Merge pull request #339 from lnis-uofu/micro_benchmark

Micro benchmark

.github/workflows/build.yml

@@ -210,6 +210,7 @@ jobs:
           - name: fpga_bitstream_reg_test
           - name: fpga_sdc_reg_test
           - name: fpga_spice_reg_test
+          - name: micro_benchmark_reg_test
           - name: quicklogic_reg_test
           - name: vtr_benchmark_reg_test
           - name: iwls_benchmark_reg_test
@@ -257,6 +258,7 @@ jobs:
           - name: fpga_bitstream_reg_test
           - name: fpga_sdc_reg_test
           - name: fpga_spice_reg_test
+          - name: micro_benchmark_reg_test
           - name: quicklogic_reg_test
           - name: vtr_benchmark_reg_test
           - name: iwls_benchmark_reg_test

openfpga_flow/benchmarks/micro_benchmark/and4/and4.v

@@ -0,0 +1,22 @@
+/////////////////////////////////////////
+//  Functionality: 4-input AND 
+//  Author:        Xifan Tang
+////////////////////////////////////////
+`timescale 1ns / 1ps
+
+module and4(
+  a,
+  b,
+  c,
+  d,
+  e);
+
+input wire a;
+input wire b;
+input wire c;
+input wire d;
+output wire e;
+
+assign e = a & b & c & d;
+
+endmodule

openfpga_flow/benchmarks/micro_benchmark/counter_128bit_async_reset/counter.v

@@ -0,0 +1,25 @@
+///////////////////////////////////////////
+//  Functionality: Counter with asynchronous reset
+//  Author:        Xifan Tang
+////////////////////////////////////////
+
+module counter (
+	clk,
+	reset,
+	result
+);
+
+	input clk;
+	input reset;
+	output [127:0] result;
+
+	reg [127:0] result;
+
+	always @(posedge clk or posedge reset)
+	begin
+		if (reset) 
+			result = 0;		
+		else 
+			result = result + 1;
+	end
+endmodule		

openfpga_flow/benchmarks/micro_benchmark/counter_128bit_async_reset/counter_tb.v

@@ -0,0 +1,25 @@
+module counter_tb;
+
+  reg clk, reset;
+  wire [127:0] result;
+
+  counter DUT(
+    .clk(clk), 
+    .reset(reset), 
+    .result(result)
+  );
+  
+  initial begin
+    #0 reset = 1'b1; clk = 1'b0;
+    #100 reset = 1'b0;
+  end
+
+  always begin
+    #10 clk = ~clk;
+  end
+  
+  initial begin
+    #5000 $stop;
+  end
+
+endmodule

openfpga_flow/benchmarks/micro_benchmark/counter_128bit_async_resetb/counter.v

@@ -0,0 +1,25 @@
+///////////////////////////////////////////
+//  Functionality: Counter with asynchronous reset
+//  Author:        Xifan Tang
+////////////////////////////////////////
+
+module counter (
+	clk,
+	resetb,
+	result
+);
+
+	input clk;
+	input resetb;
+	output [127:0] result;
+
+	reg [127:0] result;
+
+	always @(posedge clk or negedge resetb)
+	begin
+		if (~resetb) 
+			result = 0;		
+		else 
+			result = result + 1;
+	end
+endmodule		

openfpga_flow/benchmarks/micro_benchmark/counter_128bit_async_resetb/counter_tb.v

@@ -0,0 +1,25 @@
+module counter_tb;
+
+  reg clk, resetb;
+  wire [127:0] result;
+
+  counter DUT(
+    .clk(clk), 
+    .resetb(resetb), 
+    .result(result)
+  );
+  
+  initial begin
+    #0 reset = 1'b0; clk = 1'b0;
+    #100 reset = 1'b1;
+  end
+
+  always begin
+    #10 clk = ~clk;
+  end
+  
+  initial begin
+    #5000 $stop;
+  end
+
+endmodule

openfpga_flow/benchmarks/micro_benchmark/mac/mac_18/mac_18.v

@@ -0,0 +1,22 @@
+//-------------------------------------------------------
+//  Functionality: A 18-bit multiply-acculumate circuit
+//  Author:        Xifan Tang
+//-------------------------------------------------------
+
+module mac_18(a, b, c, out);
+parameter DATA_WIDTH = 18;  /* 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
+
+
+
+
+
+
+
+
+

