[mlir][SVE] Add e2e for 1D depthwise WC convolution #85225

banach-space · 2024-03-14T13:38:47Z

Follow-up for #81625

github-actions · 2024-03-14T13:41:14Z

✅ With the latest revision this PR passed the C/C++ code formatter.

llvmbot · 2024-03-16T16:45:20Z

@llvm/pr-subscribers-mlir-linalg
@llvm/pr-subscribers-mlir-sve

@llvm/pr-subscribers-mlir

Author: Andrzej Warzyński (banach-space)

Changes

Follow-up for #81625 (please only review the latest commit)

Full diff: https://github.com/llvm/llvm-project/pull/85225.diff

1 Files Affected:

(added) mlir/test/Integration/Dialect/Linalg/CPU/ArmSVE/1d-depthwise-conv.mlir (+62)

diff --git a/mlir/test/Integration/Dialect/Linalg/CPU/ArmSVE/1d-depthwise-conv.mlir b/mlir/test/Integration/Dialect/Linalg/CPU/ArmSVE/1d-depthwise-conv.mlir
new file mode 100644
index 00000000000000..97c6d1637c0c1d
--- /dev/null
+++ b/mlir/test/Integration/Dialect/Linalg/CPU/ArmSVE/1d-depthwise-conv.mlir
@@ -0,0 +1,62 @@
+// DEFINE: %{compile} =  mlir-opt %s \
+// DEFINE:    -transform-interpreter -test-transform-dialect-erase-schedule \
+// DEFINE:    -one-shot-bufferize="bufferize-function-boundaries" -lower-vector-mask -cse -canonicalize -convert-vector-to-scf -arm-sve-legalize-vector-storage \
+// DEFINE:    -convert-vector-to-llvm="enable-arm-sve" -test-lower-to-llvm -o %t
+// DEFINE: %{entry_point} = conv
+// DEFINE: %{run} = %mcr_aarch64_cmd %t -e %{entry_point} -entry-point-result=void --march=aarch64 --mattr="+sve"\
+// DEFINE:    -shared-libs=%mlir_runner_utils,%mlir_c_runner_utils
+
+// RUN: %{compile} | %{run} | FileCheck %s
+
+func.func @conv() {
+  // Define input/output tensors
+  %input_init = tensor.empty() : tensor<1x8x6xi32>
+  %output_init = tensor.empty() : tensor<1x7x6xi32>
+
+  %five = arith.constant 5 : i32
+  %zero = arith.constant 0 : i32
+  %input = linalg.fill ins(%five : i32) outs(%input_init : tensor<1x8x6xi32>) -> tensor<1x8x6xi32>
+  %output = linalg.fill ins(%zero : i32) outs(%output_init : tensor<1x7x6xi32>) -> tensor<1x7x6xi32>
+
+  // Define the filter tensor
+  %filter = arith.constant dense<[
+    [ 1,  2, 3, 4, 5, 6],
+    [ 11, 12, 13, 14, 15, 16]
+  ]> : tensor<2x6xi32>
+
+  // static sizes -> dynamic sizes
+  %input_dyn = tensor.cast %input_init : tensor<1x8x6xi32> to tensor<1x8x?xi32>
+  %output_dyn = tensor.cast %output : tensor<1x7x6xi32> to tensor<1x7x?xi32>
+  %filter_dyn = tensor.cast %filter : tensor<2x6xi32> to tensor<2x?xi32>
+
+  // Run the convolution
+  %res = linalg.depthwise_conv_1d_nwc_wc
+    {dilations = dense<1> : vector<1xi64>,
+    strides = dense<1> : vector<1xi64>}
+    ins(%input_dyn, %filter_dyn : tensor<1x8x?xi32>, tensor<2x?xi32>)
+    outs(%output_dyn : tensor<1x7x?xi32>) -> tensor<1x7x?xi32>
+
+  // Print the results
+  // CHECK: SVE: START OF TEST OUTPUT
+  vector.print str "SVE: START OF TEST OUTPUT\n"
+
+  // CHECK-NEXT: Unranked Memref base@ = {{.*}} rank = 3 offset = 0 sizes = [1, 7, 6] strides = [42, 6, 1] data =
+  // CHECK-COUNT-7: [60, 70, 80, 90, 100, 110]
+  %xf = tensor.cast %res : tensor<1x7x?xi32> to tensor<*xi32>
+  call @printMemrefI32(%xf) : (tensor<*xi32>) -> ()
+
+  // CHECK-NEXT: SVE: END OF TEST OUTPUT
+  vector.print str "SVE: END OF TEST OUTPUT\n"
+
+  return
+}
+
+module attributes {transform.with_named_sequence} {
+  transform.named_sequence @__transform_main(%arg0: !transform.any_op {transform.readonly}) {
+    %0 = transform.structured.match ops{["linalg.depthwise_conv_1d_nwc_wc"]} in %arg0 : (!transform.any_op) -> !transform.any_op
+    transform.structured.vectorize %0 vector_sizes [1, 7, [8], 2] : !transform.any_op
+    transform.yield
+  }
+}
+
+func.func private @printMemrefI32(%ptr : tensor<*xi32>) attributes { llvm.emit_c_interface }

