Repository navigation
Conversation
…0.1.33 vs 0.1.40, prefill-streaming A/B Results only: README, per-run JSONL/CSV, engine logs, RSS samples, needle results and the scripts. Main finding: 0.1.40 prefill is 6.5-7x slower than 0.1.33 on this GPU at the defaults; STRATA_PREFILL_STREAM_MIN=65536 restores and exceeds 0.1.33 speed.
…pt-length sweep Adds the DDR4-3200 re-run of the community measurements and a prompt-length sweep (16k..256k tokens, -c 262144) for 0.1.33, 0.1.40 and 0.1.40 with STRATA_PREFILL_STREAM_MIN=65536; raw runs, tables script and sweep scripts. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Maddoxkkm
added a commit
to Maddoxkkm/Strata
that referenced
this pull request
Oct 11, 2026
The four gfx1100 community issues (Niko1221#1655 Niko1221#1459 Niko1221#1406 Niko1221#1292) were read in full and triaged to this box; the useful part is in scratch/reports/COMMUNITY-KNOBS.md. Measured here at 16K, matched budgets: base 1,870.9 tok/s +STRATA_RAM_ADAPT=1 1,950.7 (+4.3%) +STRATA_PA_FAST=1 +PF_SWITCH_MIN_T=4096 1,979.0 (+5.8%) +STRATA_PF_GEMM=1 +PF_PAD=1 1,990.1 (+6.4%) +--prefill 32768 1,919.0 (+2.6%) +STRATA_PF_FUSED=1 1,898.2 (+1.5%) Combined (the three winners) 2,012.9 at 16K (+6.4%) and 945.3 at 4K (-0.1%); a 32K chunk adds nothing on top. All these are bit-changing switches, so switches-accuracy.py grades answers off vs on: the 16K needle passes in both and the one differing case ('strata' backwards) is flaky off AND on (3/5 each) - the 2-bit model, not the switches. Everything adopted; 73K real text through the server is now 1,784.3 tok/s (+7.8% today, +41% over the session's 1,264.5). Two corrections to the docs, from measurement: the WMMA/SIMT switches labelled gfx12 act on gfx1100, and on-device kernel-shape autotuning was a no-op here exactly as Niko1221#1292 reports. Also fixed, after this box hard-froze: stop-server.bat killed a fixed list of old binary names, so newer A/B engines leaked (~12 GiB each) until free RAM hit 0.0 GiB. It now kills any strata*.exe plus tune_decode, and the lessons (one engine at a time, budget as a validity gate, the RAM tier being the top lever) are in the strata-local-tuning skill.
Owner
|
Thanks for the pull request. We're resetting the tracker for 0.2.0-beta. There are more than 1,300 open issues and pull requests, too many for us to review properly, so we're closing everything that isn't being worked on right now. Closed here does not mean rejected. To reopen it takes about two minutes:
The rules are in CONTRIBUTING.md. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Issue: n/a
Summary
Results only, no engine changes. Adds one community benchmark report for an RX 7900 XTX 24 GB (gfx1100, clock capped at 2850 MHz), Ryzen 9 5900X, 64 GB DDR4 (54 GB visible to WSL2), Windows 11 + Ubuntu 26.04 in WSL2, ROCm 7.14. Model: Qwen3.8-Flash-Next GSQ-RCO IQ1_M Coder pack + MTP,
--mmap-experts,--kv int8.What changed
Only
bench/results/2026-10-06-community-rx7900xtx-wsl2/is added: README following docs/COMMUNITY_BENCHMARKS.md, per-run JSONL/CSV, engine logs, RSS samples, needle results, and the scripts that produced them.Measured:
Extra Notes
Limitations (also in the README): one machine, model and pack; sweep prompts are source code, one text per length; page-cache state and disk read rate not recorded in the sweep (compute vs file cache is open); the step in 0.1.33 prefill between 17k and 32k tokens was not investigated; long-context quality checked only with needle (to 128k).
Branch is two commits on current main; nothing outside the results folder is touched.