shylie/llvm-project
1
0
Fork 0
Code Issues Pull Requests Actions 6 Packages Projects Releases Wiki Activity
llvm-project/mlir/test/Integration/Dialect/SparseTensor/CPU/reshape_dot.mlir
Andrzej Warzynski 23e5130ebf [mlir][test] Reland: Refactor SparseTensor CPU integration tests 
CHANGES SINCE THE ORIGINAL VERSION
----------------------------------
The default test set-up was extracted from
  * SparseTensor/CPU/lit.local.cfg.
and duplicated in all tests. This is to support downstream users that
don't use these local LIT config files.

SUMMARY OF CHANGES
------------------
This patch aims to reduce test duplication. This is a direct follow-up of:
  1. https://reviews.llvm.org/D155403 (test duplication), and
  2. https://reviews.llvm.org/D155405 (code re-use),

All SVE/VLA tests are now enabled _conditionally_ and refactored to use
`mlir-cpu-runner` rather than `lli`. The former helps with test
duplication and the latter with code re-use.

A few additional refactoring changes are included.

1. The reduce verbosity, long runtime library names like:

  %mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext

are replaced with:

  %mlir_c_runner_utils

2. In order to keep the code and the comments in sync, and to maintain
   consistency across the tests, the following:

  enable-runtime-library=true

is swapped with (and vice-versa):

  enable-runtime-library=false

Note that this change won't affect test coverage. Only few tests
required such update.

3. A VLS vectorization `RUN` line is added in tests where there was a
   VLA/VLS `RUN` line, but no VLS `RUN` line (with a few exceptions of
   tests that only contained one `RUN` line to begin with).

4. A few test variables are renamed/added. Most notable example:
  * %{options}` --> %{sparse_compiler_opts}

TEST RUNTIME IMPROVEMENT
------------------------
Tl;Dr This change improves test execution time by ~25%.

At the moment, the following `llvm-lit` invocation takes ~7.30s on my
AArch64 workstation (with SVE):

  llvm-lit  <llvm-project>/mlir/test/Integration/Dialect/SparseTensor/CPU/

This timing doesn't change no matter what the value of the following
CMake variable is (that should disable some tests):

  MLIR_RUN_ARM_SVE_TESTS

With this patch, the execution time will indeed depend on the value of
the above CMake variable:
  * with `MLIR_RUN_ARM_SVE_TESTS=true` the timing remains intact,
  * with `MLIR_RUN_ARM_SVE_TESTS=false` the timing drops to ~5.40s (~25%
    improvement).
This is expected:
  * on average there are 4 `RUN` lines per test,
  * _without this change_ (and with `MLIR_RUN_ARM_SVE_TESTS=false`) the
    4th `RUN` line would in most cases duplicate the 3rd `RUN` line,
  * _with this change) (and with `MLIR_RUN_ARM_SVE_TESTS=false`) the
    4th `RUN` line becomes empty.

PATCH SIZE
----------
While rather large and touching many files, most changes in this patch
are rather mechanical. All test configurations have been preserved and
only in a handful of cases new `RUN` lines added.

Differential Revision: https://reviews.llvm.org/D156625
2023-08-11 08:16:01 +00:00

