diff --git a/docs/hrx.md b/docs/hrx.md index 82374ff05..140e8b0d8 100644 --- a/docs/hrx.md +++ b/docs/hrx.md @@ -359,7 +359,12 @@ from a device fault to passing on ordinary data (128 refined candidates, NMSE 6. CPU), with one activation at 1.06e5 (128 of 128 refined logits `inf` with only the first fault fixed; 0 after), and with activations to 1.18e5 and logits to 2.3e5. A scheduling test without a device pins the group count at widths 256, 2560 and 4096. Not run on Flash-Next itself. In about 50 runs one -refined logit came back wrong once and did not recur ([engine#2150](https://github.com/1bit-MONSTER/engine/issues/2150)). +refined logit came back wrong once ([engine#2150](https://github.com/1bit-MONSTER/engine/issues/2150)). That was a race, +fixed at pin `7a96fa179` (llama.cpp #330): the refine kernel hands each exact logit to the workitem that publishes it +through global memory, and the barrier between them fenced shared memory only, so a slot was now and then published +before the store reached it and carried the coarse scan's score instead (1543.49, 64 times the coarse 24.12, where the +exact logit is 29.15). The test `mul_mat.q6_k.token1_shortlist.repeat` recomputes the case 200 times in one process. +Measured on gfx1151 (2026-11-12), 47,988 recomputes on each side: 318 wrong before, 0 after. The routed gate/up kernels followed at pin `2b7960bae` ([engine#2127](https://github.com/1bit-MONSTER/engine/issues/2127), llama.cpp #309). The fused routed gate/up SwiGLU loads its input in four projection kernels (Q4_K, diff --git a/registry/architectures.json b/registry/architectures.json index ee338d5b8..066b48634 100644 --- a/registry/architectures.json +++ b/registry/architectures.json @@ -1,7 +1,7 @@ { "about": "HF architecture -> GGUF architecture and the backends whose code accepts it. Generated by tools/registry_build.py from the pinned sources; do not edit.", "sources": { - "llama.cpp (hrx)": "9dd1a1b72384fbda84c100911296d2a0e96674cb" + "llama.cpp (hrx)": "7a96fa179670c517db679a4b6945c7d0dd7cb3a0" }, "counts": { "hrx": 298, diff --git a/third_party/llama.cpp b/third_party/llama.cpp index 9dd1a1b72..7a96fa179 160000 --- a/third_party/llama.cpp +++ b/third_party/llama.cpp @@ -1 +1 @@ -Subproject commit 9dd1a1b72384fbda84c100911296d2a0e96674cb +Subproject commit 7a96fa179670c517db679a4b6945c7d0dd7cb3a0