align the pin indices for port of pb_type 
      * between operating and physical modes, especially when an operating 
@@ -158,14 +161,21 @@ class PbTypeAnnotation {
      *   physical pb_type bram[0].dout_a[0] with a full path memory[physical].bram[0]
      *   physical pb_type bram[0].dout_a[1] with a full path memory[physical].bram[0]
      *
-     * For example, a rotating offset of 9 is used to map 
+     * For example, a pin-level rotating offset of 9 is used to map 
+     *   operating pb_type mult_9x9[0].a[0] with a full path mult[frac].mult_9x9[0]
+     *   operating pb_type mult_9x9[0].a[1] with a full path mult[frac].mult_9x9[1]
+     * to 
+     *   physical pb_type mult_36x36.a[0] with a full path mult[physical].mult_36x36[0]
+     *   physical pb_type mult_36x36.a[9] with a full path mult[physical].mult_36x36[0]
+     *
+     * For example, a port-level rotating offset of 9 is used to map 
      *   operating pb_type mult_9x9[0].a[0:8] with a full path mult[frac].mult_9x9[0]
      *   operating pb_type mult_9x9[1].a[0:8] with a full path mult[frac].mult_9x9[1]
      * to 
      *   physical pb_type mult_36x36.a[0:8] with a full path mult[physical].mult_36x36[0]
      *   physical pb_type mult_36x36.a[9:17] with a full path mult[physical].mult_36x36[0]
      */
-    std::map<std::string, std::map<BasicPort, std::array<int, 2>>> operating_pb_type_ports_;
+    std::map<std::string, std::map<BasicPort, std::array<int, 3>>> operating_pb_type_ports_;
 
     /* Link between the interconnects under this pb_type and circuit model names */
     std::map<std::string, std::string> interconnect_circuit_model_names_;
diff --git a/libopenfpga/libarchopenfpga/src/read_xml_pb_type_annotation.cpp b/libopenfpga/libarchopenfpga/src/read_xml_pb_type_annotation.cpp
index 9831f6e23..f91dc99a8 100644
--- a/libopenfpga/libarchopenfpga/src/read_xml_pb_type_annotation.cpp
+++ b/libopenfpga/libarchopenfpga/src/read_xml_pb_type_annotation.cpp
@@ -114,6 +114,31 @@ void read_xml_pb_port_annotation(pugi::xml_node& xml_port,
                                                         std::stoi(rotate_offsets[iport]));
     }
   }
+
+  /* We have an optional attribute: physical_mode_port_rotate_offset
+   * Split based on the number of physical pb_type ports that have been defined
+   */
+  const std::string& physical_port_rotate_offset_attr = get_attribute(xml_port, "physical_mode_port_rotate_offset", loc_data, pugiutil::ReqOpt::OPTIONAL).as_string();
+
+  if (false == physical_port_rotate_offset_attr.empty()) {
+    /* Split the physical mode port attributes with space */
+    openfpga::StringToken offset_tokenizer(physical_port_rotate_offset_attr);
+    const std::vector<std::string> rotate_offsets = offset_tokenizer.split();
+   
+    /* Error out if the offset does not match the port definition */
+    if (physical_mode_ports.size() != rotate_offsets.size()) {
+      archfpga_throw(loc_data.filename_c_str(), loc_data.line(xml_port),
+                     "Defined %lu physical mode ports but only %lu physical port rotate offset are defined! Expect size matching.\n",
+                      physical_mode_ports.size(), rotate_offsets.size());
+    }
+
+    for (size_t iport = 0; iport < physical_mode_ports.size(); ++iport) {
+      openfpga::PortParser port_parser(physical_mode_ports[iport]); 
+      pb_type_annotation.set_physical_port_rotate_offset(name_attr,
+                                                         port_parser.port(),
+                                                         std::stoi(rotate_offsets[iport]));
+    }
+  }
 }
 
 /********************************************************************
diff --git a/libopenfpga/libarchopenfpga/src/write_xml_pb_type_annotation.cpp b/libopenfpga/libarchopenfpga/src/write_xml_pb_type_annotation.cpp
index e9b36522c..b401f6ca8 100644
--- a/libopenfpga/libarchopenfpga/src/write_xml_pb_type_annotation.cpp
+++ b/libopenfpga/libarchopenfpga/src/write_xml_pb_type_annotation.cpp
@@ -144,7 +144,14 @@ void write_xml_pb_port_annotation(std::fstream& fp,
     physical_mode_pin_rotate_offset_attr += std::to_string(physical_pb_port_pair.second[1]);
   }
   write_xml_attribute(fp, "physical_mode_pin_rotate_offset", physical_mode_pin_rotate_offset_attr.c_str()); 
-
+  std::string physical_mode_port_rotate_offset_attr;
+  for (const auto& physical_pb_port_pair : pb_type_annotation.physical_pb_type_port(port_name)) {
+    if (false == physical_mode_port_rotate_offset_attr.empty()) {
+      physical_mode_port_rotate_offset_attr += " ";
+    }
+    physical_mode_port_rotate_offset_attr += std::to_string(physical_pb_port_pair.second[2]);
+  }
+  write_xml_attribute(fp, "physical_mode_port_rotate_offset", physical_mode_port_rotate_offset_attr.c_str()); 
   fp << "/>" << "\n";
 }
 
