Context:
homelab
Cluster:
k3s
User:
marcos
Forge Rev:
v1.0 (C17)
K8s Rev:
v1.36.5+k3s1
CPU:
18%
MEM:
32%

describe(projects/stride)

Name:          stride
Title:         Stride
Kind:          SIMD kernels
Status:        Completed
Created:       2026-06-03 (152d ago)
Built With:    C, AVX2 assembly
Part Of:       The Optimiser's Kit
Highlights:
  - Fused single-pass AVX2 Adam, bit-for-bit identical to the C reference
  - 4.5x faster than scalar C in fp32, 2.1x in fp64 (one P-core, Core Ultra 7 155H)
  - Checked with a checkasm-style harness
Note:          Compute backend for Saddle.
Links:
  Source:      https://github.com/MarcosAsh/Stride
  Logs:        latest commits (l)
Events:        <none>