This commit is contained in:
Aurelien Alacchi 2018-10-09 15:29:18 -06:00
commit 201c0797b2
9 changed files with 471 additions and 4 deletions

View File

@ -24,9 +24,9 @@ You can find in the folder **./examples**. This will help you get in touch with
In all the examples, the CLBs are composed of LUTs, FFs and MUXs as a base.
Example 1 shows a very basic design with only 4 inputs on the LUTs and 1 Basic Element in the CLBs. It implements an inverter and allows the user to see the very core of the .xml file.
Example 1 shows a very basic design with only 4 inputs on the LUTs, a FF and a MUX in the CLB (only 1). It implements an inverter and allows the user to see the very core of the .xml file.
Example 2 increases the complexity by having 3x3 CLBs and 4 slices per CLB. The slices provide a feedback to the input structure and input MUXs are added.

View File

@ -0,0 +1,77 @@
# Examples_README
## Example_1
The goal of this example is just to make a first step into the software. The .blif contains only an inverter.
The .xml is currently on <layout auto="1.0"/> which means that the size depends on the .blif. Since the .blif is
almost empty, only 1 CLB will be generated.
<img src="https://github.com/LNIS-Projects/OpenFPGA/blob/master/examples/figures/example_1.png" alt="Example_1_FPGA" width="60%">
Schematic of the FPGA generated during example_1.
The CLB integrates a 4-inputs LUT, a FF and a MUX.
---
**Things to understand in this example**
Everything won't be explained in detail but few important structures (some common with the VPR project) are to be explained in order to build good architectures.
```xml
<architecture>
<models>
... add models such as the io pads.
</models>
<spice_settings>
... all tech and spice parameters are defined here.
<module_spice_models>
... define the Basic Elements of the architecture and the modules that cannot
be generated (i.e. the Flip-Flop) but need to be called.
</module_spice_models>
</spice_settings>
<cblocks>
... complex blocks
<complexblocklist>
... here we define the hierarchy of the primitive blocks and interconnect them
together
<pb_type>
... defines the primitive block
</pb_type>
<interconnect>
...
</interconnect>
</complexblocklist>
</cblocks>
</architecture>
```
---
## Example_2
Example_2's goal is to introduce the slices, the interconnections which can be generated from it and the manual mode of the layout.
In this case, we generate a 3x3 FPGA with 4 slices. The LUTs are 6-inputs ones similarly to the ones used in the industry.
There is a feedback-loop from the output of the slices to the input MUXs
<img src="https://github.com/LNIS-Projects/OpenFPGA/blob/master/examples/figures/example_2_the_CLB.png" alt="Example_2_CLB" width="60%">
Schematic showing the CLB generated in this example.
---
**Things to understand in this example**
```xml
<layout width="3" height="3"/> <!-- Manual mode of the layout allowing us to choose the number of CLBs -->
<complete name="crossbar" input="clb.I fle[3:0].out" output="fle[3:0].in" spice_model_name="mux_2level">
<!-- Defines how we apply the feedback on the inputs of the slices -->
<delay_constant max="53.44e-12" in_port="clb.I" out_port="fle[3:0].in" />
<delay_constant max="53.44e-12" in_port="fle[3:0].out" out_port="fle[3:0].in" />
</complete>
```
<img src="https://github.com/LNIS-Projects/OpenFPGA/blob/master/examples/figures/example_2_3x3.png" alt="Example_2_FPGA" width="60%">

View File

@ -1,7 +1,7 @@
# Baudouin Chauviere University of Utah 30 September 2018
# Benchmark doing an inverter
.model s298.bench
.model inverter.bench
.inputs I0 clk
.outputs Q0

13
examples/example_2.blif Normal file
View File

@ -0,0 +1,13 @@
# Baudouin Chauviere University of Utah 30 September 2018
# Benchmark doing an inverter
.model example_2.bench
.inputs I0 clk
.outputs Q0
.latch n0 Q0 re clk 0
.names I0 n0
0 1
.end

