Use Peano for Programming Guide and Examples (#1844)

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

.github/workflows/buildAndTestRyzenAI.yml

@@ -38,7 +38,7 @@ concurrency:
 env:
   DEBIAN_FRONTEND: noninteractive
   XILINXD_LICENSE_FILE: /opt/xilinx/Xilinx.lic
-  VITIS: /opt/Xilinx/Vitis/2023.2
+  VITIS: /opt/ryzen_ai-1.3.0/vitis_aie_essentials
   CMAKE_ARGS: |
             -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \
             -DCMAKE_C_COMPILER_LAUNCHER=ccache \
@@ -97,7 +97,7 @@ jobs:
           mkdir build
           pushd build
 
-          export PATH=/opt/Xilinx/Vitis/2023.2/bin:/opt/Xilinx/Vitis/2023.2/aietools/bin:$PATH
+          export PATH=$VITIS/bin:$VITIS/aietools/bin:$PATH
           cmake .. -G Ninja \
             -DPython3_EXECUTABLE=$(which python) \
             -DLLVM_EXTERNAL_LIT=$(which lit) \
@@ -134,11 +134,11 @@ jobs:
 
           pip cache purge
           source /opt/xilinx/xrt/setup.sh
-          export PATH=/opt/Xilinx/Vitis/2023.2/bin:/opt/Xilinx/Vitis/2023.2/aietools/bin:$PATH
+          export PATH=$VITIS/bin:$VITIS/aietools/bin:$PATH
           source utils/quick_setup.sh
           # quick_setup changes directory to programming_examples, so we need to return to mlir-aie
           cd ..
-          ./utils/build-mlir-aie-from-wheels.sh ./my_install/mlir
+          ./utils/build-mlir-aie-from-wheels.sh ./my_install/mlir build install ./my_install/llvm-aie
 
           # build is created by the build-mlir-aie-from-wheels.sh script
           pushd build

aie_kernels/generic/passThrough.cc

@@ -12,7 +12,6 @@
 #define NOCPP
 
 #include <stdint.h>
-#include <stdio.h>
 #include <stdlib.h>
 
 #include <aie_api/aie.hpp>
@@ -37,39 +36,33 @@ extern "C" {
 #if BIT_WIDTH == 8
 
 void passThroughLine(uint8_t *in, uint8_t *out, int32_t lineWidth) {
-  printf("passThroughLine BIT_WIDTH\n");
   passThrough_aie<uint8_t, 64>(in, out, 1, lineWidth);
 }
 
 void passThroughTile(uint8_t *in, uint8_t *out, int32_t tileHeight,
                      int32_t tileWidth) {
-  printf("passThroughTile BIT_WIDTH\n");
   passThrough_aie<uint8_t, 64>(in, out, tileHeight, tileWidth);
 }
 
 #elif BIT_WIDTH == 16
 
 void passThroughLine(int16_t *in, int16_t *out, int32_t lineWidth) {
-  printf("passThroughLine BIT_WIDTH\n");
   passThrough_aie<int16_t, 32>(in, out, 1, lineWidth);
 }
 
 void passThroughTile(int16_t *in, int16_t *out, int32_t tileHeight,
                      int32_t tileWidth) {
-  printf("passThroughTile BIT_WIDTH\n");
   passThrough_aie<int16_t, 32>(in, out, tileHeight, tileWidth);
 }
 
 #else // 32
 
 void passThroughLine(int32_t *in, int32_t *out, int32_t lineWidth) {
-  printf("passThroughLine BIT_WIDTH\n");
   passThrough_aie<int32_t, 16>(in, out, 1, lineWidth);
 }
 
 void passThroughTile(int32_t *in, int32_t *out, int32_t tileHeight,
                      int32_t tileWidth) {
-  printf("passThroughTile BIT_WIDTH\n");
   passThrough_aie<int32_t, 16>(in, out, tileHeight, tileWidth);
 }
 

programming_examples/basic/dma_transpose/Makefile

@@ -27,6 +27,7 @@ build/aie.mlir: ${srcdir}/aie2.py
 build/final.xclbin: build/aie.mlir
 	mkdir -p ${@D}
 	cd ${@D} && aiecc.py --aie-generate-cdo --no-compile-host --xclbin-name=${@F} \
+    			--no-xchesscc --no-xbridge --peano ${PEANO_INSTALL_DIR} \
 				--aie-generate-npu --npu-insts-name=insts.txt $(<:%=../%)
 
