# Open Kernel grouped NVFP4 experts with bit decoding | PTLLM > Cheaper bit decoding improves the earlier grouped implementation at 0.2087x but leaves every custom multi-token case slower than SGLang. The one-token… Canonical: https://posttrainllm.com/experiments/openkernel-moe-grouped-bitdecode-20261024 Experiment · runtime perf Cheaper bit decoding improves the earlier grouped implementation at 0.2087x but leaves every custom multi-token case slower than SGLang. The one-token result uses the unchanged Marlin fallback. Register pressure, spills and instruction costs remain unmeasured. Retained outcome: regressed. **Outcome:** regressed **Family:** runtime perf **Evidence confidence:** exact **Record:** infrastructure ## Evidence readout 2026-10-23 H200 check 63bd8567d57d4199 passed all five fresh correctness cases with no flags: 0.2905x aggregate, reported 95% interval 0.2861-0.2948x. This is an off-board check, not a scored submission. ## What the attempt taught Cheaper bit decoding improves the earlier grouped implementation at 0.2087x but leaves every custom multi-token case slower than SGLang. The one-token result uses the unchanged Marlin fallback. Register pressure, spills and instruction costs remain unmeasured. ## Why it stopped or stayed bounded Aggregate 0.2905x is below selected 1.2236x. At 16 tokens the candidate took 964.156 microseconds versus 189.056 for SGLang and 157.29 for the selected source. Multi-token diagnostic profiles place 98-99 percent of GPU time in the two grouped GEMMs. ## Method and scope - Experiment family: runtime perf. - Record kind: infrastructure. ## Disposition Experiment closed after scored run beb82b4224554523 finished at rank 283/749, outside the owner top-100 target. Retain the research artifact; no further jobs. This retained disposition is historical evidence, not authorization to restart the experiment. A new run needs a fresh question, frozen evaluator, explicit resource budget, and scoped tracking issue. ## How to read this dossier This canonical page separates the retained repository decision from the material that informed it. A source may describe an external claim, historical measurement, local observation, or planned procedure; those evidence classes are not interchangeable. Follow the provenance links before reusing a number or method. Status and disposition describe what PostTrainLLM retained when this record was indexed. They are not a live product promise, a newly run benchmark, or permission to restart historical work. Unknown or unmeasured fields remain unknown rather than being treated as zero. To reuse the record, first name the exact claim you need and trace it to the linked source. Then check whether the original environment, model revision, data split, evaluator, hardware, and budget match the proposed use. If they do not, treat the record as a hypothesis or design reference and run the smallest fresh comparison that can falsify it. Preserve negative outcomes and regressions beside any improvement; a local win on one slice does not silently become a general capability claim. ## Source provenance The normalized record comes from `docs/attempts.json`. The links below are the tracked evidence and explanatory sources preserved with the record. - [docs/learn/openkernel.md](https://posttrainllm.com/docs/learn/openkernel) - [experiments/openkernel/results.json](https://github.com/PostTrainLLM/posttrainllm/blob/main/experiments/openkernel/results.json) ## Related evidence - [Primary proof](https://posttrainllm.com/docs/learn/openkernel) - [Applicable recipe](https://posttrainllm.com/recipes) - [Learning path](https://posttrainllm.com/learn/paths/runtime-and-agents) Canonical record: `experiment:openkernel-moe-grouped-bitdecode-20261024`