diff --git a/openfpga/src/annotation/annotate_pb_graph.cpp b/openfpga/src/annotation/annotate_pb_graph.cpp
index 67a1baba8..bed7997cc 100644
--- a/openfpga/src/annotation/annotate_pb_graph.cpp
+++ b/openfpga/src/annotation/annotate_pb_graph.cpp
@@ -333,7 +333,7 @@ bool try_match_pb_graph_pin(t_pb_graph_pin* operating_pb_graph_pin,
      *     by the pin rotate offset value
      *     The accumulated offset will be reset to 0 when it exceeds the msb() of the physical port
      */
-    int acc_offset = vpr_device_annotation.physical_pb_pin_offset(operating_pb_graph_pin->port, candidate_port);
+    int acc_offset = vpr_device_annotation.physical_pb_pin_offset(operating_pb_graph_pin->port, candidate_port) + vpr_device_annotation.physical_pb_port_offset(operating_pb_graph_pin->port, candidate_port);
     int init_offset = vpr_device_annotation.physical_pb_pin_initial_offset(operating_pb_graph_pin->port, candidate_port);
     const BasicPort& physical_port_range = vpr_device_annotation.physical_pb_port_range(operating_pb_graph_pin->port, candidate_port);
     if (physical_pb_graph_pin->pin_number != operating_pb_graph_pin->pin_number
@@ -463,6 +463,14 @@ void annotate_physical_pb_graph_node_pins(t_pb_graph_node* operating_pb_graph_no
                                      physical_pb_graph_node, vpr_device_annotation,
                                      verbose_output);
     }
+    /* Finish a port, accumulate the port-level offset affiliated to the port */
+    if (0 == operating_pb_graph_node->num_input_pins[iport]) {
+      continue;
+    }
+    t_pb_graph_pin* operating_pb_graph_pin = &(operating_pb_graph_node->input_pins[iport][0]);
+    for (t_port* candidate_port : vpr_device_annotation.physical_pb_port(operating_pb_graph_pin->port)) {
+      vpr_device_annotation.accumulate_physical_pb_port_rotate_offset(operating_pb_graph_pin->port, candidate_port);
+    }
   }
 
   for (int iport = 0; iport < operating_pb_graph_node->num_output_ports; ++iport) {
@@ -471,6 +479,14 @@ void annotate_physical_pb_graph_node_pins(t_pb_graph_node* operating_pb_graph_no
                                      physical_pb_graph_node, vpr_device_annotation,
                                      verbose_output);
     }
+    /* Finish a port, accumulate the port-level offset affiliated to the port */
+    if (0 == operating_pb_graph_node->num_output_pins[iport]) {
+      continue;
+    }
+    t_pb_graph_pin* operating_pb_graph_pin = &(operating_pb_graph_node->output_pins[iport][0]);
+    for (t_port* candidate_port : vpr_device_annotation.physical_pb_port(operating_pb_graph_pin->port)) {
+      vpr_device_annotation.accumulate_physical_pb_port_rotate_offset(operating_pb_graph_pin->port, candidate_port);
+    }
   }
 
   for (int iport = 0; iport < operating_pb_graph_node->num_clock_ports; ++iport) {
@@ -479,6 +495,14 @@ void annotate_physical_pb_graph_node_pins(t_pb_graph_node* operating_pb_graph_no
                                      physical_pb_graph_node, vpr_device_annotation,
                                      verbose_output);
     }
+    /* Finish a port, accumulate the port-level offset affiliated to the port */
+    if (0 == operating_pb_graph_node->num_clock_pins[iport]) {
+      continue;
+    }
+    t_pb_graph_pin* operating_pb_graph_pin = &(operating_pb_graph_node->clock_pins[iport][0]);
+    for (t_port* candidate_port : vpr_device_annotation.physical_pb_port(operating_pb_graph_pin->port)) {
+      vpr_device_annotation.accumulate_physical_pb_port_rotate_offset(operating_pb_graph_pin->port, candidate_port);
+    }
   }
 }
 