 ${targetname}.exe: ${srcdir}/test.cpp

programming_examples/basic/dma_transpose/run_makefile.lit

@@ -1,7 +1,7 @@
 // (c) Copyright 2024 Advanced Micro Devices, Inc.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
-// REQUIRES: ryzen_ai, chess
+// REQUIRES: ryzen_ai, peano 
 //
 // RUN: make -f %S/Makefile clean
 // RUN: make -f %S/Makefile 

programming_examples/basic/matrix_multiplication/cascade/Makefile

@@ -25,6 +25,8 @@ target_suffix=${M}x${K}x${N}_${m}x${k}x${n}_${n_aie_cols}c
 
 include ${srcdir}/../makefile-common
 
+CHESS=true
+
 build/mm_${m}x${k}x${n}.o: ${kernels_dir}/cascade_mm.cc
 	mkdir -p ${@D}
 	cd ${@D} && xchesscc_wrapper ${CHESSCCWRAP2_FLAGS} -DBIT_WIDTH=8 -DDIM_M=${m} -DDIM_K=${k} -DDIM_N=${n} -c $< -o ${@F}

programming_examples/basic/matrix_multiplication/makefile-common

@@ -81,14 +81,16 @@ aie_py_src?=aie2.py
 runargs?=-v 2 --warmup 1 --iters 1
 aieargs+=-M $M -K $K -N $N --dtype_in ${dtype_in} --dtype_out ${dtype_out}
 
+CHESS?=false
+
 kernels_dir=${srcdir}/../../../../aie_kernels/aie2
 
 .PHONY: all
 all: ${xclbin_target} ${targetname}.exe
 
 build/%.o: ${kernels_dir}/%.cc
 	mkdir -p ${@D}
-	cd ${@D} && xchesscc_wrapper ${CHESSCCWRAP2_FLAGS} -DBIT_WIDTH=8 -c $< -o ${@F}
+	cd ${@D} && ${PEANO_INSTALL_DIR}/bin/clang++ ${PEANOWRAP2_FLAGS} -DBIT_WIDTH=8 -c $< -o ${@F}
 
 ${mlir_target}: ${srcdir}/${aie_py_src}
 	mkdir -p ${@D}
@@ -97,6 +99,9 @@ ${mlir_target}: ${srcdir}/${aie_py_src}
 ${xclbin_target}: ${mlir_target} ${kernels:%=build/%.o}
 	mkdir -p ${@D}
 	cd ${@D} && aiecc.py --aie-generate-cdo --no-compile-host --xclbin-name=${@F} \
+				$(if $(shell [ $(CHESS) != true ] && echo true), \
+					--no-xchesscc --no-xbridge --peano ${PEANO_INSTALL_DIR}, \
+				) \
 				--aie-generate-npu --npu-insts-name=${insts_target:build/%=%} $(<:%=../%)
 
 ${targetname}.exe: ${srcdir}/test.cpp ${srcdir}/../test.cpp ${srcdir}/../common.h

programming_examples/basic/matrix_multiplication/matrix_vector/Makefile

@@ -26,5 +26,5 @@ include ${SELF_DIR}../makefile-common
 
 build/mv_${m}x${k}.o: ${kernels_dir}/mv.cc
 	mkdir -p ${@D}
-	cd ${@D} && xchesscc_wrapper ${CHESSCCWRAP2_FLAGS} -DBIT_WIDTH=8 -DDIM_M=${m} -DDIM_K=${k} -c $< -o ${@F}
+	cd ${@D} && ${PEANO_INSTALL_DIR}/bin/clang++ ${PEANOWRAP2_FLAGS} -DBIT_WIDTH=8 -DDIM_M=${m} -DDIM_K=${k} -c $< -o ${@F}
 

