Yes-- it is real-world! These are results on the spark test suite on the new optimizations. We see peak 25% speedup We serve 1 pathway on the previous so there's no loss on the serving path.

New 10% to 15% Gains! We're publishing new expert gate kernel optimizations: these provide 10% to 15% speed up when 2-6 users are active (apply to 2-8 pathways). github.com/CurtisAccelerate/…

Struggle is real, even though the outcome may be predetermined... replicator or entropy... whatever the force

😂 really does not want to reopen the mixed weights lane...

Pushed the bot too hard.

Quants? you want to complain about the DeepSeek is 🔥🔥🔥 This is 60 minutes autonomous

New DeepSeek v4.1 1017 Optimizations! - Extended the new attention/drafter performance improvements from 1 user out to 8 user. The relative improvement is less with more concurrent streams because there's less slack and 7 requests fare worse due to specific padding but overall we saw gains across the board. - We added sampled drafting. Most benchmarks are tested at temp 0 but Hermes/agentic default temp is 1. This drops the acceptance rate of greedy drafters on real-world tasks where the temp is not zero. Our current recommendation is greedy for temp 0, sampled for 1 request, and greedy for 2-8. We haven't tested sampled under concurrent requests. It requires some harness/serving intelligence. You can use the new optimizations with either greedy or sampled depending on your use case. For future optimizers/optimizations, testing & optimizing drafters under real-world conditions may yield most value. github.com/CurtisAccelerate/…

DeepSeek 4.1 1017 Quality? I notice some people reporting the EXL3 weights on DeepSeek 4.1 1017 produce poor outputs in reply to my build and others. It has not been my experience with my build. It has been extremely strong in my testing. For example, I gave it a single prompt and it worked for 30 minutes building an extremely polished game that worked perfectly on a single prompt. And, I have it working on even longer run tasks. But we did something different: we use some of the FP8 weights for important aspects. I don't know if that makes the difference or if its the harness or server. But, this model has impressed me and worked extremely well.

Something to know: The largest perf gains, from the new attention + drafter, are only seen at concurrency/request depth 1 for now when more lanes are active, it does not use the new drafter and falls back to the previous optimized default. This is not an inherent limitation though the relative advantage would likely decline at higher concurrency due to having more parallel work. In order to support multiple requests, requires batch-aware buffers. The next obvious target would be request 2. Example, on this sweep only users 1 is the new attention: