Page MenuHomePhabricator
Differential D75965 Diff 249797 mlir/test/Transforms/affine-data-copy.mlir

Changeset View

Standalone View

View Options

mlir/test/Transforms/affine-data-copy.mlir

// RUN: mlir-opt %s -split-input-file -affine-data-copy-generate -affine-data-copy-generate-dma=false -affine-data-copy-generate-fast-mem-space=0 -affine-data-copy-generate-skip-non-unit-stride-loops | FileCheck %s // RUN: mlir-opt %s -split-input-file -affine-data-copy-generate -affine-data-copy-generate-dma=false -affine-data-copy-generate-fast-mem-space=0 -affine-data-copy-generate-skip-non-unit-stride-loops | FileCheck %s
// Small buffer size to trigger fine copies. // Small buffer size to trigger fine copies.
// RUN: mlir-opt %s -affine-data-copy-generate -affine-data-copy-generate-dma=false -affine-data-copy-generate-fast-mem-space=0 -affine-data-copy-generate-fast-mem-capacity=1 | FileCheck --check-prefix=CHECK-SMALL %s // RUN: mlir-opt %s -affine-data-copy-generate -affine-data-copy-generate-dma=false -affine-data-copy-generate-fast-mem-space=0 -affine-data-copy-generate-fast-mem-capacity=1 | FileCheck --check-prefix=CHECK-SMALL %s
// Test affine data copy with a memref filter. We use a test pass that invokes // Test affine data copy with a memref filter. We use a test pass that invokes
// affine data copy utility on the input loop nest. // affine data copy utility on the input loop nest.
// '-test-affine-data-copy-memref-filter' passes the first memref found in an // '-test-affine-data-copy-memref-filter' passes the first memref found in an
// affine.load op in the innermost loop as a filter. // affine.load op in the innermost loop as a filter.
// RUN: mlir-opt %s -split-input-file -test-affine-data-copy='memref-filter=1' | FileCheck %s --check-prefix=FILTER // RUN: mlir-opt %s -split-input-file -test-affine-data-copy='memref-filter' | FileCheck %s --check-prefix=FILTER
// RUN: mlir-opt %s -split-input-file -test-affine-data-copy='for-memref-region' | FileCheck %s --check-prefix=MEMREF_REGION
bondhugulaUnsubmitted
Done
ReplyInline Actions

Do you need the '=1'? It could be confusing unless it's '=true'.

bondhugula: Do you need the '=1'? It could be confusing unless it's '=true'.
// -copy-skip-non-stride-loops forces the copies to be placed right inside the // -copy-skip-non-stride-loops forces the copies to be placed right inside the
// tile space loops, avoiding the sensitivity of copy placement depth to memory // tile space loops, avoiding the sensitivity of copy placement depth to memory
// footprint -- so that one could write a definite test case and not have to // footprint -- so that one could write a definite test case and not have to
// update it each time something related to the cost functions change. // update it each time something related to the cost functions change.
#map0 = affine_map<(d0) -> (d0)> #map0 = affine_map<(d0) -> (d0)>
#map1 = affine_map<(d0) -> (d0 + 128)> #map1 = affine_map<(d0) -> (d0 + 128)>
▲ Show 20 Lines • Show All 116 Lines • ▼ Show 20 Lines
// ----- // -----
// //
// This test case will lead to single element buffers. These are eventually // This test case will lead to single element buffers. These are eventually
// expected to be turned into registers via alloca and mem2reg. // expected to be turned into registers via alloca and mem2reg.
// //
// CHECK-SMALL-LABEL: func @foo // CHECK-SMALL-LABEL: func @foo
// FILTER-LABEL: func @foo // FILTER-LABEL: func @foo
// MEMREF_REGION-LABEL: func @foo
bondhugulaUnsubmitted
Done
ReplyInline Actions

// MEMREF-REGION-LABEL

here to be safe.

bondhugula: // MEMREF-REGION-LABEL here to be safe.
func @foo(%arg0: memref<1024x1024xf32>, %arg1: memref<1024x1024xf32>, %arg2: memref<1024x1024xf32>) -> memref<1024x1024xf32> { func @foo(%arg0: memref<1024x1024xf32>, %arg1: memref<1024x1024xf32>, %arg2: memref<1024x1024xf32>) -> memref<1024x1024xf32> {
affine.for %i = 0 to 1024 { affine.for %i = 0 to 1024 {
affine.for %j = 0 to 1024 { affine.for %j = 0 to 1024 {
affine.for %k = 0 to 1024 { affine.for %k = 0 to 1024 {
%6 = affine.load %arg1[%k, %j] : memref<1024x1024xf32> %6 = affine.load %arg1[%k, %j] : memref<1024x1024xf32>
%7 = affine.load %arg2[%i, %j] : memref<1024x1024xf32> %7 = affine.load %arg2[%i, %j] : memref<1024x1024xf32>
%9 = addf %6, %7 : f32 %9 = addf %6, %7 : f32
affine.store %9, %arg2[%i, %j] : memref<1024x1024xf32> affine.store %9, %arg2[%i, %j] : memref<1024x1024xf32>
▲ Show 20 Lines • Show All 42 Lines • ▼ Show 20 Lines
// FILTER-NOT: alloc() // FILTER-NOT: alloc()
// FILTER: affine.for %{{.*}} = 0 to 1024 { // FILTER: affine.for %{{.*}} = 0 to 1024 {
// FILTER: affine.for %{{.*}} = 0 to 1024 { // FILTER: affine.for %{{.*}} = 0 to 1024 {
// FILTER: affine.for %{{.*}} = 0 to 1024 { // FILTER: affine.for %{{.*}} = 0 to 1024 {
// FILTER-NEXT: affine.for %{{.*}} = 0 to 1024 { // FILTER-NEXT: affine.for %{{.*}} = 0 to 1024 {
// FILTER-NEXT: affine.for %{{.*}} = 0 to 1024 { // FILTER-NEXT: affine.for %{{.*}} = 0 to 1024 {
// FILTER: dealloc %{{.*}} : memref<1024x1024xf32> // FILTER: dealloc %{{.*}} : memref<1024x1024xf32>
// FILTER-NOT: dealloc // FILTER-NOT: dealloc
bondhugulaUnsubmitted
Done
ReplyInline Actions

Perhaps a comment here on what is being checked for.

bondhugula: Perhaps a comment here on what is being checked for.
// CHeck that only one memref is copied, because for-memref-region is enabled
// (and the first ever encountered load is analyzed).
// MEMREF_REGION: alloc() : memref<1024x1024xf32>
// MEMREF_REGION-NOT: alloc()
// MEMREF_REGION: affine.for %{{.*}} = 0 to 1024 {
// MEMREF_REGION: affine.for %{{.*}} = 0 to 1024 {
// MEMREF_REGION: affine.for %{{.*}} = 0 to 1024 {
// MEMREF_REGION-NEXT: affine.for %{{.*}} = 0 to 1024 {
// MEMREF_REGION-NEXT: affine.for %{{.*}} = 0 to 1024 {
// MEMREF_REGION: dealloc %{{.*}} : memref<1024x1024xf32>
// MEMREF_REGION-NOT: dealloc