[wasm] AOT: System.Numerics.Tensors.Tests - aot-instances.dll with exit code 137

[radical]

Build Information

Build: https://dev.azure.com/dnceng-public/public/_build/results?buildId=517640&view=results
Build error leg or test failing: System.Numerics.Tensors.Tests
Pull request:

Error Message

Fill the error message using step by step known issues guidance.

{
  "ErrorMessage": "",
  "ErrorPattern": "aot-instances.dll.* Exec .*System.Numerics.Tensors.Tests.dll",
  "BuildRetry": false,
  "ExcludeConsoleLog": false
}

Report

Build	Definition	Test	Pull Request
521087	dotnet/runtime	normal-System.Numerics.Tensors.Tests.WorkItemExecution	#95980
519928	dotnet/runtime	normal-System.Numerics.Tensors.Tests.WorkItemExecution	#96696
519750	dotnet/runtime	normal-System.Numerics.Tensors.Tests.WorkItemExecution	#96618
519035	dotnet/runtime	normal-System.Numerics.Tensors.Tests.WorkItemExecution	#95980
518510	dotnet/runtime	normal-System.Numerics.Tensors.Tests.WorkItemExecution	#96618

Summary

24-Hour Hit Count	7-Day Hit Count	1-Month Count
1	5	5

Known issue validation

Build: 🔎 https://dev.azure.com/dnceng-public/public/_build/results?buildId=517640
Error message validated: aot-instances.dll.* Exec .*System.Numerics.Tensors.Tests.dll
Result validation: ❌ Known issue did not match with the provided build.
Validation performed at: 1/8/2024 6:11:29 PM UTC

[ghost]

Tagging subscribers to 'arch-wasm': @lewing
See info in area-owners.md if you want to be subscribed.

Issue Details

---

Build Information

Build: https://dev.azure.com/dnceng-public/public/_build/results?buildId=517640&view=results
Build error leg or test failing: System.Numerics.Tensors.Tests
Pull request:

Error Message

Fill the error message using step by step known issues guidance.

{
  "ErrorMessage": "",
  "ErrorPattern": "",
  "BuildRetry": false,
  "ExcludeConsoleLog": false
}

Author:	radical
Assignees:	-
Labels:	arch-wasm, area-System.Numerics.Tensors, blocking-clean-ci, area-Codegen-AOT-mono, Known Build Error
Milestone:	-

[radical]

This started being hit with db90a849 (build). The last passing build was 7957edc2 (build).

The diff should be within 7957edc...db90a84 .

[radical]

cc @tannergooding @vargaz

[tannergooding]

I'd presume that it's https://github.com/dotnet/runtime/tree/d3be0ecf5a384189acd82f6c5b2631c1db423873

But that's just fixing a minor issue from #94555 (7e51126) and isn't doing anything really unique here, it's just taking an already generic method and dispatching to one of two concrete implementations based on the type of T

[vargaz]

This is probably an OOM. It started with 7e51126, which seems to have a caused the aot compiler to generate way more generic instances. Will investigate.

[tannergooding]

Possibly the same issue as #96630 then. I had initially thought the OOM might be due to the new static virtuals + generics methods, but hitting that even after breaking it apart to 8 concrete APIs.

[vargaz]

It looks like the failure is actually caused by the generic classes added to the test suite i.e. TensorPrimitives.Generic.cs.

[vargaz]

Compiling this test suite with <WasmDedup>false</WasmDedup> will probably work around the failure.

[tannergooding]

Which property needs to be set to false?

[vargaz]

<WasmDedup>false</WasmDedup>

[tannergooding]

Do you expect #96630 is the same issue?

That one appears to be getting hit for System.Private.Corelib so it may not be as easy or as desirable to use that property there?

[tannergooding]

Keeping this open so that the underlying issue with LLVM can be tracked/resolved.

[vargaz]

Its tracked here:
#95791

[radical]

this is still being hit for System.Numerics.Tensors.Net8.Tests which needs WasmDedup=false also.
log.

And the failure on System.Numerics.Tensors.Tests is for the test assembly now instead of aot-instances.dll (log:

  [System.Numerics.Tensors.Tests.dll] Exec (with response file contents expanded) in /root/helix/work/workitem/e/publish: MONO_PATH=/root/helix/work/workitem/e/publish:/root/helix/work/correlation/build/microsoft.netcore.app.runtime.browser-wasm/runtimes/browser-wasm/native/:/root/helix/work/correlation/build/microsoft.netcore.app.runtime.browser-wasm/runtimes/browser-wasm/lib/net9.0 MONO_ENV_OPTIONS= /root/helix/work/correlation/build/cross/mono-aot-cross --wasm-exceptions --debug --llvm "--aot=no-opt,static,direct-icalls,deterministic,mattr=simd,dwarfdebug,llvm-path=/root/helix/work/correlation/build/emsdk/upstream/bin/,static,trimming-eligible-methods-outfile=/root/helix/work/workitem/e/wasm_build/obj/wasm/for-build/tokens/System_Numerics_Tensors_Tests_dll_compiled_methods.txt,llvmonly,interp,asmonly,llvm-outfile=/root/helix/work/workitem/e/wasm_build/obj/wasm/for-build/System.Numerics.Tensors.Tests.dll.bc.tmp,temp-path=/root/helix/work/workitem/e/wasm_build/obj/wasm/for-build/System.Numerics.Tensors.Tests.dll.tmp" "System.Numerics.Tensors.Tests.dll"

[radical]

I'll open two separate issues.