Replace hipMallocManaged with hipMalloc+hipMemcpy - #326
Open
maarquitos14 wants to merge 1 commit into
Open
Conversation
There was a problem hiding this comment.
Pull request overview
Replaces HIP managed memory with explicit host/device buffers to avoid costly system-scope atomics on unified-memory GPUs.
Changes:
- Allocates device buffers with
hipMalloc. - Initializes host buffers and transfers data with
hipMemcpy. - Copies results back for debug checks and releases host memory.
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
zjin-lcf
self-requested a review
August 27, 2026 16:01
Collaborator
|
Thank you for the pull request. A PR (#327) was submitted to revise the programs, and it will supersede this one. |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
hipMallocManagedcauses pathological performance on unified memory architectures (e.g. MI300X/gfx942) because atomics on managed memory require system-scope (CPU-GPU) coherency. Withrepeat = 10000, the benchmark takes ~700s to finish withhipMallocManaged, while it completes in ~7s withhipMalloc+hipMemcpy.