Skip to content
Navigation Menu
Sign in
Appearance settings
Platform
AI CODE CREATION
GitHub Copilot
Write better code with AI
GitHub Copilot app
Direct agents from issue to merge
MCP Registry
Integrate external tools
DEVELOPER WORKFLOWS
Actions
Automate any workflow
Codespaces
Instant dev environments
Issues
Plan and track work
Code Review
Manage code changes
Code Quality
Enforce quality at merge
APPLICATION SECURITY
GitHub Advanced Security
Find and fix vulnerabilities
Code security
Secure your code as you build
Secret protection
Stop leaks before they start
EXPLORE
Why GitHub
Documentation
Blog
Changelog
Marketplace
View all features
Solutions
BY COMPANY SIZE
Enterprises
Small and medium teams
Startups
Nonprofits
BY USE CASE
App Modernization
DevSecOps
DevOps
CI/CD
View all use cases
BY INDUSTRY
Healthcare
Financial services
Manufacturing
Government
View all industries
View all solutions
Resources
EXPLORE BY TOPIC
AI
Software Development
DevOps
Security
View all topics
EXPLORE BY TYPE
Customer stories
Events & webinars
Ebooks & reports
Business insights
GitHub Skills
SUPPORT & SERVICES
Documentation
Customer support
Community forum
Trust center
Partners
View all resources
Open Source
COMMUNITY
GitHub Sponsors
Fund open source developers
PROGRAMS
Security Lab
Maintainer Community
GitHub Stars
Archive Program
REPOSITORIES
Topics
Trending
Collections
Enterprise
ENTERPRISE SOLUTIONS
Enterprise platform
AI-powered developer platform
AVAILABLE ADD-ONS
GitHub Advanced Security
Enterprise-grade security features
Copilot for Business
Enterprise-grade AI features
Premium Support
Enterprise-grade 24/7 support
Pricing
Search
/
Sign in
Sign up
Appearance settings
You signed in with another tab or window.
Reload
to refresh your session.
You signed out in another tab or window.
Reload
to refresh your session.
You switched accounts on another tab or window.
Reload
to refresh your session.
Dismiss alert
{{ message }}
pytorch
/
helion
Public
Notifications
You must be signed in to change notification settings
Fork
174
Star
951
Code
Issues
87
Pull requests
210
Actions
Projects
Security and quality
0
Insights
Additional navigation options
Code
Issues
Pull requests
Actions
Projects
Security and quality
Insights
Commits
Branch selector
main
User selector
All users
Datepicker
All time
Commit history
Commits on Sep 23, 2026
[cutedsl] Search the KV tile width and seed the sm_103 dense shapes at 160 (#3576)
yushangdi
authored
ad71ac2
View commit details
Copy full SHA for ad71ac2
Browse repository at this point
[cutedsl] Make the flash KV tile width a real dimension (#3575)
yushangdi
authored
5a3606a
View commit details
Copy full SHA for 5a3606a
Browse repository at this point
[cutedsl] Run the dense resident softmax on the one-CTA FA4 pipeline (#3574)
yushangdi
authored
53c23de
View commit details
Copy full SHA for 53c23de
Browse repository at this point
[cutedsl] Gate the dense flash softmax lowering on its schedule, not a seed match (#3573)
yushangdi
authored
beefa1c
View commit details
Copy full SHA for beefa1c
Browse repository at this point
[config] Enable non-pow-of-2 values for num_sm_multiplier and maxnreg (#3649)
Show description for 54166b5
ethche
and
Ethan Che
authored
54166b5
View commit details
Copy full SHA for 54166b5
Browse repository at this point
Commits on Sep 22, 2026
[cute] Fix tile layouts and gather-load vectorization (#3648)
yushangdi
authored
e4ba356
View commit details
Copy full SHA for e4ba356
Browse repository at this point
Commits on Sep 21, 2026
torch_tpu: update pin to 2e3a5d3a984094911fa42c899934e5e439b44537 (#3651)
cota
authored
204098b
View commit details
Copy full SHA for 204098b
Browse repository at this point
[benchmark] Pin pretuned vLLM baseline to 0.29.0 (#3650)
yushangdi
authored
fd57d97
View commit details
Copy full SHA for fd57d97
Browse repository at this point
[cutedsl] Enable direct affine-scan schedules (#3639)
choijon5
authored
7e57e15
View commit details
Copy full SHA for 7e57e15
Browse repository at this point
[cutedsl] Replay affine scans through generated AST (#3638)
choijon5
authored
d913723
View commit details
Copy full SHA for d913723
Browse repository at this point
[cutedsl] Add affine-scan device primitives (#3637)
choijon5
authored
157d5d0
View commit details
Copy full SHA for 157d5d0
Browse repository at this point
[cutedsl] Discover and plan short affine scans (#3636)
choijon5
authored
68e7452
View commit details
Copy full SHA for 68e7452
Browse repository at this point
[cutedsl] Add transactional late-codegen provenance (#3635)
choijon5
authored
ad7fcd3
View commit details
Copy full SHA for ad7fcd3