9
examples/example_2.sh Executable file
View File

@ -0,0 +1,9 @@
#!/bin/sh
# Example of how to run vpr
# Pack, place, and route a heterogeneous FPGA
# Packing uses the AAPack algorithm
../vpr7_x2p/vpr/vpr ./example_2.xml ./example_2.blif --full_stats --nodisp --route_chan_width 30 --fpga_spice --fpga_spice_rename_illegal_port --fpga_spice_dir ./spice_test --fpga_spice_print_top_testbench --fpga_spice_print_grid_testbench --fpga_spice_print_cb_testbench --fpga_spice_print_sb_testbench --fpga_spice_print_lut_testbench --fpga_spice_print_hardlogic_testbench --fpga_spice_print_pb_mux_testbench --fpga_spice_print_cb_mux_testbench --fpga_spice_print_sb_mux_testbench --fpga_verilog --fpga_verilog_dir ./verilog_test

368
examples/example_2.xml Normal file
View File

@ -0,0 +1,368 @@
<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="io">
<input_ports>
<port name="outpad"/>
</input_ports>
<output_ports>
<port name="inpad"/>
</output_ports>
</model>
</models>
<!-- ODIN II specific config ends -->
<!-- Physical descriptions begin -->
<!-- <layout auto="1.0"/> --> <!-- Takes the smallest square possible that fits the number of LUTs needed -->
<layout width="3" height="3"/>
<spice_settings>
<parameters>
<options sim_temp="25" post="off" captab="off" fast="on"/>
<!-- Used only when doing monte_carlo simulations
<monte_carlo mc_sim="off" num_mc_points="2" cmos_variation="off" rram_variation="off">
<cmos abs_variation="0.1" num_sigma="3"/>
<rram abs_variation="0.1" num_sigma="3"/>
</monte_carlo>
-->
<measure sim_num_clock_cycle="auto" accuracy="1e-13" accuracy_type="abs">
<slew>
<rise upper_thres_pct="0.95" lower_thres_pct="0.05"/>
<fall upper_thres_pct="0.05" lower_thres_pct="0.95"/>
</slew>
<delay>
<rise input_thres_pct="0.5" output_thres_pct="0.5"/>
<fall input_thres_pct="0.5" output_thres_pct="0.5"/>
</delay>
</measure>
<stimulate>
<clock op_freq="auto" sim_slack="0.2" prog_freq="2.5e6">
<rise slew_time="20e-12" slew_type="abs"/>
<fall slew_time="20e-12" slew_type="abs"/>
</clock>
<input>
<rise slew_time="25e-12" slew_type="abs"/>
<fall slew_time="25e-12" slew_type="abs"/>
</input>
</stimulate>
</parameters>
<tech_lib lib_type="industry" transistor_type="TOP_TT" lib_path="/research/ece/lnis/USERS/tang/tangxifan-eda-tools/branches/subvt_fpga/process/tsmc40nm/toplevel.l" nominal_vdd="0.9" io_vdd="2.5"/>
<transistors pn_ratio="2" model_ref="M">
<nmos model_name="nch" chan_length="40e-9" min_width="140e-9"/>
<pmos model_name="pch" chan_length="40e-9" min_width="140e-9"/>
<io_nmos model_name="nch_25" chan_length="270e-9" min_width="320e-9"/>
<io_pmos model_name="pch_25" chan_length="270e-9" min_width="320e-9"/>
</transistors>
<module_spice_models>
<spice_model type="inv_buf" name="INVTX1" prefix="INVTX1" is_default="1">
<design_technology type="cmos" topology="inverter" size="1" tapered="off"/>
<port type="input" prefix="in" size="1"/>
<port type="output" prefix="out" size="1"/>
</spice_model>
<spice_model type="pass_gate" name="TGATE" prefix="TGATE" is_default="1">
<design_technology type="cmos" topology="transmission_gate" nmos_size="1" pmos_size="2"/>
<input_buffer exist="off"/>
<output_buffer exist="off"/>
<port type="input" prefix="in" size="1"/>
<port type="input" prefix="sel" size="1"/>
<port type="input" prefix="selb" size="1"/>
<port type="output" prefix="out" size="1"/>
</spice_model>
<spice_model type="chan_wire" name="chan_segment" prefix="track_seg" is_default="1">
<design_technology type="cmos"/>
<input_buffer exist="off"/>
<output_buffer exist="off"/>
<port type="input" prefix="in" size="1"/>
<port type="output" prefix="out" size="1"/>
<wire_param model_type="pie" res_val="101" cap_val="22.5e-15" level="1"/> <!-- model_type could be T, res_val and cap_val DON'T CARE -->
</spice_model>
<spice_model type="wire" name="direct_interc" prefix="direct_interc" is_default="1">
<design_technology type="cmos"/>
<input_buffer exist="off"/>
<output_buffer exist="off"/>
<port type="input" prefix="in" size="1"/>
<port type="output" prefix="out" size="1"/>
<wire_param model_type="pie" res_val="0" cap_val="0" level="1"/> <!-- model_type could be T, res_val cap_val should be defined -->
</spice_model>
<spice_model type="mux" name="mux_2level" prefix="mux_2level" is_default="1" dump_structural_verilog="true">
<design_technology type="cmos" structure="multi-level" num_level="2"/>
<input_buffer exist="on" spice_model_name="INVTX1"/>
<output_buffer exist="on" spice_model_name="INVTX1"/>
<!--mux2to1 subckt_name="mux2to1"/-->
<pass_gate_logic spice_model_name="TGATE"/>
<port type="input" prefix="in" size="1"/>
<port type="output" prefix="out" size="1"/>
<port type="sram" prefix="sram" size="1"/>
</spice_model>
<spice_model type="ff" name="static_dff" prefix="dff" spice_netlist="/research/ece/lnis/USERS/tang/tangxifan-eda-tools/branches/vpr7_rram/vpr/SpiceNetlists/ff.sp" verilog_netlist="/research/ece/lnis/USERS/tang/tangxifan-eda-tools/branches/vpr7_rram/vpr/VerilogNetlists/ff.v">
<design_technology type="cmos"/>
<input_buffer exist="on" spice_model_name="INVTX1"/>
<output_buffer exist="on" spice_model_name="INVTX1"/>
<pass_gate_logic spice_model_name="TGATE"/>
<port type="input" prefix="D" size="1"/>
<port type="input" prefix="Set" size="1" is_global="true" default_val="0" is_set="true"/>
<port type="input" prefix="Reset" size="1" is_global="true" default_val="0" is_reset="true"/>
<port type="output" prefix="Q" size="1"/>
<port type="clock" prefix="clk" size="1" is_global="true" default_val="0" />
</spice_model>
<spice_model type="lut" name="lut6" prefix="lut6" dump_structural_verilog="true">
<design_technology type="cmos"/>
<input_buffer exist="on" spice_model_name="INVTX1"/>
<output_buffer exist="on" spice_model_name="INVTX1"/>
<lut_input_buffer exist="on" spice_model_name="buf4"/>
<pass_gate_logic spice_model_name="TGATE"/>
<port type="input" prefix="in" size="6"/>
<port type="output" prefix="out" size="1"/>
<port type="sram" prefix="sram" size="64"/>
</spice_model>
<spice_model type="sram" name="sram6T" prefix="sram" spice_netlist="/research/ece/lnis/USERS/tang/tangxifan-eda-tools/branches/vpr7_rram/vpr/SpiceNetlists/sram.sp" verilog_netlist="/research/ece/lnis/USERS/tang/tangxifan-eda-tools/branches/vpr7_rram/vpr/VerilogNetlists/sram.v" >
<design_technology type="cmos"/>
<input_buffer exist="on" spice_model_name="INVTX1"/>