diff --git a/openfpga/src/annotation/annotate_pb_types.cpp b/openfpga/src/annotation/annotate_pb_types.cpp
index 92a6f6560..e4b5f972c 100644
--- a/openfpga/src/annotation/annotate_pb_types.cpp
+++ b/openfpga/src/annotation/annotate_pb_types.cpp
@@ -211,7 +211,7 @@ bool pair_operating_and_physical_pb_types(t_pb_type* operating_pb_type,
    * if not found, we assume that the physical port is the same as the operating pb_port
    */
   for (t_port* operating_pb_port : pb_type_ports(operating_pb_type)) {
-    std::map<BasicPort, std::array<int, 2>> expected_physical_pb_ports = pb_type_annotation.physical_pb_type_port(std::string(operating_pb_port->name));
+    std::map<BasicPort, std::array<int, 3>> expected_physical_pb_ports = pb_type_annotation.physical_pb_type_port(std::string(operating_pb_port->name));
   
     /* If not defined in the annotation, set the default pair:
      * rotate_offset is 0 by default!
@@ -243,6 +243,7 @@ bool pair_operating_and_physical_pb_types(t_pb_type* operating_pb_type,
       vpr_device_annotation.add_physical_pb_port_range(operating_pb_port, physical_pb_port, expected_physical_pb_port.first);
       vpr_device_annotation.add_physical_pb_pin_initial_offset(operating_pb_port, physical_pb_port, expected_physical_pb_port.second[0]);
       vpr_device_annotation.add_physical_pb_pin_rotate_offset(operating_pb_port, physical_pb_port, expected_physical_pb_port.second[1]);
+      vpr_device_annotation.add_physical_pb_port_rotate_offset(operating_pb_port, physical_pb_port, expected_physical_pb_port.second[2]);
     }
   }
 
diff --git a/openfpga/src/annotation/vpr_device_annotation.cpp b/openfpga/src/annotation/vpr_device_annotation.cpp
index c2eed00ab..eef4fafbb 100644
--- a/openfpga/src/annotation/vpr_device_annotation.cpp
+++ b/openfpga/src/annotation/vpr_device_annotation.cpp
@@ -220,6 +220,21 @@ int VprDeviceAnnotation::physical_pb_pin_rotate_offset(t_port* operating_pb_port
   return physical_pb_pin_rotate_offsets_.at(operating_pb_port).at(physical_pb_port);
 }
 
+int VprDeviceAnnotation::physical_pb_port_rotate_offset(t_port* operating_pb_port,
+                                                        t_port* physical_pb_port) const {
+  /* Ensure that the pb_type is in the list */
+  std::map<t_port*, std::map<t_port*, int>>::const_iterator it = physical_pb_port_rotate_offsets_.find(operating_pb_port);
+  if (it == physical_pb_port_rotate_offsets_.end()) {
+    /* Default value is 0 */
+    return 0;
+  }
+  if (0 == physical_pb_port_rotate_offsets_.at(operating_pb_port).count(physical_pb_port)) {
+    /* Default value is 0 */
+    return 0;
+  }
+  return physical_pb_port_rotate_offsets_.at(operating_pb_port).at(physical_pb_port);
+}
+
 int VprDeviceAnnotation::physical_pb_pin_offset(t_port* operating_pb_port,
                                                 t_port* physical_pb_port) const {
   /* Ensure that the pb_type is in the list */
@@ -235,6 +250,21 @@ int VprDeviceAnnotation::physical_pb_pin_offset(t_port* operating_pb_port,
   return physical_pb_pin_offsets_.at(operating_pb_port).at(physical_pb_port);
 }
 
+int VprDeviceAnnotation::physical_pb_port_offset(t_port* operating_pb_port,
+                                                t_port* physical_pb_port) const {
+  /* Ensure that the pb_type is in the list */
+  std::map<t_port*, std::map<t_port*, int>>::const_iterator it = physical_pb_port_offsets_.find(operating_pb_port);
+  if (it == physical_pb_port_offsets_.end()) {
+    /* Default value is 0 */
+    return 0;
+  }
+  if (0 == physical_pb_port_offsets_.at(operating_pb_port).count(physical_pb_port)) {
+    /* Default value is 0 */
+    return 0;
+  }
+  return physical_pb_port_offsets_.at(operating_pb_port).at(physical_pb_port);
+}
+
 t_pb_graph_pin* VprDeviceAnnotation::physical_pb_graph_pin(const t_pb_graph_pin* pb_graph_pin) const {
   /* Ensure that the pb_type is in the list */
   std::map<const t_pb_graph_pin*, t_pb_graph_pin*>::const_iterator it = physical_pb_graph_pins_.find(pb_graph_pin);
@@ -478,6 +508,28 @@ void VprDeviceAnnotation::add_physical_pb_pin_initial_offset(t_port* operating_p
   physical_pb_pin_initial_offsets_[operating_pb_port][physical_pb_port] = offset;
 }
 
