quantize: row-chunks stream to avoid thread starvation - #27830
Merged
Conversation
ngxson
changed the title
Contributor
|
Tested this out and got identical sha256sum and 20-70% speed up depending on the tensor type when using 64 threads! LGTM! |
ServeurpersoCom
approved these changes
Sep 1, 2026
fewtarius
added a commit
to fewtarius/CachyLLama
that referenced
this pull request
Sep 1, 2026
Merge upstream commits preserving all CachyLLama features: - CachyLLama Lightning Indexer shader + DSV4_HC pre/comb/post (kept over upstream cb30059's li_dict variant - c22b79e already removed the dup) - FA dequant-once scratch with host-RAM safety gate (q8_0/q4_0/q4_1/q5_0/q5_1) - concat_transpose shader (delta-net dim-0) - mmid_row_lists prepass + padded_N + use_row_lists (coexist with upstream's n_experts + hoist_row_ids; both push_constant fields kept in struct and mul_mm.comp / mul_mmq.comp dispatch tries row_lists first, then hoisted, then dense) - MoE expert residency tracking + observability (CLI args, fields, debug) - DFlash framework: decoder_laguna gate, embeddings_nextn post-final-layer lookup, dflash_conv/DFlash2 (selector + p_min) - DSpark (Nemotron) attn_sinks - Upstream's spec : fuse the DFlash encoder into the KV cache injection (ggml-org#27310) replaces the HEAD's two-step llama_encode + llama_decode path with a fused single llama_decode. Same encoder (fc + norm) runs inline as part of the decode graph. EAGLE3/MTP encode path (separate code in common_speculative_impl_draft_eagle3) is untouched. - Laguna-S-2.1 model - User isolation (user_id, conv_hash, slot affinity) - SSD-backed KV cache and page manager - Context checkpoint ring buffer with insertion-order eviction - --lazy-mode (upstream rename of CachyLLama's --tensor-read-lazy) - kv_unified_per_slot (upstream, additive) - nvfp4 scales to attention ops (ggml-org#28000) - quantize: row-slab stream to avoid thread starvation (ggml-org#27830) - test-save-load-state across all architectures (ggml-org#27755) - test-archs verbosity fix (ggml-org#28147) Verified: - llama-server builds clean - test-save-load-state passes (38s) after rebuild - DSV4_HC_COMB shaders all pass - Pre-existing failures unchanged (minimax-m3/01 numerical tolerance, test-recurrent-state-rollback)
fewtarius
pushed a commit
to fewtarius/CachyLLama
that referenced
this pull request
Sep 5, 2026
thecodacus
pushed a commit
to thecodacus/llama.cpp
that referenced
this pull request
Sep 7, 2026
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.
Overview
Alternative solution to #27770
Original idea taken from https://github.com/ngxson/quant-studio
"Stream" here means: tensor is split into row-slabs --> each slab is split into row-chunks --> chunks fed into thread
It's not real streaming like what
quant-studiodoes (no overlap between I/O, dequant and quantization), but pretty close, without invasive changes to the existing code baseWork size limit from #27795 is still being respected
Requirements