Morph MORPH
SIGN IN SIGN UP
dhruvhead / tilelang UNCLAIMED
0 0 0 C++

COMMITS

main
July 30, 2026
C
[Refactor] Remove intrinsic compatibility facade (#2812)
Chaofan Lin committed
50481cc
C
[BugFix] Reject unsupported fast-math input dtypes (#2804)
cj committed
21e8c06
K
[BugFix] Resolve partial scalar reduce barrier participation (#2777)
KellyFrog committed
0bc1913
R
[BugFix] Reject uncovered warp partitions in T.gemm instead of producing silently wrong results (#2724)
Ruinan Li committed
5b1f321
C
[CUDA] Extend the GEMM FMA fallback to SM75 (#2811)
Chennes committed
7fd9536
July 29, 2026
C
[Refactor] Remove unused tilelang.common package (#2810)
Chaofan Lin committed
dd92b78
C
[BugFix] Skip descriptor TMA for device-bound copy bases (#2803)
Chaofan Lin committed
6c3dd97
J
[BugFix] Fix warp_reduce truncating int64/uint64 to 32 bits on sm_80+ (#2782)
jpwang committed
e41fadb
C
[BugFix] Reject mixed packed x2 operand dtypes (#2802)
cj committed
2a06036
T
[Fix] Refine architecture guards (#2790)
Tong WU committed
32e02e6
M
[BugFix] Prevent autotuner cache reuse across different outputs and validation settings (#2793)
morluto committed
bc9515f
M
[BugFix] Check shared-TMEM buffer pointer types before dereference (#2794)
morluto committed
2c84f4f
G
[Metal] Add line-level threadgroup qualifier scanning (pass 5) (#2796)
Gengyuan Bai committed
28f7033
C
[BugFix] Reject floating-point predicates in vote intrinsics (#2800)
cj committed
51f88a8
C
[BugFix] Preserve alloc_var initializer dtype (#2801)
cj committed
b1b605d
Y
[CUDA] Support arbitrary TMEM layouts (#2785)
Yongqi Zhuo committed
1cb4d4f
July 28, 2026
C
[FFI] Support apache-tvm-ffi 0.1.12 (#2795)
Chaofan Lin committed
4086ba8
L
[BugFix] Pass buffer row stride to 2D scan kernel to fix silent miscomputation (#2620)
Liu Yunuo committed
7ec5adb
L
[TIR][Language] Add typed vector lane extraction API (#2789)
Lei Wang committed
6171343
Y
[Metal] M5 Cooperative Tensor T.gemm (#2252)
Yichen Yan committed
1545f00
G
[Metal] Add 16-byte alignment padding to shared/threadgroup memory (#2786)
Gengyuan Bai committed
aaf68d2
B
[BugFix] Honor nan_propagate in reduce max/min/absmax clear=False write-back (#2788)
BoHao Chen committed
940b106
P
[TIR] Inject source spans into tirx IR and surface source locations in compiler errors (#2751)
penguin_wwy committed
a42bbc3
L
[CUDA][Transform] Fix PCWS index dtype handling (#2783)
Lei Wang committed
500c368
P
[BugFix] Preserve loop step when unrolling loops (#2784)
penguin_wwy committed
eb31994
July 27, 2026
W
[BugFix] Fallback non-16B cluster bulk copies (#2683)
wcx committed
51fbfc7
T
[Enhancement] More compile-time guards for architecture-specific CUDA intrinsics (#2781)
Tong WU committed
b5e3eb9
C
[TileOP] Add SM70 GEMM FMA fallback (#2339)
cklxx committed
2a17fff
J
[BugFix] Fix fp16/bf16 `T.atomic_max`/`atomic_min` silently corrupting fp32 values (#2780)
jpwang committed
aa7df86
C
[BugFix] Fix ROCm intrinsic resolution after language dialect refactor (#2779)
Chaofan Lin committed
9fb7572