198 lines
9.5 KiB
MLIR
Raw Blame History

//--------------------------------------------------------------------------------------------------
// WHEN CREATING A NEW TEST, PLEASE JUST COPY & PASTE WITHOUT EDITS.
//
// Set-up that's shared across all tests in this directory. In principle, this
// config could be moved to lit.local.cfg. However, there are downstream users that
// do not use these LIT config files. Hence why this is kept inline.
//
// DEFINE: %{sparse_compiler_opts} = enable-runtime-library=true
// DEFINE: %{sparse_compiler_opts_sve} = enable-arm-sve=true %{sparse_compiler_opts}
// DEFINE: %{compile} = mlir-opt %s --sparse-compiler="%{sparse_compiler_opts}"
// DEFINE: %{compile_sve} = mlir-opt %s --sparse-compiler="%{sparse_compiler_opts_sve}"
// DEFINE: %{run_libs} = -shared-libs=%mlir_c_runner_utils,%mlir_runner_utils
// DEFINE: %{run_opts} = -e entry -entry-point-result=void
// DEFINE: %{run} = mlir-cpu-runner %{run_opts} %{run_libs}
// DEFINE: %{run_sve} = %mcr_aarch64_cmd --march=aarch64 --mattr="+sve" %{run_opts} %{run_libs}
//
// DEFINE: %{env} =
//--------------------------------------------------------------------------------------------------
// RUN: %{compile} | %{env} %{run} | FileCheck %s
//
// Do the same run, but now with direct IR generation.
// REDEFINE: %{sparse_compiler_opts} = enable-runtime-library=false
// RUN: %{compile} | %{env} %{run} | FileCheck %s
//
// Do the same run, but now with direct IR generation and vectorization.
// REDEFINE: %{sparse_compiler_opts} = enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true
// RUN: %{compile} | %{env} %{run} | FileCheck %s
//
// Do the same run, but now with direct IR generation and VLA vectorization.
// RUN: %if mlir_arm_sve_tests %{ %{compile_sve} | %{env} %{run_sve} | FileCheck %s %}
#COO_2D = #sparse_tensor.encoding<{ lvlTypes = [ "compressed-nu", "singleton" ], posWidth = 32, crdWidth = 32 }>
#COO_3D = #sparse_tensor.encoding<{ lvlTypes = [ "compressed-nu", "singleton-nu", "singleton" ], posWidth = 32, crdWidth = 32 }>
module {
func.func private @printMemref3dF32(%ptr : tensor<?x?x?xf32>) attributes { llvm.emit_c_interface }
func.func private @printMemref2dF32(%ptr : tensor<?x?xf32>) attributes { llvm.emit_c_interface }
func.func @test_sparse_rhs(%arg0: tensor<5x6xf32>, %arg1: tensor<6x2x3xf32, #COO_3D>) -> tensor<?x?x?xf32> {
%collapsed = tensor.collapse_shape %arg1 [[0], [1, 2]] : tensor<6x2x3xf32, #COO_3D> into tensor<6x6xf32, #COO_2D>
%0 = tensor.empty() : tensor<5x6xf32>
%cst = arith.constant 0.000000e+00 : f32
%1 = linalg.fill ins(%cst : f32) outs(%0 : tensor<5x6xf32>) -> tensor<5x6xf32>
%2 = linalg.matmul ins(%arg0, %collapsed : tensor<5x6xf32>, tensor<6x6xf32, #COO_2D>) outs(%1 : tensor<5x6xf32>) -> tensor<5x6xf32>
%expanded = tensor.expand_shape %2 [[0], [1, 2]] : tensor<5x6xf32> into tensor<5x2x3xf32>
%ret1 = tensor.cast %expanded : tensor<5x2x3xf32> to tensor<?x?x?xf32>
return %ret1 : tensor<?x?x?xf32>
}
func.func @test_sparse_all(%arg0: tensor<5x6xf32, #COO_2D>, %arg1: tensor<6x2x3xf32, #COO_3D>) -> tensor<?x?x?xf32> {
%collapsed = tensor.collapse_shape %arg1 [[0], [1, 2]] : tensor<6x2x3xf32, #COO_3D> into tensor<6x6xf32, #COO_2D>
%0 = tensor.empty() : tensor<5x6xf32>
%cst = arith.constant 0.000000e+00 : f32
%1 = linalg.fill ins(%cst : f32) outs(%0 : tensor<5x6xf32>) -> tensor<5x6xf32>
%2 = linalg.matmul ins(%arg0, %collapsed : tensor<5x6xf32, #COO_2D>, tensor<6x6xf32, #COO_2D>) outs(%1 : tensor<5x6xf32>) -> tensor<5x6xf32>
%expanded = tensor.expand_shape %2 [[0], [1, 2]] : tensor<5x6xf32> into tensor<5x2x3xf32>
%ret1 = tensor.cast %expanded : tensor<5x2x3xf32> to tensor<?x?x?xf32>
return %ret1 : tensor<?x?x?xf32>
}
func.func @test_dense(%arg0: tensor<5x6xf32>, %arg1: tensor<6x2x3xf32>) -> tensor<?x?x?xf32> {
%collapsed = tensor.collapse_shape %arg1 [[0], [1, 2]] : tensor<6x2x3xf32> into tensor<6x6xf32>
%0 = tensor.empty() : tensor<5x6xf32>
%cst = arith.constant 0.000000e+00 : f32
%1 = linalg.fill ins(%cst : f32) outs(%0 : tensor<5x6xf32>) -> tensor<5x6xf32>
%2 = linalg.matmul ins(%arg0, %collapsed : tensor<5x6xf32>, tensor<6x6xf32>) outs(%1 : tensor<5x6xf32>) -> tensor<5x6xf32>
%expanded = tensor.expand_shape %2 [[0], [1, 2]] : tensor<5x6xf32> into tensor<5x2x3xf32>
%ret1 = tensor.cast %expanded : tensor<5x2x3xf32> to tensor<?x?x?xf32>
return %ret1 : tensor<?x?x?xf32>
