Failure in checked/release asm diffs pipeline

[BruceForstall]

In the coreclr_tests spmi collection, we're seeing a bunch of:

ISSUE: <ASSERT> #303569 D:\a\_work\1\s\src\coreclr\jit\compiler.h (8722) - Assertion failed '(JitConfig.EnableHWIntrinsic() == 0) || (JitConfig.EnableSSE() == 0)' in 'System.Diagnostics.Tracing.RuntimeEventSource:.ctor():this' during 'Importation' (IL size 55; hash 0xd472ed50; FullOpts)

https://dev.azure.com/dnceng-public/public/_build/results?buildId=302887&view=results

This doesn't happen when doing a spmi replay of either Checked or Release compilers alone. Note that checked/release diffs uses the superpmi.py option --diff_with_release, which in turn passes -ignoreStoredConfig to superpmi.exe. Note the comment in superpmi.py:

# If one of the JITs is Release, ignore all the stored configuration variables.
# This isn't necessary if both are Release builds (and, in fact, it can clear variables
# that both Release compilers can interpret). However, we rarely or never compare
# two Release compilers, so this is safest.
flags += [ "-ignoreStoredConfig" ]

This appears to be a regression from #85551.

Maybe will be fixed by #87420?

[ghost]

Tagging subscribers to this area: @JulieLeeMSFT, @jakobbotsch
See info in area-owners.md if you want to be subscribed.

Issue Details

---

In the coreclr_tests spmi collection, we're seeing a bunch of:

ISSUE: <ASSERT> #303569 D:\a\_work\1\s\src\coreclr\jit\compiler.h (8722) - Assertion failed '(JitConfig.EnableHWIntrinsic() == 0) || (JitConfig.EnableSSE() == 0)' in 'System.Diagnostics.Tracing.RuntimeEventSource:.ctor():this' during 'Importation' (IL size 55; hash 0xd472ed50; FullOpts)

https://dev.azure.com/dnceng-public/public/_build/results?buildId=302887&view=results

This doesn't happen when doing a spmi replay of either Checked or Release compilers alone. Note that checked/release diffs uses the superpmi.py option --diff_with_release, which in turn passes -ignoreStoredConfig to superpmi.exe. Note the comment in superpmi.py:

# If one of the JITs is Release, ignore all the stored configuration variables.
# This isn't necessary if both are Release builds (and, in fact, it can clear variables
# that both Release compilers can interpret). However, we rarely or never compare
# two Release compilers, so this is safest.
flags += [ "-ignoreStoredConfig" ]

This appears to be a regression from #85551.

Maybe will be fixed by #87420?

Author:	BruceForstall
Assignees:	tannergooding
Labels:	area-CodeGen-coreclr, blocking-clean-ci-optional
Milestone:	8.0.0

[BruceForstall]

I kicked off https://dev.azure.com/dnceng-public/public/_build/results?buildId=305674&view=results to see if #87420 fixes it

[tannergooding]

#87420 won't fix this as the asserts weren't removed/changed.

This assert should probably just be removed. There are a few cases where Vector<T> can be unsupported even without HWIntrinsics being disabled.

[tannergooding]

Kicked off https://dev.azure.com/dnceng-public/public/_build/results?buildId=307685&view=results