ggml webgpu: faster normal quant and some k-quant matrix operations, better shader parameter handling#20173
Merged
reeselevine merged 6 commits intoggml-org:masterfrom Mar 10, 2026
Merged
Conversation
* Basic JIT compilation for mul_mat, get_rows, and scale (#17) * scale jit working * preliminary working jit for getrows and mulmat, needs refining * simplified mul_mat preprocessing switch statement * get_rows fixes, mul_mat refinement * formatted + last edits * removed some extraneous prints * fixed get_rows, fixed workgroup dispatch in mul_mat. no gibberish * small fix * some changes, working * get_rows and mul_mat jit fixed and working * Update formatting * formatting * Add header --------- Co-authored-by: Neha Abbas <nehaabbas@ReeseLevines-MacBook-Pro.local> Co-authored-by: Reese Levine <reeselevine1@gmail.com> * Start work on all-encompassing shader library * refactor argmax, set_rows * Refactor all but flashattention, mat mul * no gibberish, all k quants added, merged * vec memory fix * q6_k matching metal on my machine, tests passing * Set tile size for q6_k separately * Separate out fast shaders --------- Co-authored-by: neha-ha <137219201+neha-ha@users.noreply.github.com>
…er for parameter buffers, minor cleanups
CISC
approved these changes
Mar 6, 2026
nikhilJain17
approved these changes
Mar 9, 2026
Contributor
nikhilJain17
left a comment
There was a problem hiding this comment.
The logic changes around future submissions and waiting lgtm!
ProgenyAlpha
pushed a commit
to ProgenyAlpha/llama.cpp
that referenced
this pull request
Mar 12, 2026
…better shader parameter handling (ggml-org#20173) * K quant speedup (ggml-org#20) * Basic JIT compilation for mul_mat, get_rows, and scale (ggml-org#17) * scale jit working * preliminary working jit for getrows and mulmat, needs refining * simplified mul_mat preprocessing switch statement * get_rows fixes, mul_mat refinement * formatted + last edits * removed some extraneous prints * fixed get_rows, fixed workgroup dispatch in mul_mat. no gibberish * small fix * some changes, working * get_rows and mul_mat jit fixed and working * Update formatting * formatting * Add header --------- Co-authored-by: Neha Abbas <nehaabbas@ReeseLevines-MacBook-Pro.local> Co-authored-by: Reese Levine <reeselevine1@gmail.com> * Start work on all-encompassing shader library * refactor argmax, set_rows * Refactor all but flashattention, mat mul * no gibberish, all k quants added, merged * vec memory fix * q6_k matching metal on my machine, tests passing * Set tile size for q6_k separately * Separate out fast shaders --------- Co-authored-by: neha-ha <137219201+neha-ha@users.noreply.github.com> * Move towards writeBuffer for params * Move away from multiple buffers for set_rows errors, remove host buffer for parameter buffers, minor cleanups * Remove extra file * Formatting --------- Co-authored-by: neha-ha <137219201+neha-ha@users.noreply.github.com>
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
3 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.
TLDR: The WebGPU implementation is pretty fast now for some quantization types, and hopefully it's more stable/works on a decent number of devices now. Try this code out in wllama here: https://reeselevine.github.io/wllama/. At least on my machine, it mostly seems to outperform WebLLM and OnnxRuntimeWeb (through transformers.js) on roughly equivalent models. And if there are any issues let's try to fix them!
queue.WriteBufferto avoid having an extra buffer to copy parameters from host to device, which removes blocking operations from the encode path.set_rowsand checking it at the end of eachgraph_compute, instead of having separate error buffers for eachset_rowsoperation.