Lei Wang
efce11ed25
[CUDA][Codegen] Keep RNG state in kernel scope ( #2540 )
...
Fix CUDA RNG state scope
2026-07-07 14:41:03 +08:00
Zhengju Tang
9ddf577042
[BugFix] Stride check and fix for tensors with zero-stride argument ( #1749 )
...
* [BugFix] Stride check and fix for empty tensor in tvm_ffi
* Lint
* Lint
* Revert
* Lint
* Fix zero stride check in C host code
* Add warning for zero stride
* Modify test scripts
* Lint
* Rename test
* Refactor
* simplify
* Lint
* Fix
2026-01-29 23:59:24 +08:00
silentCoder-dev
7cfd087aea
[Feature] Add more curand operations & support vectorization ( #1582 )
...
* add more curand operations & support vectorization
* fix typo about bitwidth
* merge random float into a single operation
2026-01-06 02:20:01 +08:00
silentCoder-dev
95e3b5a716
[Refactor] Remove triton dependence in testing & move triton baseline into examples ( #1470 )
...
* remove triton dependence in testing & move triton baseline into example
* use ceildiv and handles arbitrary M correctly for triton
2025-12-19 12:34:39 +08:00
silentCoder-dev
f06726031e
[Refactor] Rename test for curand & add triton baseline in test_tilelang_language_rand.py ( #1464 )
...
* rename test for curand & add triton baseline
* add a comment for calling T.rng_rand() four times
* refactor tilelang&triton kernel
* Add boundary checks for M not divisible by 128
2025-12-19 11:08:53 +08:00
Lei Wang