-
Notifications
You must be signed in to change notification settings - Fork 173
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
[examples] Add next-mhsa-qkv-fusion example.
- Loading branch information
Showing
2 changed files
with
173 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,106 @@ | ||
| // RUN: buddy-opt %s \ | ||
| // RUN: -pass-pipeline "builtin.module(func.func(tosa-to-linalg-named),func.func(tosa-to-linalg),func.func(tosa-to-tensor),func.func(tosa-to-arith))" \ | ||
| // RUN: | buddy-opt \ | ||
| // RUN: -arith-expand \ | ||
| // RUN: -eliminate-empty-tensors \ | ||
| // RUN: -empty-tensor-to-alloc-tensor \ | ||
| // RUN: -one-shot-bufferize \ | ||
| // RUN: -convert-linalg-to-affine-loops \ | ||
| // RUN: -affine-loop-fusion \ | ||
| // RUN: -lower-affine \ | ||
| // RUN: -func-bufferize \ | ||
| // RUN: -arith-bufferize \ | ||
| // RUN: -tensor-bufferize \ | ||
| // RUN: -buffer-deallocation \ | ||
| // RUN: -finalizing-bufferize \ | ||
| // RUN: -convert-vector-to-scf \ | ||
| // RUN: -expand-strided-metadata \ | ||
| // RUN: -convert-vector-to-llvm \ | ||
| // RUN: -memref-expand \ | ||
| // RUN: -arith-expand \ | ||
| // RUN: -convert-arith-to-llvm \ | ||
| // RUN: -finalize-memref-to-llvm \ | ||
| // RUN: -convert-scf-to-cf \ | ||
| // RUN: -convert-openmp-to-llvm \ | ||
| // RUN: -convert-arith-to-llvm \ | ||
| // RUN: -convert-math-to-llvm \ | ||
| // RUN: -convert-math-to-libm \ | ||
| // RUN: -convert-func-to-llvm \ | ||
| // RUN: -reconcile-unrealized-casts \ | ||
| // RUN: | mlir-cpu-runner -e main -entry-point-result=void \ | ||
| // RUN: -shared-libs=%mlir_runner_utils_dir/libmlir_runner_utils%shlibext \ | ||
| // RUN: -shared-libs=%mlir_runner_utils_dir/libmlir_c_runner_utils%shlibext \ | ||
| // RUN: | FileCheck %s | ||
|
|
||
| func.func private @rtclock() -> f64 | ||
| func.func private @printMemrefF32(%ptr : tensor<*xf32>) | ||
|
|
||
| func.func @kernel(%t0: tensor<1x40x4096xf32>, %t1: tensor<4096x4096xf32>, %t2: tensor<4096x4096xf32>, %t3: tensor<4096x4096xf32>) { | ||
| %t_start = call @rtclock() : () -> f64 | ||
|
|
||
| %42 = tosa.reshape %t0 {new_shape = array<i64: 40, 4096>} : (tensor<1x40x4096xf32>) -> tensor<40x4096xf32> | ||
| %cst_6 = arith.constant dense<0.000000e+00> : tensor<40x4096xf32> | ||
| %43 = linalg.matmul_transpose_b {cast = #linalg.type_fn<cast_signed>} ins(%42, %t1 : tensor<40x4096xf32>, tensor<4096x4096xf32>) outs(%cst_6 : tensor<40x4096xf32>) -> tensor<40x4096xf32> | ||
| %44 = tosa.reshape %43 {new_shape = array<i64: 1, 40, 4096>} : (tensor<40x4096xf32>) -> tensor<1x40x4096xf32> | ||
|
|
||
| %45 = tosa.reshape %t0 {new_shape = array<i64: 40, 4096>} : (tensor<1x40x4096xf32>) -> tensor<40x4096xf32> | ||
| %cst_7 = arith.constant dense<0.000000e+00> : tensor<40x4096xf32> | ||
| %46 = linalg.matmul_transpose_b {cast = #linalg.type_fn<cast_signed>} ins(%45, %t2 : tensor<40x4096xf32>, tensor<4096x4096xf32>) outs(%cst_7 : tensor<40x4096xf32>) -> tensor<40x4096xf32> | ||
