akaashrp opened a new pull request, #20165: URL: https://github.com/apache/tvm/pull/20165
Avoid an intermediate `Uint8Array` copy when uploading aligned data from WASM memory to WebGPU. - Add a bounds-checked borrowed view of Wasm linear memory - Pass that view directly to `GPUQueue.writeBuffer` - Preserve padding for unaligned transfers - Round GPU allocations and readbacks to WebGPU’s four-byte copy alignment - Validate GPU copy offsets and handle zero-length readbacks `writeBuffer` snapshots its source before returning, so the borrowed WASM view does not need to remain valid after the call. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