+void VprDeviceAnnotation::add_physical_pb_port_rotate_offset(t_port* operating_pb_port,
+                                                             t_port* physical_pb_port,
+                                                             const int& offset) {
+  /* Warn any override attempt */
+  std::map<t_port*, std::map<t_port*, int>>::const_iterator it = physical_pb_port_rotate_offsets_.find(operating_pb_port);
+  if ( (it != physical_pb_port_rotate_offsets_.end())
+    && (0 < physical_pb_port_rotate_offsets_[operating_pb_port].count(physical_pb_port)) ) {
+    VTR_LOG_WARN("Override the annotation between operating pb_port '%s' and it physical pb_port '%s' port rotate offset '%d'!\n",
+                 operating_pb_port->name, offset);
+  }
+
+  physical_pb_port_rotate_offsets_[operating_pb_port][physical_pb_port] = offset;
+  /* We initialize the accumulated offset to 0 */
+  physical_pb_port_offsets_[operating_pb_port][physical_pb_port] = 0;
+}
+
+
+void VprDeviceAnnotation::accumulate_physical_pb_port_rotate_offset(t_port* operating_pb_port,
+                                                                    t_port* physical_pb_port) {
+  physical_pb_port_offsets_[operating_pb_port][physical_pb_port] += physical_pb_port_rotate_offsets_[operating_pb_port][physical_pb_port];
+}
+
 void VprDeviceAnnotation::add_physical_pb_pin_rotate_offset(t_port* operating_pb_port,
                                                             t_port* physical_pb_port,
                                                             const int& offset) {
diff --git a/openfpga/src/annotation/vpr_device_annotation.h b/openfpga/src/annotation/vpr_device_annotation.h
index b5afbda00..d1a0f8e99 100644
--- a/openfpga/src/annotation/vpr_device_annotation.h
+++ b/openfpga/src/annotation/vpr_device_annotation.h
@@ -67,6 +67,9 @@ class VprDeviceAnnotation {
     int physical_pb_pin_rotate_offset(t_port* operating_pb_port,
                                       t_port* physical_pb_port) const;
 
+    int physical_pb_port_rotate_offset(t_port* operating_pb_port,
+                                       t_port* physical_pb_port) const;
+
     /**This function returns an accumulated offset. Note that the
      * accumulated offset is NOT the pin rotate offset specified by users
      * It is an aggregation of the offset during pin pairing
@@ -76,6 +79,8 @@ class VprDeviceAnnotation {
      */
     int physical_pb_pin_offset(t_port* operating_pb_port,
                                t_port* physical_pb_port) const;
+    int physical_pb_port_offset(t_port* operating_pb_port,
+                                t_port* physical_pb_port) const;
     t_pb_graph_pin* physical_pb_graph_pin(const t_pb_graph_pin* pb_graph_pin) const;
     CircuitModelId rr_switch_circuit_model(const RRSwitchId& rr_switch) const;
     CircuitModelId rr_segment_circuit_model(const RRSegmentId& rr_segment) const;
@@ -106,6 +111,11 @@ class VprDeviceAnnotation {
     void add_physical_pb_pin_initial_offset(t_port* operating_pb_port,
                                             t_port* physical_pb_port,
                                             const int& offset);
+    void add_physical_pb_port_rotate_offset(t_port* operating_pb_port,
+                                            t_port* physical_pb_port,
+                                            const int& offset);
+    void accumulate_physical_pb_port_rotate_offset(t_port* operating_pb_port,
+                                                   t_port* physical_pb_port);
     void add_physical_pb_pin_rotate_offset(t_port* operating_pb_port,
                                            t_port* physical_pb_port,
                                            const int& offset);
@@ -166,8 +176,11 @@ class VprDeviceAnnotation {
     std::map<t_port*, std::vector<t_port*>> physical_pb_ports_;
     std::map<t_port*, std::map<t_port*, int>> physical_pb_pin_initial_offsets_;
     std::map<t_port*, std::map<t_port*, int>> physical_pb_pin_rotate_offsets_;
+    std::map<t_port*, std::map<t_port*, int>> physical_pb_port_rotate_offsets_;
 