c-rhodes

LGTM cheers

mlir/test/Integration/Dialect/Linalg/CPU/ArmSVE/1d-depthwise-conv.mlir

Follow-up for llvm#81625

Simplify test

Follow-up for llvm#81625

omjavaid · 2024-03-23T10:37:10Z

Breaks following AArch64 SVE buildbots:
https://lab.llvm.org/buildbot/#/builders/184/builds/11363
https://lab.llvm.org/buildbot/#/builders/176/builds/9331

I am temporarily reverting commit 01b1b0c

This reverts commit 01b1b0c. Breaks following AArch64 SVE buildbots: https://lab.llvm.org/buildbot/#/builders/184/builds/11363 https://lab.llvm.org/buildbot/#/builders/176/builds/9331

banach-space · 2024-03-23T15:51:46Z

Breaks following AArch64 SVE buildbots:
https://lab.llvm.org/buildbot/#/builders/184/builds/11363
https://lab.llvm.org/buildbot/#/builders/176/builds/9331

I am temporarily reverting commit 01b1b0c

Thanks for reverting and apologies for the disruption!

@omjavaid Annoyingly, I'm unable to reproduce this failure and the other buildbot that's supposed to run these tests failed for unrelated reasons:

https://lab.llvm.org/buildbot/#/builders/197/builds/13740

The limited error log suggests miscomputation, but it's impossible to tell what exactly went wrong. Would you be able to run the following and share the output?

/home/tcwg-buildbot/worker/clang-aarch64-sve-vls-2stage/stage2/bin/mlir-opt /home/tcwg-buildbot/worker/clang-aarch64-sve-vls-2stage/llvm/mlir/test/Integration/Dialect/Linalg/CPU/ArmSVE/1d-depthwise-conv.mlir -transform-interpreter -test-transform-dialect-erase-schedule -one-shot-bufferize="bufferize-function-boundaries" -lower-vector-mask -cse -canonicalize -convert-vector-to-scf -arm-sve-legalize-vector-storage -convert-vector-to-llvm="enable-arm-sve" -test-lower-to-llvm

Otherwise I'm rather stuck.

MacDue · 2024-03-23T16:55:59Z

mlir/test/Integration/Dialect/Linalg/CPU/ArmSVE/1d-depthwise-conv.mlir

+// DEFINE: %{entry_point} = conv
+// DEFINE: %{run} = %mcr_aarch64_cmd %t -e %{entry_point} -entry-point-result=void --march=aarch64 --mattr="+sve"\
+// DEFINE:    -shared-libs=%mlir_runner_utils,%mlir_c_runner_utils
+
+// RUN: %{compile} | %{run} | FileCheck %s


Here it's both using input + output files (via -o), but also piping things. This seems to work when %mcr_aarch64_cmd is an emulator, but if you replace it with mlir-cpu-runner (on SVE hardware) the test fails on the first run (but works on the second run).

P.s. -convert-vector-to-scf -arm-sve-legalize-vector-storage -convert-vector-to-llvm="enable-arm-sve" can be replaced with -convert-vector-to-scf=full-unroll.

Thanks, good catch!

Yeah, the RUN should be something like this: // RUN: rm -f %t && %{compile} && %{run} | FileCheck %s instead. However, the current RUN line, when fails, produces:

# .---command stderr------------ # | cannot open input file '/llvm-project/build/release_vls/tools/mlir/test/Integration/Dialect/Linalg/CPU/ArmSVE/Output/1d-depthwise-conv.mlir.tmp': No such file or directory # | could not parse the input IR # `-----------------------------

rather than (from https://lab.llvm.org/buildbot/#/builders/176/builds/9331):

# .---command stderr------------ # | Error: entry point not found # `-----------------------------

A guess: It may depend on the shell/system if the result is an empty file or no file (Error: entry point not found seems to indicate the input was empty).

That could be the case 🤔

There's https://github.com/llvm/llvm-project/blob/main/mlir/test/Integration/Dialect/Linalg/CPU/ArmSVE/matmul.mlir that works fine with this configuration - we probably just missed that it failed first time it was run. Also a guess :)

If I don't hear back from Omair with other suggestion, I'll merge the updated version around Monday and just monitor the bots to see what happens. Thanks Ben!

