Decision Workspace
flash-attention-cpu vs ruvector-attention-wasm vs omega-attention
Side-by-side comparison of Rust crates
40
flash-attention-cpu
experimentalv0.2.0
Pure-Rust, CPU-optimized Flash Attention: tiled online-softmax, AVX-512/AVX2/NEON/WASM-SIMD128 with scalar fallback, Rayon parallelism.
57
ruvector-attention-wasm
growingv2.3.0
High-performance WebAssembly attention mechanisms: Multi-Head, Flash, Hyperbolic, MoE, CGT Sheaf Attention with GPU acceleration for transformers and LLMs
41
omega-attention
growingv1.1.0
Attention mechanisms for ExoGenesis Omega - 39 attention types for brain-like selective processing
Core Metrics
| flash-attention-cpu | ruvector-attention-wasm | omega-attention | |
|---|---|---|---|
| Health Score | 40 | 57 | 41 |
| Total Downloads | 28 | 124 | 86 |
| 30d Downloads | 0 | 0 | 0 |
| Dependents | 0 | 0 | 0 |
| Releases | 1 | 6 | 2 |
| Last Updated | 88d ago | 13d ago | 302d ago |
| Age | 2m | 10m | 10m |
Health Breakdown
flash-attention-cpu
Maintenance
8
Quality
12
Community
6
Popularity
2
Documentation
12
ruvector-attention-wasm
Maintenance
18
Quality
17
Community
6
Popularity
3
Documentation
13
omega-attention
Maintenance
7
Quality
16
Community
6
Popularity
2
Documentation
10
Technical Details
| flash-attention-cpu | ruvector-attention-wasm | omega-attention | |
|---|---|---|---|
| Version | 0.2.0 | 2.3.0 | 1.1.0 |
| Stable (≥1.0) | ✗ No | ✓ Yes | ✓ Yes |
| License | MIT OR Apache-2.0 | MIT | MIT |
| Dependencies | 5 | 9 | 13 |
| Crate Size | 110KB | 86KB | 25KB |
| Features | 0 | 1 | 0 |
| Yanked % | 0.0% | 0.0% | 0.0% |
| Edition | 2021 | 2021 | 2021 |
| MSRV | 1.89 | — | — |
| Owners | 1 | 1 | 1 |
Links
Quick Verdict
- •ruvector-attention-wasm leads with a health score of 57/100, but none of the options score above 80.
- •flash-attention-cpu is pre-1.0 — API may change.