-    /* Accumulated offsets for a physical pb_type port, just for internal usage */
+    /* Accumulated offsets for a physical pb port, just for internal usage */
+    std::map<t_port*, std::map<t_port*, int>> physical_pb_port_offsets_;
+    /* Accumulated offsets for a physical pb_graph_pin, just for internal usage */
     std::map<t_port*, std::map<t_port*, int>> physical_pb_pin_offsets_;
 
     /* Pair a pb_port to its LSB and MSB of a physical pb_port 
diff --git a/openfpga/src/repack/repack.cpp b/openfpga/src/repack/repack.cpp
index d6cb669bc..7538c675a 100644
--- a/openfpga/src/repack/repack.cpp
+++ b/openfpga/src/repack/repack.cpp
@@ -61,9 +61,9 @@ bool rec_direct_search_sink_pb_graph_pins(const t_pb_graph_pin* source_pb_pin,
   std::vector<t_pb_graph_pin*> sink_pb_pins_to_search;
 
   /* Only support single-mode pb_type!!! */
-  if (1 != source_pb_pin->parent_node->pb_type->num_modes) {
-    return false;
-  }
+  //if (1 != source_pb_pin->parent_node->pb_type->num_modes) {
+  //  return false;
+  //}
 
   for (int iedge = 0; iedge < source_pb_pin->num_output_edges; ++iedge) {
     if (DIRECT_INTERC != source_pb_pin->output_edges[iedge]->interconnect->type) {

From a3a98fa21d6365a67f57db0c1ef2ae11a9237df8 Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Sat, 24 Apr 2021 14:56:10 -0600
Subject: [PATCH 12/17] [Arch] Bug fix for port name mismatching between
 openfpga cell library and architecture definition

---
 ...frac_dsp16_caravel_io_skywater130nm_fdhd_cc_openfpga.xml | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/openfpga_flow/openfpga_arch/k4_frac_N8_reset_softadder_register_scan_chain_frac_dsp16_caravel_io_skywater130nm_fdhd_cc_openfpga.xml b/openfpga_flow/openfpga_arch/k4_frac_N8_reset_softadder_register_scan_chain_frac_dsp16_caravel_io_skywater130nm_fdhd_cc_openfpga.xml
index 7ce8c5c28..78978c2c7 100644
--- a/openfpga_flow/openfpga_arch/k4_frac_N8_reset_softadder_register_scan_chain_frac_dsp16_caravel_io_skywater130nm_fdhd_cc_openfpga.xml
+++ b/openfpga_flow/openfpga_arch/k4_frac_N8_reset_softadder_register_scan_chain_frac_dsp16_caravel_io_skywater130nm_fdhd_cc_openfpga.xml
@@ -210,9 +210,9 @@
       <design_technology type="cmos"/>
       <input_buffer exist="true" circuit_model_name="INVTX1"/>
       <output_buffer exist="true" circuit_model_name="INVTX1"/>
-      <port type="input" prefix="A" lib_name="A" size="16"/>
-      <port type="input" prefix="B" lib_name="B" size="16"/>
-      <port type="output" prefix="Y" lib_name="Y" size="32"/>
+      <port type="input" prefix="A" lib_name="a" size="16"/>
+      <port type="input" prefix="B" lib_name="b" size="16"/>
+      <port type="output" prefix="Y" lib_name="out" size="32"/>
       <port type="sram" prefix="mode" lib_name="mode" size="1" mode_select="true" circuit_model_name="DFFRQ" default_val="0"/>
     </circuit_model>
   </circuit_library>

From 8b8096f3a860e655ebebd644f8673a3b5c2e11cc Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Sat, 24 Apr 2021 14:57:09 -0600
Subject: [PATCH 13/17] [HDL] Bug fix in HDL modeling of multi-mode 16-bit DSP
 block

---
 openfpga_flow/openfpga_cell_library/verilog/frac_mult_16x16.v | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/openfpga_flow/openfpga_cell_library/verilog/frac_mult_16x16.v b/openfpga_flow/openfpga_cell_library/verilog/frac_mult_16x16.v
index 7fedc711c..d73a1c7e1 100644
--- a/openfpga_flow/openfpga_cell_library/verilog/frac_mult_16x16.v
+++ b/openfpga_flow/openfpga_cell_library/verilog/frac_mult_16x16.v
@@ -13,7 +13,7 @@ module frac_mult_16x16 (
 	output [0:31] out,
     input [0:0] mode);
 
-    reg [0:63] out_reg;
+    reg [0:31] out_reg;
 
     always @(mode, a, b) begin 
        if (1'b1 == mode) begin

From 80f98328df34411a6baaaf6b4ac57ee59a79c762 Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Sat, 24 Apr 2021 15:16:50 -0600
Subject: [PATCH 14/17] [Test] Update test settings for architecture with
 fracturable DSP blocks

