Repository navigation
bench: RX 7900 XTX + 6800 XT helper layout, 1M and long-output results - #1459
Closed
saikiran-rs wants to merge 2 commits into
Closed
saikiran-rs wants to merge 2 commits into
saikiran-rs wants to merge 2 commits into
Conversation
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.
giridhart
pushed a commit
to giridhart/Strata
that referenced
this pull request
Oct 15, 2026
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
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.
Results-only community report for a Linux RX 7900 XTX + RX 6800 XT helper layout with a Xeon E5-2696 v4 and 121 GiB OS-visible RAM.
Measurements
Reproduction and limits
Includes exact compressed prompts, per-request engine/client timings, actual token counts, portable configs, pinned model URLs/hashes, source/build provenance and a standalone client/quality harness. Historical source revisions are preserved on the contributor's fork. Offline fixture hashes and pinned-API token counts were checked; the SSE parser was tested with a synthetic stream.
These are historical measurements from a combined0.1.40.2-derived research build, not isolated measurements of a new PR or stock0.1.40.3. All changed settings and inherited contributor work are identified. Q8-kernel-only whole-model results include regressions. The separately managed NVIDIA service's host load was not isolated. No new inference was run for publication.
This report follows docs/COMMUNITY_BENCHMARKS.md and is separate from the helper-allowance engine proposal. No universal speedup or general quality equivalence is claimed.