| %47 = tosa.reshape %46 {new_shape = array<i64: 1, 40, 4096>} : (tensor<40x4096xf32>) -> tensor<1x40x4096xf32> | ||
|
|
||
| %48 = tosa.reshape %t0 {new_shape = array<i64: 40, 4096>} : (tensor<1x40x4096xf32>) -> tensor<40x4096xf32> | ||
| %cst_8 = arith.constant dense<0.000000e+00> : tensor<40x4096xf32> | ||
| %49 = linalg.matmul_transpose_b {cast = #linalg.type_fn<cast_signed>} ins(%48, %t3 : tensor<40x4096xf32>, tensor<4096x4096xf32>) outs(%cst_8 : tensor<40x4096xf32>) -> tensor<40x4096xf32> | ||
| %50 = tosa.reshape %49 {new_shape = array<i64: 1, 40, 4096>} : (tensor<40x4096xf32>) -> tensor<1x40x4096xf32> | ||
|
|
||
| %t_end = call @rtclock() : () -> f64 | ||
| %time = arith.subf %t_end, %t_start : f64 | ||
|
|
||
| %tensor_unranked_q = tensor.cast %44 : tensor<1x40x4096xf32> to tensor<*xf32> | ||
|
|
||
| // All the elements of the MemRef are the same, | ||
| // only check the first line to verify the correctness. | ||
| // CHECK: Unranked Memref base@ = {{.*}} rank = 3 offset = 0 sizes = [1, 40, 4096] strides = [163840, 4096, 1] data = | ||
| // CHECK-NEXT: [ | ||
| // CHECK-SAME: [ | ||
| // CHECK-SAME: [24576{{(, 24576)*}}], | ||
|
|
||
| %tensor_unranked_k = tensor.cast %47 : tensor<1x40x4096xf32> to tensor<*xf32> | ||
|
|
||
| // All the elements of the MemRef are the same, | ||
| // only check the first line to verify the correctness. | ||
| // CHECK: Unranked Memref base@ = {{.*}} rank = 3 offset = 0 sizes = [1, 40, 4096] strides = [163840, 4096, 1] data = | ||
| // CHECK-NEXT: [ | ||
| // CHECK-SAME: [ | ||
| // CHECK-SAME: [32768{{(, 32768)*}}], | ||
|
|
||
| %tensor_unranked_v = tensor.cast %50 : tensor<1x40x4096xf32> to tensor<*xf32> | ||
|
|
||
| // All the elements of the MemRef are the same, | ||
| // only check the first line to verify the correctness. | ||
| // CHECK: Unranked Memref base@ = {{.*}} rank = 3 offset = 0 sizes = [1, 40, 4096] strides = [163840, 4096, 1] data = | ||
| // CHECK-NEXT: [ | ||
| // CHECK-SAME: [ | ||
| // CHECK-SAME: [40960{{(, 40960)*}}], | ||
|
|
||
| // Print results. | ||
| call @printMemrefF32(%tensor_unranked_q) : (tensor<*xf32>) -> () | ||
| call @printMemrefF32(%tensor_unranked_k) : (tensor<*xf32>) -> () | ||
| call @printMemrefF32(%tensor_unranked_v) : (tensor<*xf32>) -> () | ||
| // Print timings. | ||
| vector.print %time : f64 | ||
|
|
||
| return | ||
| } | ||
|
|
||
| func.func @main() { | ||
|
|
||
| %c0 = arith.constant dense<2.0> : tensor<1x40x4096xf32> | ||
| %c1 = arith.constant dense <3.0> : tensor<4096x4096xf32> | ||
| %c2 = arith.constant dense <4.0> : tensor<4096x4096xf32> | ||
| %c3 = arith.constant dense <5.0> : tensor<4096x4096xf32> | ||
|
|
||
| call @kernel(%c0, %c1, %c2, %c3) : (tensor<1x40x4096xf32>, tensor<4096x4096xf32>, tensor<4096x4096xf32>, tensor<4096x4096xf32>) -> () | ||
|
|
||
| return | ||
| } |