---
 .../dsp/multi_mode_mult_16x16/config/task.conf            | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/openfpga_flow/tasks/fpga_verilog/dsp/multi_mode_mult_16x16/config/task.conf b/openfpga_flow/tasks/fpga_verilog/dsp/multi_mode_mult_16x16/config/task.conf
index 0db63ef8b..b9e2c957b 100644
--- a/openfpga_flow/tasks/fpga_verilog/dsp/multi_mode_mult_16x16/config/task.conf
+++ b/openfpga_flow/tasks/fpga_verilog/dsp/multi_mode_mult_16x16/config/task.conf
@@ -24,7 +24,7 @@ yosys_cell_sim_verilog=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_yosys_techli
 yosys_dsp_map_verilog=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_yosys_techlib/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm_dsp_map.v
 yosys_dsp_map_parameters=-D DSP_A_MAXWIDTH=8 -D DSP_B_MAXWIDTH=8 -D DSP_A_MINWIDTH=2 -D DSP_B_MINWIDTH=2 -D DSP_NAME=mult_8x8
 # VPR parameter
-openfpga_vpr_device_layout=3x2
+openfpga_vpr_device_layout=3x4
 
 [ARCHITECTURES]
 arch0=${PATH:OPENFPGA_PATH}/openfpga_flow/vpr_arch/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm.xml
@@ -32,7 +32,8 @@ arch0=${PATH:OPENFPGA_PATH}/openfpga_flow/vpr_arch/k4_frac_N8_tileable_reset_sof
 [BENCHMARKS]
 bench0=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_4/mac_4.v
 bench1=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_8/mac_8.v
-bench2=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_16/mac_16.v
+bench2=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_12/mac_12.v
+bench3=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_16/mac_16.v
 
 [SYNTHESIS_PARAM]
 bench_yosys_common=${PATH:OPENFPGA_PATH}/openfpga_flow/misc/ys_tmpl_yosys_vpr_dsp_flow.ys
@@ -40,7 +41,8 @@ bench_yosys_rewrite_common=${PATH:OPENFPGA_PATH}/openfpga_flow/misc/ys_tmpl_yosy
 
 bench0_top = mac_4
 bench1_top = mac_8
-bench2_top = mac_16
+bench2_top = mac_12
+bench3_top = mac_16
 
 [SCRIPT_PARAM_MIN_ROUTE_CHAN_WIDTH]
 end_flow_with_test=

From 5adffad6023b8dad5bbbc75a9ab2de21b794e836 Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Sat, 24 Apr 2021 15:49:53 -0600
Subject: [PATCH 15/17] [Arch] Changes to the arch to avoid a bug where the
 rr_nodes at top side of a heterogenenous block have no fan-in!!!

---
 ...ac_dsp16_nonLR_caravel_io_skywater130nm.xml | 18 ++++++++++--------
 1 file changed, 10 insertions(+), 8 deletions(-)

