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 }}
Uh oh!
There was an error while loading.
Please reload this page
.
ROCm
/
TransformerEngine
Public
Notifications
You must be signed in to change notification settings
Fork
38
Star
76
Code
Issues
8
Pull requests
62
Actions
Projects
Security and quality
0
Insights
Additional navigation options
Code
Issues
Pull requests
Actions
Projects
Security and quality
Insights
Add ROCm Triton blockwise FP8 grouped GEMM
- #716
#716
Open
sudhu2k
wants to merge 18 commits into
dev
ROCm/TransformerEngine:dev
from
sudhu/blockwise_fp8_grouped_linear
ROCm/TransformerEngine:sudhu/blockwise_fp8_grouped_linear
Copy head branch name to clipboard
Conversation
Commits
18
(18)
Checks
Files changed
Open
Add ROCm Triton blockwise FP8 grouped GEMM
#716
sudhu2k
wants to merge 18 commits into
dev
ROCm/TransformerEngine:dev
from
sudhu/blockwise_fp8_grouped_linear
ROCm/TransformerEngine:sudhu/blockwise_fp8_grouped_linear
Copy head branch name to clipboard
Commits
Commits on Aug 18, 2026
Add ROCm Triton blockwise FP8 grouped GEMM path
Show description for 3247c50
sudhu2k
committed
3247c50
View commit details
Copy full SHA for 3247c50
Browse repository at this point
Commits on Aug 25, 2026
Refactor ROCm blockwise FP8 grouped GEMM gating in GroupedLinear
Show description for 3a4951d
sudhu2k
committed
3a4951d
View commit details
Copy full SHA for 3a4951d
Browse repository at this point
Merge remote-tracking branch 'origin/dev' into sudhu/blockwise_fp8_grouped_linear
sudhu2k
committed
4bd2c0d
View commit details
Copy full SHA for 4bd2c0d
Browse repository at this point
Enable fused wgrad accumulation for ROCm blockwise FP8 grouped GEMM
Show description for bfea740
sudhu2k
committed
bfea740
View commit details
Copy full SHA for bfea740
Browse repository at this point
Avoid copying already-packed expert weights in blockwise FP8 grouped GEMM
Show description for 2cc840f
sudhu2k
committed
2cc840f
View commit details
Copy full SHA for 2cc840f
Browse repository at this point
Return GPU tensor split sizes from FP8 padding when given device copy
Show description for 09e72bc
sudhu2k
committed
09e72bc
View commit details
Copy full SHA for 09e72bc
Browse repository at this point
Tighten FP8 blockwise quantization test tolerances
Show description for 1006109
sudhu2k
committed
1006109
View commit details
Copy full SHA for 1006109
Browse repository at this point
Add Triton autotuning for ROCm blockwise FP8 grouped GEMM
Show description for 4dcedb3
sudhu2k
committed
4dcedb3
View commit details
Copy full SHA for 4dcedb3
Browse repository at this point
Fix grouped weight buffer validation and relax blockwise FP8 test tolerances
Show description for 71edbb8
sudhu2k
committed
71edbb8
View commit details
Copy full SHA for 71edbb8
Browse repository at this point
Commits on Aug 26, 2026
Refactor AMD Triton compiler knob handling for blockwise FP8 grouped GEMM
Show description for 8dd6441
sudhu2k
committed
8dd6441
View commit details
Copy full SHA for 8dd6441
Browse repository at this point
Refine ROCm blockwise FP8 grouped GEMM gating, quantization, and compiler-knob scoping
Show description for 04efa5b
sudhu2k
committed
04efa5b
View commit details
Copy full SHA for 04efa5b
Browse repository at this point
Support fp8_model_params in ROCm blockwise FP8 grouped GEMM
Show description for 01f95ae
sudhu2k
committed
01f95ae
View commit details
Copy full SHA for 01f95ae
Browse repository at this point
Commits on Aug 27, 2026
Fix single_grouped_weight grad-state mirroring and cache blockwise FP8 weight quantization
Show description for d0f81c4
sudhu2k
committed
d0f81c4
View commit details
Copy full SHA for d0f81c4
Browse repository at this point
Commits on Aug 28, 2026
Fix blockwise FP8 grouped-linear weight cache reuse and add unit test
Show description for 835742d
sudhu2k
committed
835742d
View commit details
Copy full SHA for 835742d
Browse repository at this point
Commits on Sep 1, 2026
Floor blockwise FP8 quantization amax and document ROCm XCD swizzling
Show description for b53d5d3
sudhu2k
committed
b53d5d3
View commit details
Copy full SHA for b53d5d3
Browse repository at this point
Merge remote-tracking branch 'origin/dev' into sudhu/blockwise_fp8_grouped_linear
sudhu2k
committed
cf53070
View commit details
Copy full SHA for cf53070
Browse repository at this point
Commits on Sep 2, 2026
Fix grouped-linear backward to return gradients matching forward inputs
Show description for 044d942
sudhu2k
committed
044d942
View commit details
Copy full SHA for 044d942
Browse repository at this point
Commits on Sep 3, 2026
Adjust quantization alignment for ROCm blockwise FP8 GEMM
Show description for 90d4569
sudhu2k
committed
90d4569
View commit details
Copy full SHA for 90d4569
Browse repository at this point
You can’t perform that action at this time.