}
func.func @test_sparse_all_2(%arg0: tensor<5x6xf32, #COO_2D>, %arg1: tensor<2x3x6xf32, #COO_3D>) -> tensor<?x?x?xf32> {
// collapse the first two level this time, as this is the level requires coiterations.
%collapsed = tensor.collapse_shape %arg1 [[0, 1], [2]] : tensor<2x3x6xf32, #COO_3D> into tensor<6x6xf32, #COO_2D>
%0 = tensor.empty() : tensor<5x6xf32>
%cst = arith.constant 0.000000e+00 : f32
%1 = linalg.fill ins(%cst : f32) outs(%0 : tensor<5x6xf32>) -> tensor<5x6xf32>
%2 = linalg.matmul ins(%arg0, %collapsed : tensor<5x6xf32, #COO_2D>, tensor<6x6xf32, #COO_2D>) outs(%1 : tensor<5x6xf32>) -> tensor<5x6xf32>
%expanded = tensor.expand_shape %2 [[0], [1, 2]] : tensor<5x6xf32> into tensor<5x2x3xf32>
%ret1 = tensor.cast %expanded : tensor<5x2x3xf32> to tensor<?x?x?xf32>
return %ret1 : tensor<?x?x?xf32>
}
func.func @entry() {
// Setup two sparse vectors.
%d1 = arith.constant sparse<
[ [0, 0], [1, 1], [2, 2], [2, 3], [4, 5] ],
[1.0, 2.0, 3.0, 4.0, 5.0]
> : tensor<5x6xf32>
%d2 = arith.constant sparse<
[ [0, 0, 0], [1, 1, 1], [2, 1, 1] ],
[ 6.0, 7.0, 8.0]
> : tensor<6x2x3xf32>
%shape = arith.constant dense<[2, 3, 6]> : tensor<3xi32>
%d3 = tensor.reshape %d2(%shape): (tensor<6x2x3xf32>, tensor<3xi32>) -> tensor<2x3x6xf32>
%s1 = sparse_tensor.convert %d1 : tensor<5x6xf32> to tensor<5x6xf32, #COO_2D>
%s2 = sparse_tensor.convert %d2 : tensor<6x2x3xf32> to tensor<6x2x3xf32, #COO_3D>
%s3 = sparse_tensor.convert %d3 : tensor<2x3x6xf32> to tensor<2x3x6xf32, #COO_3D>
// CHECK: Memref base@ = {{.*}} rank = 3 offset = 0 sizes = [5, 2, 3] strides = [6, 3, 1] data =
// CHECK-NEXT:[
// CHECK-SAME: [
// CHECK-SAME: [6, 0, 0],
// CHECK-NEXT: [0, 0, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 14, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 24, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 0, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 0, 0]]]
%do1 = call @test_dense(%d1, %d2) : (tensor<5x6xf32>, tensor<6x2x3xf32>) -> tensor<?x?x?xf32>
call @printMemref3dF32(%do1) : (tensor<?x?x?xf32>) -> ()
// Same results.
// CHECK-NEXT: Memref base@ = {{.*}} rank = 3 offset = 0 sizes = [5, 2, 3] strides = [6, 3, 1] data =
// CHECK-NEXT:[
// CHECK-SAME: [
// CHECK-SAME: [6, 0, 0],
// CHECK-NEXT: [0, 0, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 14, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 24, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 0, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 0, 0]]]
%so1 = call @test_sparse_rhs(%d1, %s2): (tensor<5x6xf32>, tensor<6x2x3xf32, #COO_3D>) -> tensor<?x?x?xf32>
call @printMemref3dF32(%so1) : (tensor<?x?x?xf32>) -> ()
// Same results.
// CHECK-NEXT: Memref base@ = {{.*}} rank = 3 offset = 0 sizes = [5, 2, 3] strides = [6, 3, 1] data =
// CHECK-NEXT:[
// CHECK-SAME: [
// CHECK-SAME: [6, 0, 0],
// CHECK-NEXT: [0, 0, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 14, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 24, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 0, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 0, 0]]]
%so2 = call @test_sparse_all(%s1, %s2): (tensor<5x6xf32, #COO_2D>, tensor<6x2x3xf32, #COO_3D>) -> tensor<?x?x?xf32>
call @printMemref3dF32(%so2) : (tensor<?x?x?xf32>) -> ()
// Same results.
// CHECK-NEXT: Memref base@ = {{.*}} rank = 3 offset = 0 sizes = [5, 2, 3] strides = [6, 3, 1] data =
// CHECK-NEXT:[
// CHECK-SAME: [
// CHECK-SAME: [6, 0, 0],
// CHECK-NEXT: [0, 0, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 14, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 24, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 0, 0]],
// CHECK-NEXT: [
// CHECK-SAME: [0, 0, 0],
// CHECK-NEXT: [0, 0, 0]]]
%so3 = call @test_sparse_all_2(%s1, %s3): (tensor<5x6xf32, #COO_2D>, tensor<2x3x6xf32, #COO_3D>) -> tensor<?x?x?xf32>
call @printMemref3dF32(%so2) : (tensor<?x?x?xf32>) -> ()
bufferization.dealloc_tensor %s1 : tensor<5x6xf32, #COO_2D>
bufferization.dealloc_tensor %s2 : tensor<6x2x3xf32, #COO_3D>
bufferization.dealloc_tensor %s3 : tensor<2x3x6xf32, #COO_3D>
bufferization.dealloc_tensor %do1 : tensor<?x?x?xf32>
bufferization.dealloc_tensor %so1 : tensor<?x?x?xf32>
bufferization.dealloc_tensor %so2 : tensor<?x?x?xf32>
bufferization.dealloc_tensor %so3 : tensor<?x?x?xf32>
return
}
}
Reference in New Issue View Git Blame Copy Permalink