R0CKSTAR f872b59112 cuda: guard the iq4_nl dequantize row kernel against short rows (#29683) ...

dequantize_block_iq4_nl writes QK_K values per block, but a row can be shorter than that (an IQ4_NL row is only guaranteed to be a multiple of QK4_NL). Threads whose 32-value sub-block starts at or past k currently read and write past the end of the row. Skip those sub-blocks; for rows that are a multiple of QK_K the check never fires.

2026-09-30 22:33:22 +02:00

..

cmake

Enables Windows ARM64 build with MSVC cl.exe (#28362)

2026-09-28 10:07:27 +02:00

include

ci: add zdnn backend build but not test (#29541)

2026-09-29 20:45:14 +03:00

src

cuda: guard the iq4_nl dequantize row kernel against short rows (#29683)

2026-09-30 22:33:22 +02:00

.gitignore

…

CMakeLists.txt

sync : ggml (#29396)

2026-09-24 22:44:05 +03:00