record(LTX25-DECODE-SPEED): three scalar single-threaded loops in the LTX CPU path, and the number ltx2_video.cpp said was owed - #1214
Merged
Conversation
… LTX CPU path, and the number ltx2_video.cpp said was owed Three defects found while attempting the first full-model LTX-2.5 render are filed and now have an owner. None is fixed here; this change is the record. They are one story. The LTX prep path hand-rolls scalar arithmetic where the rest of the tree routes through the `vt::` GEMM seam, and it does so in three places that all sit between `Load` and the first denoise step: - #1202 `Ltx2FuseLoraIntoTensor` computes the LoRA product with a triple scalar loop at a measured ~0.53 GFLOP/s, one thread of twenty. - #1208 the text tower's `Linear` is scalar, single-threaded and accumulates in `double`, which also diverges from the `F.linear` it cites as reference. - #1210 the two-stage rebind fuses at load, un-fuses for phase 0 and re-fuses for phase 1, so the load-time pass is provably wasted. #1210 supplies a number the tree explicitly asked for. `ltx2_video.cpp:2849` records the two-rebind cost as "UNMEASURED on real weights" and says a later perf row owns it. This is that measurement. The owner is `ltx25-decode-speed.md`, whose §5 already frames "why the decode is single-threaded and on the host". These extend that finding from the decode to the load and prep path, so they belong to its `## Owed` table rather than to a new row. Scope is deliberately narrow. The rates are measured over a 10.4-minute window and extrapolated; no full-model pass has completed, so no end-to-end figure is claimed. What is measured is stated as measured and what is projected is labelled. `test_cpu_x86_llamacpp_floor` fails in the local preflight with exit 4 (`NO_QUIET_WINDOW`) instead of 2 at loadavg 68.9-79.9, which is #618 exactly. This diff touches two files under `.agents/` and cannot reach a CPU benchmark harness. `check-agent-record.py`, `check-public-doc-tables.py`, doc-checkpoint and now-current all pass. FOLLOWING_AGENTS_PROTOCOL Following-Agents-Protocol: true AI-Assisted: true Assisted-by: AGENT:claude-opus-5 [claude-code]
…b-only conflict GitHub refused the squash with "Pull Request has merge conflicts" while the local merge is clean, because GitHub ignores the `merge=union` driver that `.gitattributes` sets on `.agents/issue-index.md`. Both sides appended rows and only the forge read that as a conflict. The union result is verified rather than assumed, because that driver can interleave rows and can reinstate a deleted one: 367 rows, zero duplicate row keys, and this branch's #1202, #1208 and #1210 each present exactly once. `check-issue-index-append-only.py` and `check-agent-record.py` both pass. This branch carries no product code, so the merge cannot change a build result. FOLLOWING_AGENTS_PROTOCOL Following-Agents-Protocol: true AI-Assisted: true Assisted-by: AGENT:claude-opus-5 [claude-code]
#1209 landed and appended its own rows to `.agents/issue-index.md`, so GitHub refused the squash again while the local merge stays clean. GitHub ignores the `merge=union` driver `.gitattributes` sets on that path, and every concurrent append therefore reads as a conflict on the forge only. Verified rather than assumed, because the union driver can interleave rows and reinstate a deleted one: 376 rows, zero duplicate row keys, and this branch's #1202, #1208 and #1210 each present exactly once. `check-issue-index-append-only.py` and `check-agent-record.py` both pass. This branch carries no product code, so the merge cannot change a build result. FOLLOWING_AGENTS_PROTOCOL Following-Agents-Protocol: true AI-Assisted: true Assisted-by: AGENT:claude-opus-5 [claude-code]
localai-bot
pushed a commit
that referenced
this pull request
Aug 18, 2026
`origin/main` moved from `ae581da3e` to `1dac4f9a7` while this row was in review, and the five new commits touch four of the same record files. None of them touches the depth schedule, the music3 sources, or this row's spec section, so this merge carries records only and no code. `docs/STATUS.md` conflicted on two adjacent rows, because #1214 and #1252 rewrote the LTX-2.5 and MiniMax-Music3 rows while this branch held older text for both. Main's complete version is taken here, byte for byte, including its new music3 wording. This row's own claim is added back in the commit that owns it rather than inside a merge resolution, so the record edit and the measurement it describes stay in one reviewable commit. `.agents/benchmark-record.md` and `.agents/issue-index.md` were rebuilt the way both files require: main's version verbatim, then this row's entry and its two index rows appended at the end. Main's text is an exact prefix of each result, verified by comparison rather than by reading. FOLLOWING_AGENTS_PROTOCOL Following-Agents-Protocol: true AI-Assisted: true Assisted-by: AGENT:claude-opus-5 [Claude Code]
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.
record(LTX25-DECODE-SPEED): three scalar single-threaded loops in the LTX CPU path, and the number ltx2_video.cpp said was owed
Three defects found while attempting the first full-model LTX-2.5 render are
filed and now have an owner. None is fixed here; this change is the record.
They are one story. The LTX prep path hand-rolls scalar arithmetic where the
rest of the tree routes through the
vt::GEMM seam, and it does so in threeplaces that all sit between
Loadand the first denoise step:Ltx2FuseLoraIntoTensorcomputes the LoRA product with a triplescalar loop at a measured ~0.53 GFLOP/s, one thread of twenty.
Linearis scalar, single-threaded and accumulatesin
double, which also diverges from theF.linearit cites as reference.for phase 1, so the load-time pass is provably wasted.
#1210 supplies a number the tree explicitly asked for.
ltx2_video.cpp:2849records the two-rebind cost as "UNMEASURED on real weights" and says a later
perf row owns it. This is that measurement.
The owner is
ltx25-decode-speed.md, whose §5 already frames "why the decodeis single-threaded and on the host". These extend that finding from the decode
to the load and prep path, so they belong to its
## Owedtable rather than toa new row.
Scope is deliberately narrow. The rates are measured over a 10.4-minute window
and extrapolated; no full-model pass has completed, so no end-to-end figure is
claimed. What is measured is stated as measured and what is projected is
labelled.
test_cpu_x86_llamacpp_floorfails in the local preflight with exit 4(
NO_QUIET_WINDOW) instead of 2 at loadavg 68.9-79.9, which is #618 exactly.This diff touches two files under
.agents/and cannot reach a CPU benchmarkharness.
check-agent-record.py,check-public-doc-tables.py, doc-checkpointand now-current all pass.
FOLLOWING_AGENTS_PROTOCOL
Following-Agents-Protocol: true
AI-Assisted: true
Assisted-by: AGENT:claude-opus-5 [claude-code]