openfpga_flow/benchmarks/micro_benchmark/mac/mac_20/mac_20.v

@@ -0,0 +1,22 @@
+//-------------------------------------------------------
+//  Functionality: A 20-bit multiply-acculumate circuit
+//  Author:        Xifan Tang
+//-------------------------------------------------------
+
+module mac_20(a, b, c, out);
+parameter DATA_WIDTH = 20;  /* 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
+
+
+
+
+
+
+
+
+

openfpga_flow/benchmarks/micro_benchmark/mac/mac_36/mac_36.v

@@ -0,0 +1,22 @@
+//-------------------------------------------------------
+//  Functionality: A 36-bit multiply-acculumate circuit
+//  Author:        Xifan Tang
+//-------------------------------------------------------
+
+module mac_36(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
+
+
+
+
+
+
+
+
+

openfpga_flow/benchmarks/micro_benchmark/mac/mac_8_9/mac_8_9.v

@@ -0,0 +1,25 @@
+//-------------------------------------------------------
+//  Functionality: 
+//    - A 8-bit multiply-acculumate circuit 
+//    - A 9-bit multiply-acculumate circuit
+//  Author:        Xifan Tang
+//-------------------------------------------------------
+
+module mac_8_9(a, b, c, out);
+parameter DATA_WIDTH = 18;  /* declare a parameter. default required */
+input [DATA_WIDTH - 1 : 0] a, b, c;
+output [DATA_WIDTH - 1 : 0] out;
+
+assign out[8:0] = a[8:0] * b[8:0] + c[8:0];
+assign out[17:9] = a[17:9] * b[17:9] + c[17:9];
+
+endmodule
+
+
+
+
+
+
+
+
+

openfpga_flow/benchmarks/micro_benchmark/mac/mac_9/mac_9.v

@@ -0,0 +1,22 @@
+//-------------------------------------------------------
+//  Functionality: A 9-bit multiply-acculumate circuit
+//  Author:        Xifan Tang
+//-------------------------------------------------------
+
+module mac_9(a, b, c, out);
+parameter DATA_WIDTH = 9;  /* 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
+
+
+
+
+
+
+
+
+

openfpga_flow/openfpga_shell_scripts/mcnc_example_script.openfpga

@@ -37,7 +37,7 @@ repack #--verbose
 
 # Build the bitstream
 #  - Output the fabric-independent bitstream to a file
-build_architecture_bitstream --verbose --write_file fabric_indepenent_bitstream.xml
+build_architecture_bitstream --verbose
 
 # Build fabric-dependent bitstream
 build_fabric_bitstream --verbose
@@ -46,9 +46,7 @@ build_fabric_bitstream --verbose
 #  - Enable the use of explicit port mapping in Verilog netlist
 write_fabric_verilog --file ./SRC \
   --explicit_port_mapping \
-  --include_timing \
-  --include_signal_init
-# --support_icarus_simulator
+  --include_timing
 
 # Write the Verilog testbench for FPGA fabric
 #  - We suggest the use of same output directory as fabric Verilog netlists
@@ -59,16 +57,6 @@ write_fabric_verilog --file ./SRC \
 write_preconfigured_fabric_wrapper --file ./SRC --support_icarus_simulator
 write_preconfigured_testbench --file ./SRC --reference_benchmark_file_path ${REFERENCE_VERILOG_TESTBENCH} --support_icarus_simulator
 
-# Write the SDC files for PnR backend
-#  - Turn on every options here
-write_pnr_sdc --file ./SDC
-
-# Write SDC to disable timing for configure ports
-write_sdc_disable_timing_configure_ports --file ./SDC/disable_configure_ports.sdc
-
-# Write the SDC to run timing analysis for a mapped FPGA fabric
-write_analysis_sdc --file ./SDC_analysis
-
 # Finish and exit OpenFPGA
 exit
 