<output_buffer exist="on" spice_model_name="INVTX1"/>
<pass_gate_logic spice_model_name="TGATE"/>
<port type="input" prefix="in" size="1"/>
<port type="output" prefix="out" size="2"/>
</spice_model>
<spice_model type="sram" name="sram6T_blwl" prefix="sram_blwl" spice_netlist="/research/ece/lnis/USERS/tang/tangxifan-eda-tools/branches/vpr7_rram/vpr/SpiceNetlists/sram.sp" verilog_netlist="/research/ece/lnis/USERS/tang/tangxifan-eda-tools/branches/vpr7_rram/vpr/VerilogNetlists/sram.v">
<design_technology type="cmos"/>
<input_buffer exist="on" spice_model_name="INVTX1"/>
<output_buffer exist="on" spice_model_name="INVTX1"/>
<pass_gate_logic spice_model_name="TGATE"/>
<port type="input" prefix="in" size="1"/>
<port type="output" prefix="out" size="2"/>
<port type="bl" prefix="bl" size="1" default_val="0" inv_spice_model_name="INVTX1"/>
<port type="blb" prefix="blb" size="1" default_val="1" inv_spice_model_name="INVTX1"/>
<port type="wl" prefix="wl" size="1" default_val="0" inv_spice_model_name="INVTX1"/>
</spice_model>
<spice_model type="iopad" name="iopad" prefix="iopad" spice_netlist="/research/ece/lnis/USERS/tang/tangxifan-eda-tools/branches/vpr7_rram/vpr/SpiceNetlists/io.sp" verilog_netlist="/research/ece/lnis/USERS/tang/tangxifan-eda-tools/branches/vpr7_rram/vpr/VerilogNetlists/io.v">
<design_technology type="cmos"/>
<input_buffer exist="on" spice_model_name="INVTX1"/>
<output_buffer exist="on" spice_model_name="INVTX1"/>
<pass_gate_logic spice_model_name="TGATE"/>
<port type="inout" prefix="pad" size="1"/>
<port type="sram" prefix="en" size="1" mode_select="true" spice_model_name="sram6T_blwl" default_val="1"/>
<port type="input" prefix="outpad" size="1"/>
<port type="input" prefix="zin" size="1" is_global="true" default_val="0" />
<port type="output" prefix="inpad" size="1"/>
</spice_model>
</module_spice_models>
</spice_settings>
<device>
<sizing R_minW_nmos="8926" R_minW_pmos="16067" ipin_mux_trans_size="9"/>
<timing C_ipin_cblock="596e-18" T_ipin_cblock="77.93e-12"/>
<area grid_logic_tile_area="0"/>
<sram area="6">
<verilog organization="memory_bank" spice_model_name="sram6T_blwl"/>
<spice organization="standalone" spice_model_name="sram6T" />
</sram>
<chan_width_distr>
<io width="1.000000"/>
<x distr="uniform" peak="1.000000"/>
<y distr="uniform" peak="1.000000"/>
</chan_width_distr>
<switch_block type="wilton" fs="3"/>
</device>
<cblocks>
<switch type="mux" name="cb_mux" R="0" Cin="596e-18" Cout="0" Tdel="77.93e-12" mux_trans_size="3" buf_size="63" spice_model_name="mux_2level_tapbuf" structure="multi-level" num_level="2">
</switch>
</cblocks>
<switchlist>
<switch type="mux" name="sb_mux_L4" R="128" Cin="596e-18" Cout="0e-15" Tdel="47.2e-12" mux_trans_size="3" buf_size="63" spice_model_name="mux_1level_tapbuf" structure="multi-level" num_level="1">
</switch>
</switchlist>
<segmentlist>
<segment freq="0.4" length="4" type="unidir" Rmetal="101" Cmetal="22.5e-15" spice_model_name="chan_segment">
<mux name="sb_mux_L4"/>
<sb type="pattern">1 1 1 1 1</sb>
<cb type="pattern">1 1 1 1</cb>
</segment>
<segment freq="0.3" length="2" type="unidir" Rmetal="101" Cmetal="22.5e-15" spice_model_name="chan_segment">