programming_examples/basic/matrix_multiplication/matrix_vector/Makefile.chess

@@ -0,0 +1,32 @@
+##===- Makefile -----------------------------------------------------------===##
+# 
+# This file licensed under the Apache License v2.0 with LLVM Exceptions.
+# See https://llvm.org/LICENSE.txt for license information.
+# SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+# 
+# Copyright (C) 2024, Advanced Micro Devices, Inc.
+# 
+##===----------------------------------------------------------------------===##
+
+subdir=matrix_vector
+targetname=matrixVectorMultiplication
+
+# Currently does not accept reconfiguring size via these variables; must change
+# in source at aie2.py as well as here
+M=288
+K=288
+N=1
+m=32
+k=32
+
+kernels=mv_${m}x${k}
+
+SELF_DIR := $(dir $(lastword $(MAKEFILE_LIST)))
+include ${SELF_DIR}../makefile-common
+
+CHESS=true
+
+build/mv_${m}x${k}.o: ${kernels_dir}/mv.cc
+	mkdir -p ${@D}
+	cd ${@D} && xchesscc_wrapper ${CHESSCCWRAP2_FLAGS} -DBIT_WIDTH=8 -DDIM_M=${m} -DDIM_K=${k} -c $< -o ${@F}
+

programming_examples/basic/matrix_multiplication/matrix_vector/run_makefile.lit

@@ -1,7 +1,7 @@
 // (c) Copyright 2024 Advanced Micro Devices, Inc.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
-// REQUIRES: ryzen_ai, chess
+// REQUIRES: ryzen_ai, peano 
 //
 // RUN: make -f %S/Makefile clean
 // RUN: make -f %S/Makefile 

programming_examples/basic/matrix_multiplication/matrix_vector/run_makefile_chess.lit

@@ -0,0 +1,9 @@
+// (c) Copyright 2024 Advanced Micro Devices, Inc.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+// REQUIRES: ryzen_ai, chess 
+//
+// RUN: make -f %S/Makefile.chess clean
+// RUN: make -f %S/Makefile.chess
+// RUN: %run_on_npu make -f %S/Makefile.chess run | FileCheck %s
+// CHECK: PASS!

programming_examples/basic/matrix_multiplication/single_core/Makefile

@@ -33,5 +33,5 @@ include ${srcdir}/../makefile-common
 
 build/mm_${m}x${k}x${n}.o: ${kernels_dir}/mm.cc
 	mkdir -p ${@D}
-	cd ${@D} && xchesscc_wrapper ${CHESSCCWRAP2_FLAGS} -DBIT_WIDTH=8 -D${dtype_in}_${dtype_out}_ONLY -DDIM_M=${m} -DDIM_K=${k} -DDIM_N=${n} -c $< -o ${@F}
+	cd ${@D} && ${PEANO_INSTALL_DIR}/bin/clang++ ${PEANOWRAP2_FLAGS} -DBIT_WIDTH=8 -D${dtype_in}_${dtype_out}_ONLY -DDIM_M=${m} -DDIM_K=${k} -DDIM_N=${n} -c $< -o ${@F}
 

programming_examples/basic/matrix_multiplication/single_core/Makefile.chess

@@ -0,0 +1,39 @@
+##===- Makefile -----------------------------------------------------------===##
+# 
+# This file licensed under the Apache License v2.0 with LLVM Exceptions.
+# See https://llvm.org/LICENSE.txt for license information.
+# SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+#
+# Copyright (C) 2024, Advanced Micro Devices, Inc.
+# 
+##===----------------------------------------------------------------------===##
+
+srcdir := $(shell dirname $(realpath $(firstword $(MAKEFILE_LIST))))
+subdir=single_core
+targetname=matrixMultiplication
+kernels=mm
+
+M?=256
+K?=256
+N?=256
+m?=64
+k?=64
+n?=32
+
+kernels=mm_${m}x${k}x${n}
+aieargs+=-m $m -k $k -n $n
+target_suffix=${M}x${K}x${N}_${m}x${k}x${n}
+use_alt?=0
+
+ifeq (${use_alt}, 1)
+aie_py_src=aie2_alt.py
+endif
+
+include ${srcdir}/../makefile-common
+
+CHESS=true
+
+build/mm_${m}x${k}x${n}.o: ${kernels_dir}/mm.cc
+	mkdir -p ${@D}
+	cd ${@D} && xchesscc_wrapper ${CHESSCCWRAP2_FLAGS} -DBIT_WIDTH=8 -D${dtype_in}_${dtype_out}_ONLY -DDIM_M=${m} -DDIM_K=${k} -DDIM_N=${n} -c $< -o ${@F}
+