openfpga_flow/regression_test_scripts/fpga_verilog_reg_test.sh

@@ -136,9 +136,3 @@ run-task fpga_verilog/fully_connected_output_crossbar --debug --show_thread_logs
 echo -e "Testing through channels in tileable routing";
 run-task fpga_verilog/thru_channel/thru_narrow_tile --debug --show_thread_logs
 run-task fpga_verilog/thru_channel/thru_wide_tile --debug --show_thread_logs
-
-# Verify MCNC big20 benchmark suite with ModelSim
-# Please make sure you have ModelSim installed in the environment
-# Otherwise, it will fail
-#run-task fpga_verilog/mcnc_big20 --debug --show_thread_logs --maxthreads 20
-#python3 openfpga_flow/scripts/run_modelsim.py mcnc_big20 --run_sim

openfpga_flow/regression_test_scripts/micro_benchmark_reg_test.sh

@@ -0,0 +1,17 @@
+#!/bin/bash
+
+set -e
+source openfpga.sh
+PYTHON_EXEC=python3.8
+###############################################
+# OpenFPGA Shell with VPR8
+##############################################
+echo -e "Micro benchmark regression tests";
+run-task benchmark_sweep/counter --debug --show_thread_logs
+run-task benchmark_sweep/mac_units --debug --show_thread_logs
+
+# Verify MCNC big20 benchmark suite with ModelSim
+# Please make sure you have ModelSim installed in the environment
+# Otherwise, it will fail
+run-task benchmark_sweep/mcnc_big20 --debug --show_thread_logs
+#python3 openfpga_flow/scripts/run_modelsim.py mcnc_big20 --run_sim

openfpga_flow/tasks/benchmark_sweep/counter/config/task.conf

@@ -9,29 +9,45 @@
 [GENERAL]
 run_engine=openfpga_shell
 power_tech_file = ${PATH:OPENFPGA_PATH}/openfpga_flow/tech/PTM_45nm/45nm.xml
-power_analysis = true
+power_analysis = false
 spice_output=false
 verilog_output=true
-timeout_each_job = 20*60
-fpga_flow=vpr_blif
+timeout_each_job = 5*60
+fpga_flow=yosys_vpr
 
 [OpenFPGA_SHELL]
-openfpga_shell_template=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_shell_scripts/implicit_verilog_example_script.openfpga
-openfpga_arch_file=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_arch/k6_N10_40nm_openfpga.xml
-openfpga_sim_setting_file=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_simulation_settings/auto_sim_openfpga.xml
+openfpga_shell_template=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_shell_scripts/iwls_benchmark_example_script.openfpga
+openfpga_arch_file=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_arch/k6_frac_N10_adder_chain_dpram8K_dsp36_fracff_40nm_openfpga.xml
+openfpga_sim_setting_file=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_simulation_settings/fixed_sim_openfpga.xml
+# Yosys script parameters
+yosys_cell_sim_verilog=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_yosys_techlib/k6_frac_N10_tileable_adder_chain_dpram8K_dsp36_fracff_40nm_cell_sim.v
+yosys_dff_map_verilog=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_yosys_techlib/k6_frac_N10_tileable_adder_chain_dpram8K_dsp36_fracff_40nm_dff_map.v
+yosys_bram_map_rules=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_yosys_techlib/k6_frac_N10_tileable_adder_chain_dpram8K_dsp36_40nm_bram.txt
+yosys_bram_map_verilog=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_yosys_techlib/k6_frac_N10_tileable_adder_chain_dpram8K_dsp36_40nm_bram_map.v
+yosys_dsp_map_verilog=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_yosys_techlib/k6_frac_N10_tileable_adder_chain_dpram8K_dsp36_40nm_dsp_map.v
+yosys_dsp_map_parameters=-D DSP_A_MAXWIDTH=36 -D DSP_B_MAXWIDTH=36 -D DSP_A_MINWIDTH=2 -D DSP_B_MINWIDTH=2 -D DSP_NAME=mult_36x36
+# VPR parameters
+# # Use a fixed routing channel width to save runtime
+vpr_route_chan_width=50
 
 [ARCHITECTURES]
-arch0=${PATH:OPENFPGA_PATH}/openfpga_flow/vpr_arch/k6_N10_40nm.xml
+arch0=${PATH:OPENFPGA_PATH}/openfpga_flow/vpr_arch/k6_frac_N10_tileable_adder_chain_dpram8K_dsp36_fracff_40nm.xml
 