<mux name="sb_mux_L4"/>
<sb type="pattern">1 1 1</sb>
<cb type="pattern">1 1 </cb>
</segment>
<segment freq="0.3" length="1" type="unidir" Rmetal="101" Cmetal="22.5e-15" spice_model_name="chan_segment">
<mux name="sb_mux_L4"/>
<sb type="pattern">1 1</sb>
<cb type="pattern">1</cb>
</segment>
</segmentlist>
<complexblocklist>
<pb_type name="io" capacity="8" area="0" idle_mode_name="inpad" physical_mode_name="io_phy">
<input name="outpad" num_pins="1"/>
<output name="inpad" num_pins="1"/>
<!-- physical design description -->
<mode name="io_phy" disabled_in_packing="false">
<pb_type name="iopad" blif_model=".subckt io" num_pb="1" spice_model_name="iopad">
<input name="outpad" num_pins="1"/>
<output name="inpad" num_pins="1"/>
</pb_type>
<interconnect>
<direct name="inpad" input="iopad.inpad" output="io.inpad">
<delay_constant max="0e-11" in_port="iopad.inpad" out_port="io.inpad"/>
</direct>
<direct name="outpad" input="io.outpad" output="iopad.outpad">
<delay_constant max="0e-11" in_port="io.outpad" out_port="iopad.outpad"/>
</direct>
</interconnect>
</mode>
<mode name="inpad">
<pb_type name="inpad" blif_model=".input" num_pb="1" spice_model_name="iopad" mode_bits="1">
<output name="inpad" num_pins="1"/>
</pb_type>
<interconnect>
<direct name="inpad" input="inpad.inpad" output="io.inpad">
<delay_constant max="0e-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" spice_model_name="iopad" mode_bits="0">
<input name="outpad" num_pins="1"/>
</pb_type>
<interconnect>
<direct name="outpad" input="io.outpad" output="outpad.outpad">
<delay_constant max="0e-11" in_port="io.outpad" out_port="outpad.outpad"/>
</direct>
</interconnect>
</mode>
<!-- 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 -->
<fc default_in_type="frac" default_in_val="0.15" default_out_type="frac" default_out_val="0.10"/>
<!-- IOs go on the periphery of the FPGA, for consistency,
make it physically equivalent on all sides so that only one definition of I/Os is needed.
If I do not make a physically equivalent definition, then I need to define 4 different I/Os, one for each side of the FPGA
-->
<pinlocations pattern="custom">
<loc side="left">io.outpad io.inpad</loc>
<loc side="top">io.outpad io.inpad</loc>
<loc side="right">io.outpad io.inpad</loc>
<loc side="bottom">io.outpad io.inpad</loc>
</pinlocations>
<!-- Place I/Os on the sides of the FPGA -->
<gridlocations>
<loc type="perimeter" priority="10"/>
</gridlocations>
<power method="ignore"/>
</pb_type>
<!-- Define I/O pads ends -->
<pb_type name="clb" area="53894" opin_to_cb="false">
<pin_equivalence_auto_detect input_ports ="off" output_ports="off"/>
<input name="I" num_pins="20" equivalent="true"/> <!-- Each slice will have 6 inputs but we pass the output of a slice to the next one hence 4x(6-1)= 20 inputs (one input from the first slice is hanging) -->
<output name="O" num_pins="1" equivalent="false"/>
<clock name="clk" num_pins="1"/>
<!-- fle basically refers to a slice. In this example, we just want one of them but we
leave it in so that we know this architecture can and should be used -->
<pb_type name="fle" num_pb="4" idle_mode_name="n1_lut6" physical_mode_name="n1_lut6">
