tzhouamandClaude Opus 5
9264ff81ab
The packaging commit shipped the code and a summary of the numbers. It left out
everything the numbers are measured *against*, which is most of what makes them
checkable.
Added:
baselines/ llama.cpp runs (thread sweeps, NUMA sweeps, the
interleaved A/B passes behind the headline decode
figure) and the quality references: each build
against its own bf16 over 12 greedy prompts.
benchmarks/.../results/ 102 artifacts backing the published figures, each
with a provenance block naming the model,
quantization, thread set and build SHAs.
docs/engines/ the engine analyses -- llama.cpp, ExecuTorch, MNN,
ncnn, vLLM, vLLM-Omni -- and the cross-engine and
device-simulation comparisons.
docs/analysis/ the full study documents the summaries distil.
docs/edge/comparisons.md where this fork sits against llama.cpp and why:
decode is llama.cpp's, prefill is ours by ~2x, and
since neither engine exceeds 39% of the measured
222 GB/s ceiling the decode gap is the execution
model rather than the kernels.
docs/edge/environment.md pins and machine, and the correction below.
On "the modified vllm": there isn't one, and the doc says so plainly because it
is the obvious assumption. The vLLM checkout is clean upstream (0 commits
ahead, 0 uncommitted) and the installed wheel is stock 0.28.0+cpu. This fork
changes vLLM's CPU *behaviour* through vLLM's own extension points at import
time -- three CPU patches in vllm_omni/patch.py, the quantization-config
registry, the IR op registry, platform hooks -- specifically so the changes
survive a wheel upgrade rather than being lost on it. One of those patches is
why "KV sizing is not a memory lever" looked true for months: CpuPlatform
overwrote kv_cache_memory_bytes unconditionally, so the setting had no effect.
There is no patch series to apply; pin the version and let patch.py do the rest.
.gitignore gains explicit exceptions: upstream's blanket `*.json` and `results/`
rules were silently eating the measurement artifacts, which in this fork are the
substance rather than build output. Binary captures stay ignored.
No vLLM or llama.cpp source is redistributed -- only output produced with them.
NOTICE records both.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01L1ax5wYLA6XbqEL1i8L9gZ
62 lines
1019 B
JSON
62 lines
1019 B
JSON
{
|
|
"prompt": "<\uff5cstart\u2581of\u2581sentence\uff5c><|System|>\nyou are a helpful assistant.<\uff5cend\u2581of\u2581sentence\uff5c><\uff5cstart\u2581of\u2581sentence\uff5c><|User|>What is the capital of France? Answer in one word.<\uff5cend\u2581of\u2581sentence\uff5c><\uff5cstart\u2581of\u2581sentence\uff5c><|Bot|><think>",
|
|
"prompt_ids": [
|
|
0,
|
|
130972,
|
|
198,
|
|
18178,
|
|
599,
|
|
259,
|
|
18639,
|
|
31627,
|
|
27,
|
|
1,
|
|
0,
|
|
130973,
|
|
4766,
|
|
390,
|
|
275,
|
|
11862,
|
|
314,
|
|
17349,
|
|
44,
|
|
74288,
|
|
304,
|
|
1135,
|
|
5240,
|
|
27,
|
|
1,
|
|
0,
|
|
130976,
|
|
3
|
|
],
|
|
"new_ids": [
|
|
3066,
|
|
599,
|
|
5509,
|
|
39,
|
|
518,
|
|
4766,
|
|
390,
|
|
275,
|
|
11862,
|
|
314,
|
|
17349,
|
|
44,
|
|
74288,
|
|
304,
|
|
1135,
|
|
5240,
|
|
1848,
|
|
4358,
|
|
275,
|
|
6787,
|
|
2029,
|
|
417,
|
|
275,
|
|
5133
|
|
],
|
|
"text": "We are asked: \"What is the capital of France? Answer in one word.\" So the answer should be the single",
|
|
"dtype": "bfloat16",
|
|
"top1": 3066
|
|
} |