Skip to content

TransferBench v1.69#337

Closed
AtlantaPepsi wants to merge 5 commits into
developfrom
candidate
Closed

TransferBench v1.69#337
AtlantaPepsi wants to merge 5 commits into
developfrom
candidate

Conversation

@AtlantaPepsi

Copy link
Copy Markdown
Contributor
  • Fixes for non-zero bytes offset in DMA executor
  • Updated documentation together with ROCm 7.14 release
  • Fixes invalid host access to GPU subExecParam on platforms without large BAR, for GFX executor
  • Fixes a missing device synchronize after input hipMemcpy in RunTransfers()
  • Adds an option to disable validation

gilbertlee-amd and others added 3 commits July 6, 2026 15:36
## Motivation

Adds a performance debug option that disables validation, so only the initialization and actual transfers execute.

Previously, `ALWAYS_VALIDATE` was a boolean (0 = validate once at the end, 1 = validate every iteration), with no way to skip validation entirely. This makes it difficult to isolate pure transfer/initialization cost from validation overhead.

## Technical Details

Adds a "disable" option to `ALWAYS_VALIDATE` while preserving the existing 0/1 behavior (backward compatible):

- `ALWAYS_VALIDATE=-1` — validation **disabled** (init + transfers only)
- `ALWAYS_VALIDATE=0` — validate **once at the end** (default, unchanged)
- `ALWAYS_VALIDATE=1` — validate **after each iteration** (unchanged)

Co-authored-by: Cursor <cursoragent@cursor.com>
Copilot AI review requested due to automatic review settings July 15, 2026 19:13
@AtlantaPepsi
AtlantaPepsi requested review from a team as code owners July 15, 2026 19:13

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This PR bumps TransferBench to v1.69 and adjusts transfer/validation behavior to address byte-offset DMA correctness, improve interactive validation handling, and expose a way to disable validation.

Changes:

  • Fix DMA executor copies to respect non-zero BYTE_OFFSET by applying the offset to src/dst pointers.
  • Extend ALWAYS_VALIDATE semantics to support disabling validation (< 0) and update interactive validation prompts/flows accordingly.
  • Update version strings and changelog for the v1.69 release.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.

Show a summary per file
File Description
src/header/TransferBench.hpp Updates version, fixes DMA byte-offset handling, and adds <0 validation-disable mode with updated interactive validation behavior.
src/client/Utilities.hpp Prints a clear notice when validation is disabled via ALWAYS_VALIDATE < 0.
src/client/EnvVars.hpp Documents and reports the new ALWAYS_VALIDATE modes (<0/0/>0).
src/client/Client.cpp Modifies the version banner formatting for the client output.
CHANGELOG.md Adds a v1.69.00 entry noting the DMA byte-offset fix.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Comment thread src/client/Client.cpp Outdated
Comment thread src/header/TransferBench.hpp

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.

Comment thread src/client/Client.cpp Outdated
Copilot AI review requested due to automatic review settings July 15, 2026 22:27

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

Comment thread CHANGELOG.md
Comment on lines +6 to +9
## v1.69.00
### Added
- Added support for `ALWAYS_VALIDATE=-1` to disable validation
### Fixed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants