Add docs/VLLM_PARITY_FINAL.md: the standing, never-re-litigate record of the
exhaustive GB10 (sm_121) vLLM-parity investigation for the Qwen3.6 NVFP4 hybrid
models. Captures the definitive same-session both-engine benchmark (prefill
S_PP, decode/serving per-seq + aggregate, TTFT, PEAK_GB, paged-as-%-of-vLLM for
both the MoE 35B-A3B and dense 27B models), the complete lever map (every
prefill-GEMM, prefill-GDN, decode and serving/engine attempt with its verdict
and key number), the structural floors (LPDDR5x bandwidth, FP4-MMQ optimality,
GDN O(C^2) intra-chunk + serial recurrence, vLLM's HBM-tuned FLA/Marlin), the
shipped bit-exact wins, and the parity verdict: parity is a hardware ceiling on
GB10, not missing optimizations; the path to parity is datacenter Blackwell.
Every number cites its artifact (dgx:~/bench/COMBINED_DEFINITIVE.txt, the
marlin_gate / gdn_p1_ab A/Bs, PREFILL_GEMM_RESULTS, VLLM_PARITY_LEVER_MAP,
DECODE_SERVING_SCOPE, the patch headers); figures not pinned to an artifact are
marked estimated. Add a section-9 summary + link in the backend README.
Assisted-by: Claude:opus-4.8 [Claude Code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
The 0044/0045 patches were exported from the old bf16/hybrid dev tree and no
longer apply on the f32-only series (0026 ssm_bf16_tau is dropped), so the
build broke at `git apply`. Re-sync the vendored series to the now
feature-complete fork branch mudler/llama.cpp:localai-paged, which is the
canonical source (pin 0ed235ea + the paged patch commits in order).
- git rm the dev-tree-based 0044 (GDN M5, bf16-machinery base) and 0045
(Marlin W4A16 offline-repack, not part of the fork branch).
- Add the fork branch's newest commit (2c32ab8b7, "GDN M5 tensor-core
chunked-scan prefill, f32-only re-port") as 0047, generated with a single
git format-patch off that branch. It sequences after 0046 (its parent on
the branch) and recovers the prefill win 0044 encoded (+3.5% S_PP @npp512,
+17.7% @npp2048), bit-exact per-path (test-backend-ops GATED_DELTA_NET
46/46 default and force-M5; greedy md5 default-on == M5-forced == canonical).
- Track patch 0046 (dense-prefill geometry gate), which was on disk but never
committed, so the series is complete in git.
- README: patch-table header 0001-0046 -> 0001-0047, replace the 0044 row with
the f32-only 0047 row, fix the dangling 0044 prose references, note the
bf16 M6/M7/M8 variants are not part of this f32-only series, and add a
maintenance bullet that the series is now generated from the fork branch so
there is no more patch-export drift.
Verified: on a pristine llama.cpp at pin 0ed235ea the full series 0001-0043,
0046, 0047 applies clean in sorted order with the Makefile's exact
`git apply --verbose` method (37/37 OK), and the resulting tree is
byte-identical to the fork branch tip 2c32ab8b7.
Assisted-by: Claude:opus-4.8 [Claude Code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Land the tensor-core forms of the chunked gated-DeltaNet prefill scan (0031)
as a single GDN_TC-selected build and ship the M5 variant (full TC form-T
solve + state-update mma) default-ON when LLAMA_KV_PAGED is set.
The dispatch defaults GDN_TC=5 and GDN_CHUNK_MIN=64 under paged KV (both
env-overridable; OFF/INT_MAX when not paged, so stock/non-paged stays
regression-free). GDN_CHUNK_MIN is the per-call engage threshold and stays > 1
so decode (1 tok/call) keeps the sequential recurrence; 64 was tuned from a
{1,32,64,128,256} sweep (32/64/128 all win on prefill, 256 barely fires because
the MoE-prefill per-call count is < 256, 1 collapses decode S_TG ~25%).
Measured GB10, q36-35b-a3b-nvfp4, LLAMA_KV_PAGED=1 LLAMA_MOE_FORCE_GRAPHS=1,
llama-batched-bench -ngl 99 -fa on -ntg 4 -npl 32:
-npp 512 S_PP 2208.96 -> 2286.5 t/s (+3.5%, mean of 3 interleaved A/B)
-npp 2048 S_PP 2021.5 -> 2379.8 t/s (+17.7%)
Decode S_TG unchanged (~399 vs ~397 t/s, within noise).
Bit-exactness (per-path greedy md5, n=48 --temp 0 --seed 1, paged): default-on
== M5-forced == canonical on the gate prompt - MoE 8cb0ce23, dense 5951a5b4.
test-backend-ops GATED_DELTA_NET 94/94 vs CPU with M5 forced (incl. multi-chunk
up to n_tokens=256). On a long MoE prompt the default (M5 fires at >=64 tokens)
and the sequential path agree word-for-word until one benign greedy token-flip;
dense is byte-identical. The chunked scan is a NEW per-path result (different FP
reduction order), NMSE-validated benign.
CUDA-only, gencode arch=compute_121a,code=sm_121a (GB10 / sm_121a). README
sections 3 (0044 row, 0031 superseded note) and 5 (dev-notes verdict) updated.
Assisted-by: Claude:opus-4.8 [Claude Code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Patch 0041 (LLAMA_PAGED_DECODE_STABLE) was made default-on-when-paged, but a
measured end-to-end A/B proved that is a serving mistake. S3 defers prefill
admission on the period-8 cadence, which delays prompt admission: 2.5x worse
TTFT (60s vs 24s at N=256) and 20-29% lower end-to-end throughput, with no
end-to-end win at any concurrency. Its apparent decode_agg gain was a metric
artifact (faster per-step decode bought by starving prefill).
Flip the s3_enabled default so an unset LLAMA_PAGED_DECODE_STABLE means OFF; the
mechanism stays available as an explicit opt-in (LLAMA_PAGED_DECODE_STABLE=1) for
decode-dominated, low-arrival traffic where TTFT is not a concern. The default now
prefers prompt prefill admission for good TTFT. S1 (patch 0040) keeps shipping
default-on; only S3's default changes.
Re-exports patch 0041 (change folded into its source commit) and updates the
README 0041 row plus the decode-serving narrative to record the A/B finding.
Greedy md5 gate unchanged (single-sequence llama-completion path, not
update_slots): paged MoE 8cb0ce23, dense 5951a5b4.
Assisted-by: Claude:opus-4.8 [Claude Code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
The S1 section-(a) padded/fixed-slot decode shape (the scoped follow-up to push
serving graph reuse from ~72% toward ~100%) was implemented in an isolated
worktree off the committed S1/S3/tail base, built CUDA-only, and benched on GB10.
Verdict: REJECTED. It is bit-exact and provably inert, but it regresses serving
throughput at every concurrency and does not close the vLLM gap.
Implementation (default-off, LLAMA_PAGED_PAD_DECODE): on a pure-decode step
(n_prompt_budgeted == 0) emit a masked-inert dummy decode for every idle slot so
n_tokens / n_seqs / n_seqs_unq / n_outputs and the seq-id set stay constant; a
release()-side guard keeps a finished slot warm under padding. Each dummy is its
own sequence (private recurrent state, per-stream paged attention, logits
discarded), so it cannot perturb a real stream.
Gates: single-seq greedy md5 bit-exact (dense 5951a5b4, paged-MoE 8cb0ce23). The
literal per-stream ON-vs-OFF identity gate is unachievable - concurrent cuBLAS/FA
decode is not bit-reproducible run-to-run even with padding off (OFF-vs-OFF
diverging streams: dense 3/16, MoE 8/16). The achievable inertness gate passed:
ON-vs-OFF per-stream prefix-agreement equals the OFF-vs-OFF noise floor exactly
(MoE 0.940/0.940, dense 0.812/0.812), so the dummy slots leak nothing.
Bench (MoE Qwen3.6-35B-A3B-NVFP4, GB10), burst decode tok/s/seq: n=8 S1+S3 28.16
/ PAD 6.05 / vLLM 44.8; n=128 S1+S3 4.53 / PAD 4.32 / vLLM 6.87. Staggered
aggregate tok/s: baseline (reuse 0%) 757.6, S1+S3 (reuse 72%) 763.3, PAD
(reuse 38%) 558.0.
