Repository Issue Activity (beta)

NVIDIA/cutlass

Current issue state, recent activity, and per-issue timelines from the indexed issue data.

Open Issues
430
New in 7 Days
6
Closed in 7 Days
11
Average Open Age
360 days
Stale 30+ Days
378
Stale 90+ Days
335
Last 2 Weeks
DateOpenedClosedCommentsEventsOpen Backlog
2026-08-2400000
2026-08-2310122
2026-08-2200123
2026-08-2101364
2026-08-2014171
2026-08-1901232
2026-08-1803246
2026-08-17425266
2026-08-1602134
2026-08-1500455
2026-08-1420131
2026-08-1310397
2026-08-1231001
2026-08-1111594
This Week

Opened: 2

Closed: 9

Comments: 10

Events: 24

Top Labels
? - Needs Triage (1127)
question (839)
inactive-30d (695)
inactive-90d (478)
bug (419)
CuTe DSL (205)
feature request (133)
CUTLASS C++ (76)
Issue Explorer
IssueAuthorStateLabelsCommentsReactionsUpdated

#3481 [DOC/QST] `StreamKScheduler` silently degrades to data-parallel when k-tiles per output tile <= `min_iters_per_sk_unit_` (8); with the sm103 FP4 "ultra" kernel's 768-wide K tile every K <= 6144 GEMM gets no stream-K and no warning. Plus two sm_103a build notes.

Opened 1 day ago
Infatoshi
open
No labels
001 day ago

#3406 [CuTe DSL] setmaxregister_* can be silently dropped by ptxas (C7508) — nothing surfaces to Python; min_blocks_per_mp=1 prevents it

Opened 1 month ago
zkyue
open
inactive-30d
101 day ago

#3340 Stale comment in `examples/cute/tutorial/hopper/wgmma_sm90.cu`

Opened 2 months ago
gambiTarun
open
documentation
? - Needs Triage
inactive-30d
402 days ago

#3469 [BUG] [CuTeDSL] cute.coalesce doesn't do anything on dynamic shapes

Opened 7 days ago
lw
open
bug
? - Needs Triage
CuTe DSL
1013 days ago

#3398 [Feature] cute.compile hardcodes no_cache=True, disabling the DSL's content-addressed compile cache

Opened 1 month ago
arbi-dev
open
inactive-30d
203 days ago

#3390 [QST] Public SIMT SGEMM is slower than cuBLAS on sm_89

Opened 1 month ago
Mryange
closed - completed
question
? - Needs Triage
inactive-30d
203 days ago

#3330 [QST] Why is the accumulation dtype in the CUTLASS FP4 GEMV example in FP16, instead of FP32?

Opened 2 months ago
b8zhong
open
question
? - Needs Triage
inactive-30d
203 days ago

#3182 [DOC] Wrong stride in Layout Algebra docs, composition worked example

Opened 4 months ago
tea-more
open
documentation
? - Needs Triage
inactive-30d
204 days ago

#3478 [BUG] pycute: logical_product with an integer tiler computes logical_divide

Opened 4 days ago
arcusbuilds
open
No labels
104 days ago

#2915 [FEA] Support `grid_constant` annotation in CuTe DSL

Opened 8 months ago
Gin-Sin
closed - not_planned
feature request
? - Needs Triage
inactive-30d
704 days ago

#3447 [BUG] cuteDSL export_to_c outputs i32 dynamic shapes in headers for i64 inputs

Opened 13 days ago
slayton58
closed - completed
bug
? - Needs Triage
CuTe DSL
004 days ago

#3443 [BUG][CuTe DSL] Constructing cutlass Numeric scalars from Python values costs ~0.3-0.8us each

Opened 14 days ago
reubenconducts
closed - completed
bug
? - Needs Triage
CuTe DSL
005 days ago

#3444 [BUG][CuTe DSL] TVM-FFI env stream detection fails when all GPU tensors are nested in a tuple

Opened 14 days ago
reubenconducts
closed - completed
bug
? - Needs Triage
CuTe DSL
105 days ago

#2711 [QST] [CuTeDSL] numeric_conversion?

Opened 10 months ago
HanGuo97
closed - not_planned
question
? - Needs Triage
inactive-30d
205 days ago

#3395 [BUG][CuTe DSL] Stale source after import breaks lazy jit preprocessing silently

Opened 1 month ago
yunweili3
open
bug
? - Needs Triage
inactive-30d
305 days ago

#2708 [QST]question about cutlass epilogue customization

Opened 10 months ago
Shiroha-Key
closed - not_planned
question
? - Needs Triage
inactive-30d
206 days ago

#2974 [QST] Missing tcgen05.fence::before_thread_sync

Opened 7 months ago
Edenzzzz
open
question
? - Needs Triage
406 days ago

#2833 [BUG] cute SM75 mma instruction missing

Opened 9 months ago
jinzhen-lin
closed - completed
bug
? - Needs Triage
inactive-30d
307 days ago

#3101 [BUG] ScatterD issue in predicated_tile_iterator with unguarded pointer updates in load_with_byte_offset and operator+=

Opened 5 months ago
pengpeng-yu
closed - completed
bug
? - Needs Triage
CUTLASS C++
407 days ago

#3463 [BUG] v4.7.0 regression: bf16 tile stores scalarized to `st.global.b16` (was `st.global.v4.b32` in 4.5.3)

Opened 10 days ago
joydddd
open
bug
? - Needs Triage
CuTe DSL
227 days ago

#3471 [BUG] [CuTeDSL] cute.right_inverse refuses layouts with dynamic sizes

Opened 7 days ago
lw
open
bug
? - Needs Triage
CuTe DSL
107 days ago

#3470 [BUG] [CuTeDSL] cute.composition throws when rhs has a dynamic shape

Opened 7 days ago
lw
open
bug
? - Needs Triage
CuTe DSL
107 days ago

#3472 [BUG] Installed headers return CMake / configure_file placeholders for version numbers

Opened 7 days ago
BillyONeal
open
bug
? - Needs Triage
CUTLASS C++
007 days ago

#3452 CuTeDSL kernel name prefix leaks into later compilations

Opened 12 days ago
drisspg
closed - completed
No labels
327 days ago

#2947 [BUG] [Blackwell] Enable FP4/tcgen05 support for sm_121 (DGX Spark) in CuTe DSL

Opened 8 months ago
Teora
closed - completed
bug
? - Needs Triage
CuTe DSL
907 days ago

Rows per page:

1–25 of 1,491