Skip to content

Pull requests: NVIDIA/cutlass

New pull request New
34 Open 511 Closed
34 Open 511 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

bugfix generic-k code in top-k with softmax
#1993 opened Dec 17, 2024 by t4c1 Loading…
6
fix bug: arch/mma_sm60.h Mma<2,2,1> calculate wrong
#1989 opened Dec 16, 2024 by wangdongxuking61 Loading…
Add half->int8 saturate conversion to promise valid range
#1983 opened Dec 12, 2024 by MARD1NO Loading…
Update 0x_gemm_tutorial.md
#1982 opened Dec 11, 2024 by foolusion Loading…
Fix assertion in integer_subbytes.h
#1961 opened Nov 25, 2024 by foreverlms Loading…
90 inactive-30d
#1935 opened Nov 11, 2024 by ckl117 Loading…
1
Blockwise Scaling for FP8
#1932 opened Nov 8, 2024 by manishucsd Loading…
15
add support for int4 in MMA_Traits inactive-30d
#1925 opened Nov 7, 2024 by luliyucoordinate Loading…
1
Distributed GEMM
#1907 opened Oct 29, 2024 by alihassanijr Loading…
28
Make CuTe and PyCuTe idx2crd Behaviors The Same inactive-30d
#1891 opened Oct 22, 2024 by leimao Loading…
10
fix Copy_Traits<SM80_CP_ASYNC_*_ZFILL> without definition of with(pred) function inactive-30d
#1887 opened Oct 20, 2024 by CalebDu Loading…
4
Fix for Host code binary operation. inactive-30d
#1855 opened Oct 8, 2024 by moradza Loading…
1
Add BF16 dtype mapping for python EVT inactive-30d
#1832 opened Sep 20, 2024 by gau-nernst Loading…
1
2
Fixing std::numeric_limits<half_t>::digits to include the implicit one inactive-30d inactive-90d
#1702 opened Aug 10, 2024 by akamiru Loading…
2
Avoid LDGSTS routing by changing default copy to be universalcopy inactive-30d inactive-90d
#1674 opened Aug 1, 2024 by ZelboK Loading…
13
feat: allow print_latex(TiledMMA) to colorize sliced thread and add print_latex(ThrMMA) inactive-30d inactive-90d
#1656 opened Jul 26, 2024 by cloudhan Loading…
2
Add atomic_add and BlockStripedReduce to bfloat162 inactive-30d inactive-90d
#1653 opened Jul 24, 2024 by yzh119 Loading…
3
Make mainloop schedule type available as GemmKernel::Schedule inactive-30d inactive-90d
#1638 opened Jul 16, 2024 by manishucsd Loading…
3
Make integer_subbyte Fully Compliant with is_integral inactive-30d inactive-90d
#1632 opened Jul 15, 2024 by osayamenja Loading…
9
Fix cute gemm dispatch-5 when C != D inactive-30d inactive-90d
#1618 opened Jul 6, 2024 by sjfeng1999 Loading…
5
Add Ampere GEMM example using Cute and CUTLASS 3.x inactive-30d inactive-90d
#1604 opened Jun 27, 2024 by aacostadiaz Loading…
7
Fix illegal memory accesses in multistage Mma's for k=0 inactive-90d
#1593 opened Jun 18, 2024 by dfyz Loading…
16
Remove duplicate function calls inactive-90d
#1584 opened Jun 12, 2024 by 103yiran Loading…
@d-k-b @hwu36
3
correct debug info when profiling inactive-30d inactive-90d
#1554 opened May 29, 2024 by yuukidach Loading…
2
Allow scalar broadcasting in VisitorRowBroadcast and VisitorColBroadcast feature request New feature or request inactive-90d
#1539 opened May 16, 2024 by tlrmchlsmth Loading…
25
ProTip! Filter pull requests by the default branch with base:main.