 [BENCHMARKS]
-bench0=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/counter/counter_post_yosys.blif
+bench0=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/counter/counter.v
+bench1=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/counter_async_reset/counter.v
+bench2=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/counter_128bit_async_reset/counter.v
+bench3=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/counter_128bit_async_resetb/counter.v
 
 [SYNTHESIS_PARAM]
+bench_yosys_common=${PATH:OPENFPGA_PATH}/openfpga_flow/misc/ys_tmpl_yosys_vpr_bram_dsp_dff_flow.ys
+#bench_yosys_rewrite_common=${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 = counter
-bench0_act = ${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/counter/counter_pre_vpr.act
-bench0_verilog = ${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/counter/counter_output_verilog.v
-bench0_chan_width = 300
+bench1_top = counter
+bench2_top = counter
+bench3_top = counter
 
 [SCRIPT_PARAM_MIN_ROUTE_CHAN_WIDTH]
-end_flow_with_test=
+#end_flow_with_test=
 #vpr_fpga_verilog_formal_verification_top_netlist=

openfpga_flow/tasks/benchmark_sweep/mac_units/config/task.conf

@@ -0,0 +1,67 @@
+# = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
+# 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/iwls_benchmark_example_script.openfpga
+openfpga_arch_file=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_arch/k6_frac_N10_adder_chain_dpram8K_dsp36_fracff_40nm_openfpga.xml
+openfpga_sim_setting_file=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_simulation_settings/fixed_sim_openfpga.xml
+# Yosys script parameters
+yosys_cell_sim_verilog=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_yosys_techlib/k6_frac_N10_tileable_adder_chain_dpram8K_dsp36_fracff_40nm_cell_sim.v
+yosys_dff_map_verilog=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_yosys_techlib/k6_frac_N10_tileable_adder_chain_dpram8K_dsp36_fracff_40nm_dff_map.v
+yosys_bram_map_rules=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_yosys_techlib/k6_frac_N10_tileable_adder_chain_dpram8K_dsp36_40nm_bram.txt
+yosys_bram_map_verilog=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_yosys_techlib/k6_frac_N10_tileable_adder_chain_dpram8K_dsp36_40nm_bram_map.v
+yosys_dsp_map_verilog=${PATH:OPENFPGA_PATH}/openfpga_flow/openfpga_yosys_techlib/k6_frac_N10_tileable_adder_chain_dpram8K_dsp36_40nm_dsp_map.v
+yosys_dsp_map_parameters=-D DSP_A_MAXWIDTH=36 -D DSP_B_MAXWIDTH=36 -D DSP_A_MINWIDTH=2 -D DSP_B_MINWIDTH=2 -D DSP_NAME=mult_36x36
+# VPR parameters
+# Use a fixed routing channel width to save runtime
+vpr_route_chan_width=300
+
+[ARCHITECTURES]
+arch0=${PATH:OPENFPGA_PATH}/openfpga_flow/vpr_arch/k6_frac_N10_tileable_adder_chain_dpram8K_dsp36_fracff_40nm.xml
+
+[BENCHMARKS]
+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
+bench4=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_8_9/mac_8_9.v
+bench5=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_9/mac_9.v
+bench6=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_12/mac_12.v
+bench7=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_16/mac_16.v
+bench8=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_18/mac_18.v
+bench9=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_20/mac_20.v
+bench10=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_32/mac_32.v
+bench11=${PATH:OPENFPGA_PATH}/openfpga_flow/benchmarks/micro_benchmark/mac/mac_36/mac_36.v
+
+[SYNTHESIS_PARAM]
+bench_yosys_common=${PATH:OPENFPGA_PATH}/openfpga_flow/misc/ys_tmpl_yosys_vpr_bram_dsp_dff_flow.ys
+bench0_top = mac_2
+bench1_top = mac_4
+bench2_top = mac_6
+bench3_top = mac_8
+bench4_top = mac_8_9
+bench5_top = mac_9
+bench6_top = mac_12
+bench7_top = mac_16
+bench8_top = mac_18
+bench9_top = mac_20
+bench10_top = mac_32
+bench11_top = mac_36
+
+[SCRIPT_PARAM_MIN_ROUTE_CHAN_WIDTH]
+#end_flow_with_test=
+#vpr_fpga_verilog_formal_verification_top_netlist=