SPMI: Switch back to unpinned 9.0 SDK for micro benchmarks #109992
Conversation
dotnet-issue-labeler
bot
added
the
area-CodeGen-coreclr
|
Tagging subscribers to this area: @JulieLeeMSFT, @jakobbotsch |
This comment was marked as outdated.
This comment was marked as outdated.
|
Not really sure why it's failing.. I cannot reproduce the failure locally. |
|
/azp run runtime-coreclr superpmi-collect-test |
|
Azure Pipelines successfully started running 1 pipeline(s). |
| @@ -187,12 +187,13 @@ def build_and_run(coreclr_args, output_mch_name): | |||
| env_copy["NUGET_EXPERIMENTAL_CHAIN_BUILD_RETRY_POLICY"] = "9,2000" | |||
|
|
|||
| # If `dotnet restore` fails, retry. | |||
| tfm = "net9.0" | |||
There was a problem hiding this comment.
Another option is to turn it into superpmi_benchmarks.py.template with markers like {TargetFramework} which some .proj would replace at build time to generate the real superpmi_benchmarks.py somewhere under artificats/obj. linker testing uses that mechanism to avoid updating the files each time the TFM version is bumped:
template:
There was a problem hiding this comment.
That's a good idea, but a bit more work than I'd like to take on here -- I opened #110043 about it
|
/azp run runtime-coreclr superpmi-collect-test |
|
Azure Pipelines successfully started running 1 pipeline(s). |
|
/azp run runtime-coreclr superpmi-collect-test |
|
Azure Pipelines successfully started running 1 pipeline(s). |
|
It looks like the fix here was to set Looks like a small bug when collect does it replay also snuck in with #109481, so I also pushed a fix for that. Hopefully this should be the last superpmi-collect-test run... |
|
/azp run runtime-coreclr superpmi-collect-test |
|
Azure Pipelines successfully started running 1 pipeline(s). |
…9992) Also set PERFLAB_TARGET_FRAMEWORKS and fix superpmi.py when it replays during a collection.
cc @dotnet/jit-contrib