feat(hip_runtime): add device-side allocation + multi-instance tests - #59
Open
y1k2chaitanya wants to merge 1 commit into
Open
feat(hip_runtime): add device-side allocation + multi-instance tests#59y1k2chaitanya wants to merge 1 commit into
y1k2chaitanya wants to merge 1 commit into
Conversation
Contributor
Author
CI validation — e2e-tests.yml on
|
y1k2chaitanya
marked this pull request as ready for review
July 28, 2026 10:52
- test_deviceside_malloc.py: 5 parametrized in-kernel allocation scenarios (malloc/new/per_thread/per_block/across_kernels) via a re-authored public device_side_alloc HIP workload. - test_multi_instance.py: concurrent multi-instance execution via HIP_VISIBLE_DEVICES (same-gpu + two-gpu@gpu_count(2)). New src: device_side_alloc.cpp, hip_multi_instance_app.cpp (gated CMake options). Validated 7/7 on quanta MI325 (gfx942).
y1k2chaitanya
force-pushed
the
users/kyenumul/hip-runtime-deviceside-multiinstance
branch
from
July 30, 2026 20:02
5c180a0 to
d477fc2
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds two functional HIP-runtime suites (shift-left from TMS):
test_deviceside_malloc.py(TMS 1000024HIP_Deviceside_Malloc) — 5 parametrized in-kernel dynamic-allocation scenarios:malloc,new,per_thread,per_block,across_kernels. Each allocates inside the kernel, writes host-verified values, asserts a PASSED sentinel with no allocation fault.test_multi_instance.py(TMS 1002308 / 1002309) — two concurrent instances of a self-verifying HIP vector-add viaHIP_VISIBLE_DEVICES:same_gpu(both on device 0) andtwo_gpu(@gpu_count(2), one per GPU).New vendored public sources under
src/:device_side_alloc.cpp,hip_multi_instance_app.cpp, gated behind dedicated CMake options so existing fixtures are untouched.Provenance / OSS-safety
Re-authored from the standard public HIP device-side-allocation feature and a trivial public vector-add — no internal/NDA payload reused.
Test plan
pytest --collect-only --no-gpu(marker lint) — 7 collectedlinux-gfx94x-MI325: passed — run 30346057930 (12 passed, 288 deselected;test_multi_instance_two_gpuishw.multi_gpuand correctly deselected on the 1-GPU runner, covered by the quanta 2-GPU run)