Browse repository at this point
benchmarks: extend recurrent KDA comparisons (#3634)
choijon5
authored
2ff9935
View commit details
Copy full SHA for 2ff9935
Browse repository at this point
Commits on Sep 20, 2026
[megakernel] Add cross_loop_pipeline = dynamic (#3645)
Show description for 95b31a0
ethche
and
Ethan Che
authored
95b31a0
View commit details
Copy full SHA for 95b31a0
Browse repository at this point
[megakernel] refactor: enable the scheduler to handle symbolic shapes (#3644)
Show description for e0e2d33
ethche
and
Ethan Che
authored
e0e2d33
View commit details
Copy full SHA for e0e2d33
Browse repository at this point
[ROCM] FlyDSL: Support wavefront cordination along the row dimension( W>1) for reduction kernels. (#3630)
umechand-amd
authored
7868a43
View commit details
Copy full SHA for 7868a43
Browse repository at this point
Commits on Sep 19, 2026
[ROCm][CI] Update ROCm version to 10.0 (#3626)
Show description for ebb05c9
akashveramd
and
cursoragent
authored
ebb05c9
View commit details
Copy full SHA for ebb05c9
Browse repository at this point
[Pallas] Make pallas_emit_pipeline_group_size autotunable (#3647)
AmesingFlank
authored
0b66239
View commit details
Copy full SHA for 0b66239
Browse repository at this point
Commits on Sep 18, 2026
[Pallas] Fold narrowing left-input casts into dot (#3633)
AmesingFlank
authored
98a7e52
View commit details
Copy full SHA for 98a7e52
Browse repository at this point
[Pallas] Expose the TPU low-level scheduler for pipelines (#3632)
AmesingFlank
authored
8a68615
View commit details
Copy full SHA for 8a68615
Browse repository at this point
[Pallas] Decouple emit-pipeline DMA and compute tiles (#3631)
AmesingFlank
authored
0cbee6f
View commit details
Copy full SHA for 0cbee6f
Browse repository at this point
Commits on Sep 17, 2026
[Pallas] Share nested DMA address setup across streams (#3566)
AmesingFlank
authored
786792f
View commit details
Copy full SHA for 786792f
Browse repository at this point
[Pallas] Pipeline nested DMA panels across outer loops (#3565)
AmesingFlank
authored
04600d9
View commit details
Copy full SHA for 04600d9
Browse repository at this point
Commits on Sep 16, 2026
[autotuner] Fix FROM_BEST_AVAILABLE never matching cached configs (#3642)
fulvius31
authored
9b6deab
View commit details
Copy full SHA for 9b6deab
Browse repository at this point
[cutedsl] flash-bwd: group-local S-t2r barriers, diagonal-only causal mask, cute_flash_bwd_exp2_f32 (#3625)
Show description for 8deae06
jansel
authored
8deae06
View commit details
Copy full SHA for 8deae06
Browse repository at this point
[cutedsl] flash-bwd: CLC dynamic persistent tile scheduling (cute_flash_bwd_persistent) (#3624)
Show description for 04f5c79
jansel
authored
04f5c79
View commit details
Copy full SHA for 04f5c79
Browse repository at this point
[cutedsl] flash-bwd: dK/dV epilogue via smem staging + TMA store; dV overlaps the tail gemms (#3623)
Show description for b1029b8
jansel
authored
b1029b8
View commit details
Copy full SHA for b1029b8
Browse repository at this point
[cutedsl] flash-bwd 2-CTA: rebuild the cluster core on the FA4 structure (#3622)
Show description for 585d900
jansel
authored
585d900
View commit details
Copy full SHA for 585d900
Browse repository at this point
[cutedsl] flash-bwd: fix the persistent tile-scheduler deadlock and two cross-tile races (#3621)
Show description for f49ec56
jansel
authored
f49ec56
View commit details
Copy full SHA for f49ec56
Browse repository at this point
[cutedsl] flash-bwd 2-CTA: balance the setmaxnreg budgets (fixes the fresh-compile hang) (#3620)
Show description for 5bb041f
jansel
authored
5bb041f
View commit details
Copy full SHA for 5bb041f
Browse repository at this point
[cutedsl] flash-bwd: 4-deep Q/dO TMA rings at head_dim 64 (#3619)
Show description for 69f4eed
jansel
authored
69f4eed
View commit details
Copy full SHA for 69f4eed
Browse repository at this point
[cutedsl] flash-bwd: TMA tensor reduce-add dQ drain, honored register split, lean MMA warp (#3618)
Show description for b23876b
jansel
authored
b23876b
View commit details
Copy full SHA for b23876b
Browse repository at this point
[cutedsl] flash-bwd: conflict-free padded dQ staging; persistent forced off (#3617)
Show description for 9edd910
jansel
authored
9edd910
View commit details
Copy full SHA for 9edd910
Browse repository at this point
[cutedsl] flash-bwd: persistent tile scheduler knob, segmented causal walk (#3616)
Show description for be3c4be
jansel
authored
be3c4be
View commit details
Copy full SHA for be3c4be
Browse repository at this point
Previous
Next
You can’t perform that action at this time.