<input name="in" num_pins="6"/>
<output name="out" num_pins="1"/>
<clock name="clk" num_pins="1"/>
<!-- 4-LUT mode definition begin -->
<mode name="n1_lut6">
<!-- Define 6-LUT mode -->
<pb_type name="ble6" num_pb="1">
<input name="in" num_pins="6"/>
<output name="out" num_pins="1"/>
<clock name="clk" num_pins="1"/>
<!-- Define LUT -->
<pb_type name="lut6" blif_model=".names" num_pb="1" class="lut" spice_model_name="lut6">
<input name="in" num_pins="6" port_class="lut_in"/>
<output name="out" num_pins="1" port_class="lut_out"/>
<!-- LUT timing using delay matrix, one delay per input -->
<delay_matrix type="max" in_port="lut6.in" out_port="lut6.out">
127e-12
127e-12
127e-12
127e-12
127e-12
127e-12
</delay_matrix>
</pb_type>
<!-- Define flip-flop -->
<pb_type name="ff" blif_model=".latch" num_pb="1" class="flipflop" spice_model_name="static_dff">
<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="29e-12" port="ff.D" clock="clk"/>
<T_clock_to_Q max="16e-12" port="ff.Q" clock="clk"/>
</pb_type>
<interconnect>
<direct name="direct1" input="ble6.in" output="lut6.in"/>
<direct name="direct2" input="lut6.out" output="ff.D">
<!-- Advanced user option that tells CAD tool to find LUT+FF pairs in netlist-->
<pack_pattern name="ble6" in_port="lut6.out" out_port="ff.D"/>
</direct>
<direct name="direct3" input="ble6.clk" output="ff.clk"/>
<mux name="mux1" input="ff.Q lut6.out" output="ble6.out" spice_model_name="mux_1level_tapbuf">
<delay_constant max="42.06e-12" in_port="lut6.out" out_port="ble6.out" />
<delay_constant max="42.06e-12" in_port="ff.Q" out_port="ble6.out" />
</mux>
</interconnect>
</pb_type>
<interconnect>
<direct name="direct1" input="fle.in" output="ble6.in"/>
<direct name="direct2" input="ble6.out" output="fle.out"/>
<direct name="direct3" input="fle.clk" output="ble6.clk"/>
</interconnect>
</mode>
<!-- 6-LUT mode definition end -->
</pb_type>
<interconnect>
<complete name="crossbar" input="clb.I fle[3:0].out" output="fle[3:0].in" spice_model_name="mux_2level">
<delay_constant max="53.44e-12" in_port="clb.I" out_port="fle[3:0].in" />
<delay_constant max="53.44e-12" in_port="fle[3:0].out" out_port="fle[3:0].in" />
</complete>
<complete name="clks" input="clb.clk" output="fle[3:0].clk">
<complete/>
<direct name="clbouts1" input="fle[3:0].out" output="clb.O[3:0]"/>
</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 -->
<fc default_in_type="frac" default_in_val="0.15" default_out_type="frac" default_out_val="0.10"/>
<pinlocations pattern="spread"/>
<!-- Place this general purpose logic block in any unspecified column -->
<gridlocations>
<loc type="fill" priority="1"/>
</gridlocations>
</pb_type>
<!-- Define general purpose logic block (CLB) ends -->
</complexblocklist>
<power>
<local_interconnect C_wire="0"/>
<mux_transistor_size mux_transistor_size="5"/>
<FF_size FF_size="4"/>
<LUT_transistor_size LUT_transistor_size="5"/>
</power>
<clocks>
<clock buffer_size="auto" C_wire="0"/>
</clocks>
</architecture>

Binary file not shown.

After

Width:  |  Height:  |  Size: 85 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 149 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 130 KiB