Date: 2026-07-16
Status: method validated; first bundle positive but not promoted
The 0.50 ms/cycle integration threshold now applies to a predeclared,
compatible optimization portfolio, not necessarily to every exact component
in isolation. This is the supported way to evaluate small improvements without
lowering the gate or benchmark-shopping.
The first portfolio combined:
The same-binary B-A-B crossover was positive, but neither bundle run exceeded the qualified 63.349928 tok/s record. Keep the record recipe at M=1 for the QNorm selector and N64 for MXFP4. No LocalMaxxing submission was made.
A bundle may receive a TP4 service load only when all of the following hold:
0.50 ms/cycle;A bundle can establish that sub-gate changes add up, but promotion still requires a new qualified record, exact canaries, and the normal reproducibility and quality gates. One favorable row is not sufficient.
4a6fd874725312c53883b1d53970af1d0eccfc3f;3e600bf5dc8c96ca5a1bc4156c78cf7c10c08b36;48fda4f0e074db005596d6899d5227d3f0316c12;codex/deepseek-v4-m2-portfolio;VLLM_XPU_V4_FUSED_QNORM_ROPE_KV_INSERT_MAX_M=2;VLLM_XPU_MXFP4_SMALL_M_N=128;The clean rebuild produced package-identical artifacts:
_xpu_C.abi3.so: d24308602f8360771aff5c66ffcf7202d763a3b95a56c4cef3bca20833910be6;_moe_C.abi3.so: 490f687d9fae4747d4e2477c9293ec5458e12c8584c2a38c21ba0272acfaed01;libgrouped_gemm_xe_2.so: 9184bcd85b91fc5b92fe4f61c0458123e139c8235742484be5aee3aae95fc1a3.The prior M=2 QNorm gate passed 40/40 changed eager inputs and 8/8 graph
replays on each B70. It projected 1.68-1.90 ms/cycle saved over 43 layers.
The portfolio rebuild repeated the N128 gate on all four cards. Every card passed 48/48 bitwise-exact route cases. Projected savings versus N64 were:
| Physical B70 | Saved ms / 43 layers |
|---|---|
| 0 | 0.3111 |
| 1 | 0.2782 |
| 2 | 0.2961 |
| 3 | 0.2677 |
The measured combined range is therefore approximately 1.95-2.21 ms/cycle.
This clears the portfolio gate even though N128 does not clear it alone.
Raw N128 summaries are in
../../../data/deepseek-v4-reap-m2-portfolio-20260716/.
All three suites used the fixed 12-prompt realistic suite once per prompt, returned token IDs, generated 128 tokens, and reported zero cached prompt tokens.
| Order | Variant | Median tok/s | p10 tok/s | Mean tok/s |
|---|---|---|---|---|
| B1 | QNorm M2 + N128 | 62.446116 | 59.707219 | 62.918715 |
| A | QNorm M1 + N64 | 61.895036 | 58.898587 | 62.193886 |
| B2 | QNorm M2 + N128 | 62.767570 | 59.935732 | 63.137313 |
The two bundle medians average 62.606843 tok/s, +0.711806 tok/s over
the same-binary control. Their p10 average is 59.821475, +0.922888 over
control. The direction is positive, but both bundle medians remain below the
63.349928 record. This is useful evidence for the portfolio method, not a
promotable speed record.
Raw service evidence:
/mnt/fast-ai/bench-results/deepseek-v4-flash-xpu/mtp1-m2-portfolio-qnorm2-n128-screen-20260716TportfolioB1;/mnt/fast-ai/bench-results/deepseek-v4-flash-xpu/mtp1-m2-portfolio-control-qnorm1-n64-20260716TportfolioA;/mnt/fast-ai/bench-results/deepseek-v4-flash-xpu/mtp1-m2-portfolio-qnorm2-n128-confirm-20260716TportfolioB2.After B2, 10/10 ordered exact capture suites passed: 60/60 expected answers, all cached-zero.
Do not discard exact micro-wins solely because each is below 0.50 ms. Keep a
portfolio ledger and combine only compatible candidates whose conservative,
non-overlapping lower bounds clear the threshold. Do not revive candidates
with negative service evidence merely to inflate the projected sum.
The next useful portfolio should add a newly measured exact component to this method, then repeat a frozen same-binary crossover. The present two-component bundle is closed below the record and should not be rerun until another non-overlapping component materially raises its conservative ceiling.