No worries, sounds good to me :)

In the matmul test there's a lone // RUN: %{compile}, with the run steps separate, which would not have this issue.

Follow-up for #81625 Relands #85225 with a minor update to the RUN line to fix buildbot failures: ```diff -// RUN: %{compile} | %{run} | FileCheck %s +// RUN: rm -f %t && %{compile} && %{run} | FileCheck %s ``` Failing buildbots after landing #85225: * https://lab.llvm.org/buildbot/#/builders/184/builds/11363 * https://lab.llvm.org/buildbot/#/builders/176/builds/9331

banach-space · 2024-03-25T16:07:08Z

Re-landed: 0bdb883

Buildbot seems fine: https://lab.llvm.org/buildbot/#/builders/184/builds/11408

banach-space requested review from dcaballe, nicolasvasilache and hanhanW as code owners March 14, 2024 13:38

banach-space force-pushed the andrzej/scalable_depthwise_conv_e2e branch from 02008d4 to f3ec5e7 Compare March 14, 2024 13:40

banach-space changed the title ~~andrzej/scalable depthwise conv e2e~~ [mlir][SVE] Add e2e for 1D depthwise WC convolution Mar 14, 2024

banach-space requested a review from c-rhodes March 14, 2024 13:41

banach-space mentioned this pull request Mar 14, 2024

[mlir][linalg] Enable masked vectorisation for depthwise convolutions #81625

Merged

banach-space force-pushed the andrzej/scalable_depthwise_conv_e2e branch from f3ec5e7 to ee401a9 Compare March 16, 2024 16:44

llvmbot added mlir:linalg mlir mlir:sve labels Mar 16, 2024

c-rhodes approved these changes Mar 22, 2024

View reviewed changes

mlir/test/Integration/Dialect/Linalg/CPU/ArmSVE/1d-depthwise-conv.mlir Outdated Show resolved Hide resolved

banach-space added 2 commits March 22, 2024 13:14

[mlir][SVE] Add e2e for 1D depthwise WC convolution

a5f7480

Follow-up for llvm#81625

fixup! [BOLT] Add BB index to BAT (llvm#86044)

41de1a7

Simplify test

banach-space force-pushed the andrzej/scalable_depthwise_conv_e2e branch from ee401a9 to 41de1a7 Compare March 22, 2024 14:07

banach-space merged commit 01b1b0c into llvm:main Mar 22, 2024

chencha3 pushed a commit to chencha3/llvm-project that referenced this pull request Mar 23, 2024

[mlir][SVE] Add e2e for 1D depthwise WC convolution (llvm#85225)

7bc56bc

Follow-up for llvm#81625

MacDue reviewed Mar 23, 2024

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[mlir][SVE] Add e2e for 1D depthwise WC convolution #85225

[mlir][SVE] Add e2e for 1D depthwise WC convolution #85225

banach-space commented Mar 14, 2024 •

edited

Loading

github-actions bot commented Mar 14, 2024 •

edited

Loading

llvmbot commented Mar 16, 2024 •

edited

Loading

c-rhodes left a comment

omjavaid commented Mar 23, 2024

banach-space commented Mar 23, 2024

MacDue Mar 23, 2024 •

edited

Loading

MacDue Mar 23, 2024

banach-space Mar 23, 2024

MacDue Mar 23, 2024

banach-space Mar 23, 2024

MacDue Mar 23, 2024

banach-space commented Mar 25, 2024

[mlir][SVE] Add e2e for 1D depthwise WC convolution #85225

[mlir][SVE] Add e2e for 1D depthwise WC convolution #85225

Conversation

banach-space commented Mar 14, 2024 • edited Loading

github-actions bot commented Mar 14, 2024 • edited Loading

llvmbot commented Mar 16, 2024 • edited Loading

c-rhodes left a comment

Choose a reason for hiding this comment

omjavaid commented Mar 23, 2024

banach-space commented Mar 23, 2024

MacDue Mar 23, 2024 • edited Loading

Choose a reason for hiding this comment

MacDue Mar 23, 2024

Choose a reason for hiding this comment

banach-space Mar 23, 2024

Choose a reason for hiding this comment

MacDue Mar 23, 2024

Choose a reason for hiding this comment

banach-space Mar 23, 2024

Choose a reason for hiding this comment

MacDue Mar 23, 2024

Choose a reason for hiding this comment

banach-space commented Mar 25, 2024

banach-space commented Mar 14, 2024 •

edited

Loading

github-actions bot commented Mar 14, 2024 •

edited

Loading

llvmbot commented Mar 16, 2024 •

edited

Loading

MacDue Mar 23, 2024 •

edited

Loading