Merge branch 'main' into AXI-INFER

.github/actions/perf-check/action.yml

@@ -68,9 +68,11 @@ runs:
           if [[ ! -z "${{ inputs.score }}" ]]; then
             score="-s ${{ inputs.score }}"
           fi
-          worse_count="0"
-          python ./etc/scripts/perf_diff.py -d ${{inputs.datapoints}} $score -o $COMPARE_CSV --returnfail "main-reports/${{inputs.test-id}}.csv" "${{inputs.new-results}}" || worse_count=$?
-          if [[ "$worse_count" != "0" ]]; then
+          
+          worse_count=$(python ./etc/scripts/perf_diff.py -d ${{inputs.datapoints}} $score -o $COMPARE_CSV "main-reports/${{inputs.test-id}}.csv" "${{inputs.new-results}}" 3>&2 2>&1 1>&3)
+          if [[ "$worse_count" == "-1" ]]; then
+            echo "::warning:: Benchmark suite is different from main"
+          elif [[ "$worse_count" != "0" ]]; then
             outcsv="$COMPARE_CSV"
             echo "::warning:: Performance are worse than main for ${{inputs.test-id}}: $worse_count benchmarks have degraded performance"
           fi

.github/workflows/reusable-examples.yml

@@ -41,25 +41,25 @@ jobs:
         include:
           - basename: softfloat
             script: ./examples/softfloat/generic_softfloat.sh
-            batch: "5CSEMA5F31C6_6.6,5CSEMA5F31C6_10.0,5SGXEA7N2F45C1_3.3,5SGXEA7N2F45C1_10.0,EP2C70F896C6-R_10.0,5CSEMA5F31C6_6.6 --flopoco,5CSEMA5F31C6_10.0 --flopoco,5SGXEA7N2F45C1_3.3 --flopoco,5SGXEA7N2F45C1_10.0 --flopoco,EP2C70F896C6-R_10.0 --flopoco"
+            batch: "5CSEMA5F31C6_6.6,5CSEMA5F31C6_10.0,5SGXEA7N2F45C1_3.3,5SGXEA7N2F45C1_10.0,5CSEMA5F31C6_6.6 --flopoco,5CSEMA5F31C6_10.0 --flopoco,5SGXEA7N2F45C1_3.3 --flopoco,5SGXEA7N2F45C1_10.0 --flopoco"
           - basename: truefloat
             script: ./examples/truefloat/generic_truefloat.sh
             args: -c=--evaluation=CYCLES
-            batch: "5CSEMA5F31C6_6.6,5CSEMA5F31C6_10.0,5SGXEA7N2F45C1_3.3,5SGXEA7N2F45C1_10.0,EP2C70F896C6-R_10.0"
+            batch: "5CSEMA5F31C6_6.6,5CSEMA5F31C6_10.0,5SGXEA7N2F45C1_3.3,5SGXEA7N2F45C1_10.0"
           - basename: libm
             script: ./examples/libm/generic_libm.sh
-            batch: "5SGXEA7N2F45C1_3.3,5SGXEA7N2F45C1_5.0,5SGXEA7N2F45C1_10.0,EP2C70F896C6-R_10.0"
+            batch: "5SGXEA7N2F45C1_3.3,5SGXEA7N2F45C1_5.0,5SGXEA7N2F45C1_10.0"
           - basename: libm-flopoco
             script: ./examples/libm/generic_libm.sh
-            batch: "5SGXEA7N2F45C1_3.3 --flopoco,5SGXEA7N2F45C1_5.0 --flopoco,5SGXEA7N2F45C1_10.0 --flopoco,EP2C70F896C6-R_10.0 --flopoco"
+            batch: "5SGXEA7N2F45C1_3.3 --flopoco,5SGXEA7N2F45C1_5.0 --flopoco,5SGXEA7N2F45C1_10.0 --flopoco"
           - basename: CHStone_O2
             script: ./examples/CHStone/generic_CHStone.sh
             args: -c=--experimental-setup=BAMBU-BALANCED-MP -lchstone_aa_list
-            batch: "5CSEMA5F31C6_10.0,5CSEMA5F31C6_15.0,5SGXEA7N2F45C1_5.0,5SGXEA7N2F45C1_10.0,EP2C70F896C6-R_10.0,EP2C70F896C6-R_15.0"
+            batch: "5CSEMA5F31C6_10.0,5CSEMA5F31C6_15.0,5SGXEA7N2F45C1_5.0,5SGXEA7N2F45C1_10.0"
           - basename: CHStone_O3
             script: ./examples/CHStone/generic_CHStone.sh
             args: -c=--experimental-setup=BAMBU-PERFORMANCE-MP -lchstone_altera_O3_list
-            batch: "5CSEMA5F31C6_10.0,5CSEMA5F31C6_15.0,5SGXEA7N2F45C1_5.0,5SGXEA7N2F45C1_10.0,EP2C70F896C6-R_10.0,EP2C70F896C6-R_15.0"
+            batch: "5CSEMA5F31C6_10.0,5CSEMA5F31C6_15.0,5SGXEA7N2F45C1_5.0,5SGXEA7N2F45C1_10.0"
           - basename: hls_study
             script: ./examples/hls_study/generic_hls_study.sh
             batch: "5SGXEA7N2F45C1_2.5 --aligned-access"

