Skip to content

[Docs][AMD] Update GLM-5.2 MI355X daily image #41597

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
HaiShaw merged 1 commit into from
Sep 28, 2026
Merged

[Docs][AMD] Update GLM-5.2 MI355X daily image #41597

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion docs/cookbook/autoregressive/GLM/GLM-5.2.mdx
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -124,7 +124,7 @@ import { Playground } from "/src/snippets/_playground.jsx";
**gfx950 block-FP8 accuracy: fixed as of the pinned MI355X image (`v0.5.13.post1-rocm720-mi35x-20260618`).** Earlier SGLang ROCm images miscompiled AMD aiter's `gemm_a8w8_blockscale_bpreshuffle` GEMM on gfx950 (ROCm 7.2): the error was small per layer but compounded across all 78 layers and silently corrupted output — in-context reasoning broke (GSM8K ≈ 0) while short factual prompts still looked fine. The root cause was a gfx950/ROCm-7.2 miscompile of the CK kernel (a packed illegal-type FMA that relied on an LLVM coercion pass removed in ROCm 7.2; non-deterministic wrong rows near tile boundaries). This is resolved in the pinned image and newer: GLM-5.2-FP8 on MI350X/MI355X (gfx950) was re-validated at TP4 and TP8 — **GSM8K ≈ 0.96 (0% invalid)** and **15/15 needle-in-haystack retrieval to ~118K tokens**. **MI300X / MI325X (gfx942) were never affected.** If you must run an older image, treat gfx950 FP8 output as unverified. Background: [sgl-project/sglang#28685](https://github.com/sgl-project/sglang/issues/28685) (analysis) and the upstream CK fix [ROCm/rocm-libraries#8639](https://github.com/ROCm/rocm-libraries/pull/8639) (scalar FMA + accumulator anchor; restores correctness and determinism at -O3).
</Note>

- **MTP / EAGLE speculative decoding on AMD.** Five-step MTP is validated for `amd/GLM-5.2-MXFP4` on MI355X with `lmsysorg/sglang-rocm:v0.5.20-rocm720-mi35x-20260920`. Choose **Low-Latency** for TP8/EP1, or **High-Throughput** for TP4/EP4 with HiCache (`--enable-hierarchical-cache --hicache-ratio 1.0`). HiCache uses the default `kernel`, `page_first`, and `write_through` settings. MTP remains unvalidated for GLM-5.2 on MI300X/MI325X and for the other MI355X checkpoint precisions.
- **MTP / EAGLE speculative decoding on AMD.** Five-step MTP is validated for `amd/GLM-5.2-MXFP4` on MI355X with `lmsysorg/sglang-rocm:v0.5.20-rocm720-mi35x-20260924`. Choose **Low-Latency** for TP8/EP1, or **High-Throughput** for TP4/EP4 with HiCache (`--enable-hierarchical-cache --hicache-ratio 1.0`). HiCache uses the default `kernel`, `page_first`, and `write_through` settings. MTP remains unvalidated for GLM-5.2 on MI300X/MI325X and for the other MI355X checkpoint precisions.

## 3. Advanced Usage

Expand Down
Loading