programming_examples/basic/matrix_multiplication/single_core/run_makefile.lit

@@ -1,7 +1,7 @@
 // (c) Copyright 2024 Advanced Micro Devices, Inc.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
-// REQUIRES: ryzen_ai, chess
+// REQUIRES: ryzen_ai, peano 
 //
 // RUN: mkdir -p %S/test_1
 // RUN: cd %S/test_1

programming_examples/basic/matrix_multiplication/single_core/run_makefile_1.lit

@@ -1,7 +1,7 @@
 // (c) Copyright 2024 Advanced Micro Devices, Inc.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
-// REQUIRES: ryzen_ai, chess
+// REQUIRES: ryzen_ai, peano 
 //
 // RUN: mkdir -p %S/test_2
 // RUN: cd %S/test_2

programming_examples/basic/matrix_multiplication/single_core/run_makefile_alt.lit

@@ -1,7 +1,7 @@
 // (c) Copyright 2024 Advanced Micro Devices, Inc.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
-// REQUIRES: ryzen_ai, chess
+// REQUIRES: ryzen_ai, peano 
 //
 // RUN: mkdir -p %S/test_alt
 // RUN: cd %S/test_alt

programming_examples/basic/matrix_multiplication/single_core/run_makefile_chess.lit

@@ -0,0 +1,9 @@
+// (c) Copyright 2024 Advanced Micro Devices, Inc.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+// REQUIRES: ryzen_ai, chess 
+//
+// RUN: make -f %S/Makefile.chess clean
+// RUN: make -f %S/Makefile.chess
+// RUN: %run_on_npu make -f %S/Makefile.chess run | FileCheck %s
+// CHECK: PASS!

programming_examples/basic/matrix_multiplication/single_core/run_makefile_i8.lit

@@ -1,7 +1,7 @@
 // (c) Copyright 2024 Advanced Micro Devices, Inc.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
-// REQUIRES: ryzen_ai, chess
+// REQUIRES: ryzen_ai, peano 
 //
 // RUN: mkdir -p %S/test_i8
 // RUN: cd %S/test_i8

programming_examples/basic/matrix_multiplication/whole_array/Makefile

@@ -29,11 +29,11 @@ include ${srcdir}/../makefile-common
 
 build/mm_b_row_maj_${m}x${k}x${n}.o: ${kernels_dir}/mm.cc
 	mkdir -p ${@D}
-	cd ${@D} && xchesscc_wrapper ${CHESSCCWRAP2_FLAGS} -DBIT_WIDTH=8 -D${dtype_in}_${dtype_out}_ONLY -DDIM_M=${m} -DDIM_K=${k} -DDIM_N=${n} -c $< -o ${@F}
+	cd ${@D} && ${PEANO_INSTALL_DIR}/bin/clang++ ${PEANOWRAP2_FLAGS} -DBIT_WIDTH=8 -D${dtype_in}_${dtype_out}_ONLY -DDIM_M=${m} -DDIM_K=${k} -DDIM_N=${n} -c $< -o ${@F}
 
 build/mm_b_col_maj_${m}x${k}x${n}.o: ${kernels_dir}/mm_b_col_maj.cc
 	mkdir -p ${@D}
-	cd ${@D} && xchesscc_wrapper ${CHESSCCWRAP2_FLAGS} -DBIT_WIDTH=8 -DDIM_M=${m} -DDIM_K=${k} -DDIM_N=${n} -c $< -o ${@F}
+	cd ${@D} && ${PEANO_INSTALL_DIR}/bin/clang++ ${PEANOWRAP2_FLAGS} -DBIT_WIDTH=8 -DDIM_M=${m} -DDIM_K=${k} -DDIM_N=${n} -c $< -o ${@F}
 
 ifneq (${b_col_maj}, 1)
 build/mm_${m}x${k}x${n}.o: build/mm_b_row_maj_${m}x${k}x${n}.o 