Why it fails: (1) serving decode here is GPU-compute-bound, not host-rebuild-bound
- baseline reuse 0% ~= S1+S3 reuse 72% on aggregate tok/s, so closing reuse buys
~nothing (the earlier 542->762 host-bound delta did not reproduce); (2) padding
adds dummy-row compute proportional to pad_width - real_load, catastrophic at low
load; (3) in continuous serving padding cannot hold a constant width (perpetual
prefill churn) so reuse drops 72% -> 38%; (4) the completion-driven batch shrink
padding prevents is itself a throughput win in a compute-bound regime. The
residual burst gap is GPU-compute, which a host-side reuse lever cannot close.
Patch series unchanged: this rejected lever is NOT added to patches/paged/.
Assisted-by: Claude:opus-4.8 [Claude Code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
FIX A (patch 0031 compose break): the chunked GDN prefill patch carried
'#include <cuda_bf16.h>' and '#include <type_traits>' as CONTEXT lines, but
those were introduced by the dropped bf16-tau patch 0026, so on the
bf16-tau-free 0001-0030 base only '#include <cstdlib>' is present and 'git
apply' failed. The same 0026 drop also shifted 0031's later hunks off their
context (the ', hyb' kernel-launch arg, the 'STATE_BF16, HYBRID' template
params, and the GDN_LAUNCH_ARGS list). Regenerated 0031 against a fresh
pin(0ed235ea) + 0001-0030 tree: the chunked kernel now SELF-PROVIDES the
cuda_bf16.h / type_traits includes (adds them, plus the climits it needs for
INT_MAX) and the dispatch guard is the 2-param 'if constexpr (!KDA &&
!keep_rs_t)' form. Behaviour is unchanged: 0031 stays opt-in, default OFF
(GDN_CHUNK_MIN), a recorded negative. The full 0001-0042 series now applies
clean on 0ed235ea ('git apply --check' green for every patch).
FIX B (patch 0041 S3 default): the decode-shape-stable scheduler defaulted OFF.
Make it default ON whenever paged KV is active (LLAMA_KV_PAGED set), still
overridable to off via LLAMA_PAGED_DECODE_STABLE=0. Minimal host-side change in
update_slots(); re-exported from the dev tree, README 0041 row updated to match.
Assisted-by: Claude:opus-4.8 [Claude Code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Add the two decode-serving graph-reuse levers (validated on GB10) that close the
host-bound serving gap (paged dropped to ~3.7 vs vLLM ~5.9 tok/s/seq in real
continuous serving while tying it in static batched-bench).
- 0040 S1 paged decode-graph reuse: the paged decode inputs never overrode
llm_graph_input_i::can_reuse (defaults false), so the host rebuilt the ggml
graph on EVERY decode step (layer-A reuse 0%). Add a 256-bucketed-shape
can_reuse + a live-mctx refresh from the owning attn input. Bit-exact (md5
byte-identical reuse on/off). Static batched-bench: paged reuse 0% -> 95.5%.
- 0041 S3 decode-shape-stable scheduling: keep co-batched prefill out of decode
steps so the scheduler emits the reuse-stable pure-decode shape S1 can reuse.
Default-off policy on top of 0016; bit-exact (per-stream independent).
S1+S3 together (128-client staggered serving, MoE Qwen3.6-35B-A3B-NVFP4): graph
reuse 0% -> 72.2%, hostproc 15.98 -> 6.31 ms/step, decode 4.05 -> 5.52 tok/s/seq
median (4.24 -> 5.96 mean, at vLLM's ~5.9). S1 alone is insufficient (13.8%);
S3 is the multiplier. S2 (double-buffer set_inputs) dropped: Phase-0 put
set_inputs at ~0.05 ms/step, so it has nothing to recover. README patch table +
DECODE_SERVING_SCOPE.md updated with results and the padded/fixed-slot follow-up.
Assisted-by: Claude:opus-4.8 [Claude Code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Adds patch 0031 to the paged llama.cpp series: an FLA-style chunked
parallel-scan prefill kernel for gated DeltaNet (the upstream
gated_delta_net.cu "Add chunked kernel for even faster pre-fill" TODO).
Scope: non-KDA scalar gate, f32 state, final-state-only, homogeneous.
Bit-exact-benign (NEW per-path): test-backend-ops GATED_DELTA_NET 91/91 within
the 1e-7 NMSE gate vs the CPU reference (patch adds 8 S_v=128 prefill cases:
exact-multiple / tail / multi-seq / GQA / permuted); numpy prototype confirms
f32 chunked-vs-sequential NMSE ~1e-13.
OPT-IN, default OFF: GB10's 99KB dynamic-smem opt-in forces C=16 (the 128x128
f32 state is 64KB of the all-shared layout), pinning the kernel to 1 block/SM
with serial dk-reductions. Measured ~761 t/s chunked vs ~971 t/s sequential
(~22%% slower) on q36-27b-nvfp4 prefill, so it defaults OFF (enable with
GDN_CHUNK_MIN=<n>); the backend default is regression-free. Beating the
84.7%-of-peak sequential scan needs tensor-core matmuls / register-resident
state with larger chunks (recorded in README section 5).
Assisted-by: Claude:opus-4.8 [Claude Code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
The opt-in hybrid per-head bf16 SSM-state lever (ssm_bf16_tau, patch 0026) is
removed from the llama-cpp-localai-paged patch series. Clean re-measurement after
the decode fusions (0028 recurrent-state gather-fusion + 0029 block-table cache)
landed shows it buys nothing: forcing ALL gated-DeltaNet heads to bf16
(tau=100000, the most aggressive setting) gives flat decode throughput, 780.6 vs
780.0 t/s. The mode engages but adds zero speed because it is subsumed by the
fusions. The earlier "+12%" was measured before the fusions completed. bf16-tau
was a precision trade (not bit-exact, ~91% same-top-p) plus extra bug surface and
extra CUDA template-instantiation compile cost with no offsetting benefit.
Dependency check: no later patch (0028/0029/0030) depends on 0026. 0030's only
mention is a description comment; its code keys off fused_gdn_ar/ch/auto_fgdn,
which originate in 0018/0019/0021 (before 0026). The remaining series (0001-0025,
0028-0030) applies clean with git apply --check against the pin
0ed235ea2c17a19fc8238668653946721ed136fd. The Makefile applies the series by glob
(patches/paged/0*.patch); the resulting gap at 0026 is tolerated (0005/0027 are
already absent).
Removed:
- patches/paged/0026-qwen35-hybrid-perhead-ssm-state.patch
- the dead ssm_bf16_tau / ssm_hybrid_tau option handler in the shared
grpc-server.cpp (it only set LLAMA_SSM_BF16_TAU, now a no-op the library no
longer reads)
- the patched+bf16-tau benchmark columns and llama-patched-bf16tau rows
(README + final_benchmark.csv), the ssm_bf16_tau option text in backend
index.yaml, the gallery NOTE block, and the docs/features/backends.md mention.
The rejected-lever lesson is kept (why it was dropped: subsumed, tau=100000 flat)
in the backend README section 5, the paged-backend agent guide, and the
vLLM-parity methodology, so it is not re-tried.
Assisted-by: Claude:opus-4.8 [Claude Code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Per request, the plots now show all four series: llama.cpp (standard), vLLM,
LocalAI's llama.cpp patches (bit-exact hero), and LocalAI's patches + bf16-tau
(opt-in ceiling, +3% to +17% over the patches, ahead of vLLM at every dense width
and MoE npl>=32). Subtitle flags bf16-tau as opt-in / not bit-exact.
Assisted-by: Claude:opus-4.8 [Claude Code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Rebuild the two committed decode plots from the re-measured CSV and add a combined
overview. Three series per the comparison that matters: llama.cpp (standard) vs
vLLM vs LocalAI's llama.cpp patches; x-over-standard called out at npl128. bf16-tau
stays out of the plot (it remains in the CSV + the README table as the opt-in row).
Assisted-by: Claude:opus-4.8 [Claude Code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Re-run the GB10/DGX-Spark llama-batched-bench matrix (dense q36-27b + MoE
q36-35b-a3b, npl 8/32/64/128, -fa on -ngl 99 -npp 128 -ntg 128) so the CSV and
README section 4 carry a single consistent set of llama numbers with all three
configs:
- stock: separately-built unpatched llama.cpp at this backend's exact pin
9d5d882d (toggling LLAMA_KV_PAGED on the patched binary does NOT reproduce
stock - the SSM decode fusions are compiled in, not env-gated).
