LoreonLabsPlatform

Overview

Intelligence

Markets
Builders
Research
Ecosystems
Launchpads

Search

Other

fmha_native

rocking5566/fmha_native

OtherEmerging

Stars

—

Forks

—

Contributors

1

Last push

14d ago

Recent commits

Latest commits.

docs: refresh README perf tables (post -enable-post-misched=1)
dfa9d41Po Yen Chen15d ago
fmha_native: enable post-RA misched (-enable-post-misched=1) to hide dep latency
e10ae57Po Yen Chen15d ago
docs: refresh README perf tables (paired native+CK run, post address-math fixes)
2af6d83Po Yen Chen15d ago
fmha_native: strength-reduce K async-copy address math (single-recurrence IV)
7e710eePo Yen Chen15d ago
fmha_native: strength-reduce V-load address math (induction variable)
e5c6d61Po Yen Chen

15d ago

docs: refresh perf tables with consecutive same-GPU benchmark run

61d30bcPo Yen Chen16d ago

docs: fix CK perf numbers - use 6-run avg framework (was best-of-5, biased high)

551bf92Po Yen Chen16d ago

docs: drop src/fused/README.md, keep perf table in root README only

1b6dd46Po Yen Chen16d ago

Top contributors

Builders behind this project.