… and decode attention kernel (sgl-project#9)

* Add intel_amx backend for attention, including extend and decode

* update

Signed-off-by: Ivan Butygin <ivan.butygin@gmail.com>

…agregation

add llm-datadist feature to realize pd disaggregate

Signed-off-by: Ivan Butygin <ivan.butygin@gmail.com>

Signed-off-by: Ivan Butygin <ivan.butygin@gmail.com>

Merge in 0 4 9 post2

…unked prefill size for OOM scenarios (sgl-project#9)

[Fix] fix fuse share expert function in MI35X

* test: add EPD disaggregation integration tests

* fix comment for encoder-only

* revert http_server warmup for vlm

Devstral tools

nit: modify weight loader warning msg

…s 28 xfers

v9 pool-binding-shift trace produces real differentiation:
- Phase B (KV-bound 8K random): L1+L2 -37% mean TTFT vs stock
- Phase C (mixed 4K random):     L1+L2 -38% median E2E vs stock
- Cross-pool transfers: stock 0, L1-only 0, L2-only 0, L1+L2 28

Two surprising findings documented:
1. Layer 2 alone fires zero transfers — Layer 1 retention is what
   makes Layer 2 cross the firing threshold.
2. Phase A regresses with L1 (-20% TPS) because K_big=8192 hurts on
   prefix-friendly GSP. Consistent with A2's K_big=0-wins finding.
   Adaptive K_big control marked as follow-up.

Settings status: Setting 1 marked **DONE v6 NULL + v9 PASS**.
All 4 user-requested follow-ups (sgl-project#9 Q3.A 4-arm, sgl-project#10 Sweep 1
multi-seed, sgl-project#11 Setting 4 fallback rule, sgl-project#12 Setting 1 v9 trace)
now complete.