Use pto-arch a3 for Ascend910B handling#234
Use pto-arch a3 for Ascend910B handling#234wangqilin18 wants to merge 14 commits intozhangstevenunity:mainfrom
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 4ff6b7bf16
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| run_sh = run_sh.replace("@EXECUTABLE@", testcase) | ||
| run_sh = run_sh.replace("@RUN_MODE@", run_mode) | ||
| run_sh = run_sh.replace("@SOC_VERSION@", soc_version) | ||
| run_sh = run_sh.replace("@SOC_VERSION@", arch_for_runsh) |
There was a problem hiding this comment.
generate_testcase now fills @SOC_VERSION@ with pto_arch (a3/a5), but the generated run.sh still treats SOC_VERSION as a simulator/target SoC name (it builds simulator library paths and forwards -DSOC_VERSION to CMake in test/npu_validation/templates/run_sh_template.sh). In GOLDEN_MODE=sim, this causes lookups like .../simulator/a3/lib instead of a valid SoC directory, so sim-golden builds/runs can fail for otherwise valid testcases.
Useful? React with 👍 / 👎.
wangqilin18
force-pushed
the
codex/ptoas-release
branch
2 times, most recently
from
f32a77a to
2337737
Compare
wangqilin18
force-pushed
the
codex/ptoas-release
branch
from
ed0da8d to
a8c6755
Compare
2 participants
Summary
Testing