programming_examples/basic/matrix_multiplication/whole_array/Makefile.chess

@@ -0,0 +1,48 @@
+##===- Makefile -----------------------------------------------------------===##
+# 
+# This file licensed under the Apache License v2.0 with LLVM Exceptions.
+# See https://llvm.org/LICENSE.txt for license information.
+# SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+#
+# Copyright (C) 2024, Advanced Micro Devices, Inc.
+# 
+##===----------------------------------------------------------------------===##
+srcdir := $(shell dirname $(realpath $(firstword $(MAKEFILE_LIST))))
+subdir=whole_array
+targetname=matrixMultiplication
+
+M?=640
+K?=896
+N?=768
+m?=16
+k?=32
+n?=48
+n_aie_cols?=2
+b_col_maj?=0
+
+kernels=mm_${m}x${k}x${n}
+aieargs+=-m $m -k $k -n $n --n-aie-cols ${n_aie_cols} --b-col-maj ${b_col_maj}
+runargs+=--b_col_maj ${b_col_maj}
+target_suffix=${M}x${K}x${N}_${m}x${k}x${n}_${n_aie_cols}c
+
+include ${srcdir}/../makefile-common
+
+CHESS=true
+
+build/mm_b_row_maj_${m}x${k}x${n}.o: ${kernels_dir}/mm.cc
+	mkdir -p ${@D}
+	cd ${@D} && xchesscc_wrapper ${CHESSCCWRAP2_FLAGS} -DBIT_WIDTH=8 -D${dtype_in}_${dtype_out}_ONLY -DDIM_M=${m} -DDIM_K=${k} -DDIM_N=${n} -c $< -o ${@F}
+
+build/mm_b_col_maj_${m}x${k}x${n}.o: ${kernels_dir}/mm_b_col_maj.cc
+	mkdir -p ${@D}
+	cd ${@D} && xchesscc_wrapper ${CHESSCCWRAP2_FLAGS} -DBIT_WIDTH=8 -DDIM_M=${m} -DDIM_K=${k} -DDIM_N=${n} -c $< -o ${@F}
+
+ifneq (${b_col_maj}, 1)
+build/mm_${m}x${k}x${n}.o: build/mm_b_row_maj_${m}x${k}x${n}.o 
+	mkdir -p ${@D}
+	ar rvs $@ $^
+else
+build/mm_${m}x${k}x${n}.o: build/mm_b_col_maj_${m}x${k}x${n}.o
+	mkdir -p ${@D}
+	ar rvs $@ $^
+endif

programming_examples/basic/matrix_multiplication/whole_array/run_makefile_1_col.lit

@@ -1,7 +1,7 @@
 // (c) Copyright 2024 Advanced Micro Devices, Inc.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
-// REQUIRES: ryzen_ai, chess
+// REQUIRES: ryzen_ai, peano 
 //
 // RUN: mkdir -p %S/test_1_col
 // RUN: cd %S/test_1_col

programming_examples/basic/matrix_multiplication/whole_array/run_makefile_2_col.lit

@@ -1,7 +1,7 @@
 // (c) Copyright 2024 Advanced Micro Devices, Inc.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
-// REQUIRES: ryzen_ai, chess
+// REQUIRES: ryzen_ai, peano 
 //
 // RUN: mkdir -p %S/test_2_col
 // RUN: cd %S/test_2_col

programming_examples/basic/matrix_multiplication/whole_array/run_makefile_4_col.lit

@@ -1,7 +1,7 @@
 // (c) Copyright 2024 Advanced Micro Devices, Inc.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
-// REQUIRES: ryzen_ai, chess
+// REQUIRES: ryzen_ai, peano 
 //
 // RUN: mkdir -p %S/test_4_col
 // RUN: cd %S/test_4_col

programming_examples/basic/matrix_multiplication/whole_array/run_makefile_4_col_i8.lit

@@ -1,7 +1,7 @@
 // (c) Copyright 2024 Advanced Micro Devices, Inc.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
-// REQUIRES: ryzen_ai, chess
+// REQUIRES: ryzen_ai, peano 
 //
 // RUN: mkdir -p %S/test_4_col_i8
 // RUN: cd %S/test_4_col_i8

programming_examples/basic/matrix_multiplication/whole_array/run_makefile_chess.lit

@@ -0,0 +1,9 @@
+// (c) Copyright 2024 Advanced Micro Devices, Inc.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+// REQUIRES: ryzen_ai, chess 
+//
+// RUN: make -f %S/Makefile.chess clean
+// RUN: make -f %S/Makefile.chess
+// RUN: %run_on_npu make -f %S/Makefile.chess run | FileCheck %s
+// CHECK: PASS!