Morph MORPH
SIGN IN SIGN UP
dhruvhead / cuda-optimized-skill UNCLAIMED
0 0 0 Python

COMMITS

main
April 22, 2026
K
update README.md
KernelFlow-ops committed
f805c72
K
Upgrade optimizer with roofline-guided budgeting, branch exploration, ablation attribution, SASS verification, and richer state tracking for bottleneck history, implementation failures, frontier candidates, and better iterative learning.
KernelFlow-ops committed
f5ffe7a
April 20, 2026
K
update README.md
KernelFlow-ops committed
74ee9ec
K
The entire skill has been refactored, resulting in a more complete structure, and the README.md file has been updated
KernelFlow-ops committed
b729ffc
April 18, 2026
K
修改了reference
KernelFlow-ops committed
7017ca9
April 16, 2026
K
添加了持续优化记忆，可以在之前的基础上继续迭代，完善README.md，进一步完善了reference中cuda知识库
KernelFlow-ops committed
ba52759
April 12, 2026
K
README.md中加入了框架图
KernelFlow-ops committed
dad497e
K
完善README.md
KernelFlow-ops committed
a4ce8c0
K
将策略记忆机制真正落地到了 loop 脚本，并且实现了 Exa MCP 官方文档检索策略
KernelFlow-ops committed
112a23c
April 11, 2026
K
适配优化cuda cutlass triton的算子
KernelFlow-ops committed
b75c651
M
补上了这个 skill 的运行前环境检查
mark-liu committed
e87be8f
April 10, 2026
M
v1.5
mark-liu committed
6528b77
M
修改README.md
mark-liu committed
1f828f6
M
Merge branch 'main' of https://github.com/KernelFlow-ops/cuda-optimized-skill
mark-liu committed
76e8710
A
Initial commit
aa24aa committed
ee242de
M
v1.5
mark-liu committed
84b0f82
M
v1.2
mark-liu committed
33b2791
M
v1.1
mark-liu committed
9c2b8b8
April 9, 2026
M
v1.0
mark-liu committed
4a00fb5