documentation/tutorial_fpl_2022/01-introduction/Exercise1/bambu.sh

@@ -0,0 +1,11 @@
+#!/bin/bash
+script=$(readlink -e $0)
+root_dir=$(dirname $script)
+
+rm -rf  icrc1
+mkdir -p icrc1
+cd icrc1
+echo "#synthesis of icrc1"
+bambu ../icrc.c --top-fname=icrc1 \
+   --generate-tb=../test_icrc1.xml --simulator=VERILATOR --simulate \
+   -v2 --print-dot --pretty-print=a.c "$@" |& tee icrc1.log

documentation/tutorial_fpl_2022/01-introduction/Exercise1/icrc.c

@@ -0,0 +1,14 @@
+unsigned short icrc1(unsigned short crc, unsigned char onech)
+{
+    int i;
+    unsigned short ans=(crc^onech << 8);
+
+    for (i=0;i<8;i++) {
+        if (ans & 0x8000)
+            ans = (ans <<= 1) ^ 4129;
+        else
+            ans <<= 1;
+    }
+    return ans;
+}
+

documentation/tutorial_fpl_2022/01-introduction/Exercise2/solution/minmax.c

@@ -0,0 +1,19 @@
+void min_max(int * input, int num_elements, int * max, int * min)
+{
+   int local_max = input[0];
+   int local_min = input[0];
+   int i = 0;
+   for(i = 0; i < num_elements; i++)
+   {
+      if(input[i] > local_max)
+      {
+         local_max = input[i];
+      }
+      else if(input[i] < local_min)
+      {
+         local_min = input[i];
+      }
+   }
+   *min = local_min;
+   *max = local_max;
+}

documentation/tutorial_fpl_2022/01-introduction/Exercise2/solution/synthesize.sh

@@ -0,0 +1,2 @@
+#!/bin/bash
+bambu minmax.c --generate-tb=testbench.xml --simulate "$@" |& tee log.txt

documentation/tutorial_fpl_2022/01-introduction/Exercise2/solution/testbench.xml

@@ -0,0 +1,7 @@
+<?xml version="1.0"?>
+<function>
+   <testbench input="{0,1,2,3,4}" num_elements="5" max="{0}" min="{0}"/>
+   <testbench input="{0,1,2,3,4,5,6,7,8,9}" num_elements="10" max="{0}" min="{0}"/>
+   <testbench input="{0,0,0,0,0,0,0,0,0,0}" num_elements="10" max="{0}" min="{0}"/>
+   <testbench input="{0}" num_elements="1" max="{0}" min="{0}"/>
+</function>

documentation/tutorial_fpl_2022/01-introduction/Exercise3/bambu.sh

@@ -0,0 +1,2 @@
+#!/bin/bash
+bambu matmul.ll --top-fname=main_kernel --generate-tb=test.xml --simulate --simulator=VERILATOR --compiler=I386_CLANG12 "$@" |& tee log.txt