feat: tok/s regression guards + README factual update#23
Merged
dexwritescode merged 2 commits intomainfrom May 7, 2026
Merged
Conversation
Integration tests: - Add GenerateThroughput test to Mistral, Llama3, Gemma, Qwen3 MoE - 128-token greedy run with mx::compile warmup pass before timing - Floors set to ~50% of measured debug-build baseline to catch regressions back to pre-pipelining speeds README: - Replace ASCII architecture diagram with Mermaid - Add Qwen3 (dense) and Qwen3 MoE to supported models table - Fix model class names (GemmaModelMLX, Qwen3MoeModelMLX) - Simplify backend section — MLX is the only backend today - Remove cuBLAS from intro (not yet implemented) - Add Performance section with release-build tok/s measurements - Update roadmap: mark Phase O (MLX perf) done, Qwen3 MoE in Phase F
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
GenerateThroughputtest to Mistral 7B, Llama-3.1 8B, Gemma 3 1B, and Qwen3 MoE integration test suites. Each test runs a 128-token greedy generation with an 8-token warmup pass to triggermx::compilebefore timing, then asserts a floor of ~50% of measured baseline — enough to catch any regression back to pre-pipelining speeds (~10-12 tok/s)Throughput floors (debug build, post-warmup):
Performance numbers in README (release build, greedy):
TinyLlama 1.1B ~265 tok/s · Gemma 3 1B ~190 tok/s · Llama-3.1 8B ~61 tok/s · Mistral 7B ~57 tok/s · Qwen3.6 35B-A3B ~77 tok/s