diff --git a/openfpga_flow/vpr_arch/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm.xml b/openfpga_flow/vpr_arch/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm.xml
index 98bae4fa2..2d817e7cd 100644
--- a/openfpga_flow/vpr_arch/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm.xml
+++ b/openfpga_flow/vpr_arch/k4_frac_N8_tileable_reset_softadder_register_scan_chain_frac_dsp16_nonLR_caravel_io_skywater130nm.xml
@@ -111,7 +111,7 @@
          These clocks can be handled in back-end
      -->
     <!-- Top-side has 1 I/O per tile -->
-    <tile name="io_top" capacity="1" area="0">
+    <tile name="io_top" capacity="9" area="0">
       <equivalent_sites>
         <site pb_type="io"/>
       </equivalent_sites>
@@ -123,7 +123,7 @@
       </pinlocations>
     </tile>
     <!-- Right-side has 1 I/O per tile -->
-    <tile name="io_right" capacity="1" area="0">
+    <tile name="io_right" capacity="9" area="0">
       <equivalent_sites>
         <site pb_type="io"/>
       </equivalent_sites>
@@ -147,7 +147,7 @@
       </pinlocations>
     </tile>
     <!-- Left-side has 1 I/O per tile -->
-    <tile name="io_left" capacity="1" area="0">
+    <tile name="io_left" capacity="9" area="0">
       <equivalent_sites>
         <site pb_type="io"/>
       </equivalent_sites>
@@ -217,10 +217,12 @@
       <!--  Highly recommand to customize pin location when direct connection is used!!! -->
       <!--pinlocations pattern="spread"/-->
       <pinlocations pattern="custom">
-        <loc side="left">mult_16.a[0:5] mult_16.b[0:5] mult_16.out[0:10]</loc>
-        <loc side="top">mult_16.a[6:7] mult_16.b[6:7] mult_16.out[11:15]</loc>
-        <loc side="right">mult_16.a[8:13] mult_16.b[8:13] mult_16.out[16:26]</loc>
-        <loc side="bottom">mult_16.a[14:15] mult_16.b[14:15] mult_16.out[27:31]</loc>
+        <loc side="left" yoffset="0">mult_16.a[0:2] mult_16.b[0:2] mult_16.out[0:5]</loc>
+        <loc side="left" yoffset="1">mult_16.a[3:5] mult_16.b[3:5] mult_16.out[6:10]</loc>
+        <loc side="top"></loc>
+        <loc side="right" yoffset="0">mult_16.a[8:10] mult_16.b[8:10] mult_16.out[16:21]</loc>
+        <loc side="right" yoffset="1">mult_16.a[11:13] mult_16.b[11:13] mult_16.out[22:26]</loc>
+        <loc side="bottom">mult_16.a[6:7] mult_16.b[6:7] mult_16.out[11:15] mult_16.a[14:15] mult_16.b[14:15] mult_16.out[27:31]</loc>
       </pinlocations>
     </tile>
   </tiles>
@@ -239,7 +241,7 @@
       <!--Column of 'mult_8' with 'EMPTY' blocks wherever a 'mult_8' does not fit. Vertical offset by 1 for perimeter.-->
       <col type="mult_16" startx="2" starty="1" repeatx="8" priority="20"/>
     </auto_layout>
-    <fixed_layout name="3x2" width="5" height="4">
+    <fixed_layout name="3x4" width="5" height="6">
       <!--Perimeter of 'io' blocks with 'EMPTY' blocks at corners-->
       <row type="io_top" starty="H-1" priority="100"/>
       <row type="io_bottom" starty="0" priority="100"/>

From b7da22501c4ff35e56b6af623aceadaad967331f Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Sat, 24 Apr 2021 15:55:05 -0600
Subject: [PATCH 16/17] [Test] Deply new test to regression test

---
 openfpga_flow/regression_test_scripts/fpga_verilog_reg_test.sh | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/openfpga_flow/regression_test_scripts/fpga_verilog_reg_test.sh b/openfpga_flow/regression_test_scripts/fpga_verilog_reg_test.sh
index 0033ad8ac..9168cbd1e 100755
--- a/openfpga_flow/regression_test_scripts/fpga_verilog_reg_test.sh
+++ b/openfpga_flow/regression_test_scripts/fpga_verilog_reg_test.sh
@@ -47,6 +47,9 @@ run-task fpga_verilog/bram/wide_dpram16k --debug --show_thread_logs
 echo -e "Testing Verilog generation with heterogeneous fabric using 8-bit single-mode multipliers ";
 run-task fpga_verilog/dsp/single_mode_mult_8x8 --debug --show_thread_logs
 
+echo -e "Testing Verilog generation with heterogeneous fabric using 16-bit multi-mode multipliers ";
+run-task fpga_verilog/dsp/multi_mode_mult_16x16 --debug --show_thread_logs
+
 echo -e "Testing Verilog generation with different I/O capacities on each side of an FPGA ";
 run-task fpga_verilog/io/multi_io_capacity --debug --show_thread_logs
 

From 62dc5a3856035a009363f3862a0053b66908b72d Mon Sep 17 00:00:00 2001
From: tangxifan <tangxifan@gmail.com>
Date: Sat, 24 Apr 2021 16:02:24 -0600
Subject: [PATCH 17/17] [Doc] Update documentation about the new syntax
 introduced for pin binding between operating modes and physical modes

---
 .../manual/arch_lang/annotate_vpr_arch.rst    | 27 ++++++++++++++++---
 1 file changed, 23 insertions(+), 4 deletions(-)

diff --git a/docs/source/manual/arch_lang/annotate_vpr_arch.rst b/docs/source/manual/arch_lang/annotate_vpr_arch.rst
index 239d10dfb..f4767c938 100644
--- a/docs/source/manual/arch_lang/annotate_vpr_arch.rst
+++ b/docs/source/manual/arch_lang/annotate_vpr_arch.rst
@@ -223,7 +223,9 @@ The ``circuit_model_name`` should match the given name of a ``circuit_model`` de
   .. note:: A ``<pb_type name="<string>">`` parent XML node is required for the interconnect-to-circuit bindings whose interconnects are defined under the ``pb_type`` in VPR architecture description. 
 
 .. option:: <port name="<string>" physical_mode_port="<string>"
-             physical_mode_pin_initial_offset="<int>" physical_mode_pin_rotate_offset="<int>"/>
+             physical_mode_pin_initial_offset="<int>"
+             physical_mode_pin_rotate_offset="<int>"/>
+             physical_mode_port_rotate_offset="<int>"/>
 
    Link a port of an operating ``pb_type`` to a port of a physical ``pb_type``
 
@@ -233,7 +235,6 @@ The ``circuit_model_name`` should match the given name of a ``circuit_model`` de
 
     .. note:: Users can define multiple ports. For example: ``physical_mode_pin="a[0:1] b[2:2]"``. When multiple ports are used, the ``physical_mode_pin_initial_offset`` and ``physical_mode_pin_rotate_offset`` should also be adapt. For example: ``physical_mode_pin_rotate_offset="1 0"``)
 
-
   - ``physical_mode_pin_initial_offset="<int>"`` aims to align the pin indices for ``port`` of ``pb_type`` between operating and physical modes, especially when part of port of operating mode is mapped to a port in physical ``pb_type``. When ``physical_mode_pin_initial_offset`` is larger than zero, the pin index of ``pb_type`` (whose index is large than 1) will be shifted by the given offset. 
 
     .. note:: A quick example to understand the initial offset
@@ -249,7 +250,24 @@ The ``circuit_model_name`` should match the given name of a ``circuit_model`` de
 
     .. note:: If not defined, the default value of ``physical_mode_pin_initial_offset`` is set to ``0``.
 
-  - ``physical_mode_pin_rotate_offset="<int>"`` aims to align the pin indices for ``port`` of ``pb_type`` between operating and physical modes, especially when an operating mode contains multiple ``pb_type`` (``num_pb``>1) that are linked to the same physical ``pb_type``. When ``physical_mode_pin_rotate_offset`` is larger than zero, the pin index of ``pb_type`` (whose index is large than 1) will be shifted by the given offset.
+  - ``physical_mode_pin_rotate_offset="<int>"`` aims to align the pin indices for ``port`` of ``pb_type`` between operating and physical modes, especially when an operating mode contains multiple ``pb_type`` (``num_pb``>1) that are linked to the same physical ``pb_type``. When ``physical_mode_pin_rotate_offset`` is larger than zero, the pin index of ``pb_type`` (whose index is large than 1) will be shifted by the given offset, **each time a pin in the operating mode is binded to a pin in the physical mode**.
+  
+    .. note:: A quick example to understand the rotate offset
+              For example, a rotating offset of 9 is used to map 
+
+              - operating pb_type ``mult_9x9[0].a[0]`` with a full path ``mult[frac].mult_9x9[0]``
+              - operating pb_type ``mult_9x9[1].a[1]`` with a full path ``mult[frac].mult_9x9[1]``
+
+               to 
+
+              - physical pb_type ``mult_36x36.a[0]`` with a full path ``mult[physical].mult_36x36[0]``
+              - physical pb_type ``mult_36x36.a[9]`` with a full path ``mult[physical].mult_36x36[0]``
+
+    .. note:: If not defined, the default value of ``physical_mode_pin_rotate_offset`` is set to ``0``.
+
+  .. warning:: The result of using ``physical_mode_pin_rotate_offset`` is fundementally different than ``physical_mode_port_rotate_offset``!!! Please read the examples carefully and pick the one fitting your needs.
+
+  - ``physical_mode_port_rotate_offset="<int>"`` aims to align the port indices for ``port`` of ``pb_type`` between operating and physical modes, especially when an operating mode contains multiple ``pb_type`` (``num_pb``>1) that are linked to the same physical ``pb_type``. When ``physical_mode_port_rotate_offset`` is larger than zero, the pin index of ``pb_type`` (whose index is large than 1) will be shifted by the given offset, **only when all the pins of a port in the operating mode is binded to all the pins of a port in the physical mode**.
   
     .. note:: A quick example to understand the rotate offset
               For example, a rotating offset of 9 is used to map 
@@ -262,7 +280,8 @@ The ``circuit_model_name`` should match the given name of a ``circuit_model`` de
               - physical pb_type ``mult_36x36.a[0:8]`` with a full path ``mult[physical].mult_36x36[0]``
               - physical pb_type ``mult_36x36.a[9:17]`` with a full path ``mult[physical].mult_36x36[0]``
 
-    .. note:: If not defined, the default value of ``physical_mode_pin_rotate_offset`` is set to ``0``.
+    .. note:: If not defined, the default value of ``physical_mode_port_rotate_offset`` is set to ``0``.
+
 
 .. note::
   It is highly recommended that only one physical mode is defined for a multi-mode configurable block. Try not to use nested physical mode definition. This will ease the debugging and lead to clean XML description.