- patched: paged binary, LLAMA_KV_PAGED=1 (+LLAMA_MOE_FORCE_GRAPHS=1 for MoE).
- patched+bf16-tau: patched plus --ssm-bf16-tau 64 (opt-in, NOT bit-exact,
~91% same-top-p).
final_benchmark.csv now has stock + patched + bf16-tau + vllm rows for both
models at all four widths (the prior CSV had no stock and no bf16-tau rows).
peak_gb is dropped: the GB10's unified LPDDR5x reports [N/A] to nvidia-smi and
the bench does not print it, so per-run peak could not be captured this session.
Patch series gives up to 2.46x (dense) / 2.26x (MoE) over true-stock; opt-in
bf16-tau adds a further +3% to +17% on top of patched (growing with width).
vLLM column is kept from the prior session (not re-run) and labeled as such.
Assisted-by: Claude:opus-4.8 [Claude Code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
The paged backend's llama.cpp pin was reverted from c299a92c back to
9d5d882d (== stock), so docs/PIN_SYNC_c299a92c.md (a blow-by-blow of the
reverted sync) is dead weight. The pin-sync PROCESS stays documented in
the three live places: the Makefile comment, README section 7 (Pin +
maintenance policy), and .agents/llama-cpp-localai-paged-backend.md.
Delete the doc and repoint every reference to it (Makefile, README,
.agents, canary script + workflow) at README section 7. No functional
paths change: the canary's patches-dir glob (patches/paged/0*.patch)
is untouched.
Assisted-by: Claude:opus-4.8 [Claude Code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
The c299a92c bump diverged 23 commits ahead of the stock llama-cpp pin.
grpc-server.cpp is SHARED with the stock backend and tracks the stock pin;
c299a92c's upstream server-API refactor pulled stream_* helpers into the headers
grpc-server.cpp includes, whose definitions the stock-aligned build does not
compile -> every paged variant failed to LINK (undefined reference to
stream_aware_should_stop / stream_pipe_producer::cleanup /
stream_session_attach_pipe). The bump was greedy-md5 bit-exact, but the bit-exact
gate never exercises the full grpc-server build, so it slipped through.
Revert LLAMA_VERSION to 9d5d882d (== stock pin, where the patches are bit-exact
AND grpc-server links - the original DGX-proven baseline). Document the hard
constraint in the Makefile, README, PIN_SYNC record, and the .agents guide: the
paged pin must track the stock pin, and a pin-sync must pass the full CI
grpc-server build, not only the bit-exact gate.
Assisted-by: Claude:opus-4.8 [Claude Code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
The llama-cpp-localai-paged patches/ dir had accumulated docs, plots, a csv,
dev .cpp harnesses, and a dead FP4-MoE kernel scaffold after an earlier git-mv.
Restore the invariant that patches/ holds only the .patch series.
Moves:
- patches/paged/README.md -> README.md (canonical doc at the backend root)
- patches/paged/{PIN_SYNC_c299a92c,PAGED_BITEXACT_NOTE,LOCALAI_LLAMACPP_BACKEND_PLAN,UPSTREAM_LAYER2_SCOPE}.md,
final_benchmark.csv, qwen36_*.png, paged-burst-bench.cpp, paged-reclaim-unit.cpp -> docs/
- patches/README.md -> docs/PATCH_MAINTENANCE.md (unique patch-regen recipe not in the canonical README)
Deletes:
- patches/BENCHMARKS.md (superseded by README section 4 + the dev-notes section)
- patches/kernel/ (dead FP4-MoE scaffold, never in the 0001-0030 apply glob, zero refs repo-wide)
Repoint every reference to the moved files: README internal links (docs/ + the
.github links drop from 5x ../ to 3x ../), .agents/llama-cpp-localai-paged-backend.md,
.github/scripts/paged-canary-apply.sh, .github/workflows/llama-cpp-paged-canary.yml,
the wrapper Makefile, backend/cpp/llama-cpp/grpc-server.cpp, backend/index.yaml,
docs/content/features/backends.md, gallery/index.yaml.
The build apply glob PAGED_PATCHES_DIR/0*.patch (PAGED_PATCHES_DIR := .../patches/paged)
is unchanged and still resolves to the 28 patches.
Assisted-by: Claude:opus-4.8 [Claude Code]
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
