Francois Dugast af1fe65ced sycl: Increase minimum buffer size for USM system allocations (llama/25525) ...

Raise the threshold for minimum buffer size from 1 GiB to 4 GiB, based
on real-world experiments of overcommitting device memory with model
weights larger than available VRAM, for example Qwen3.5-35B-A3B-Q8
running on a B70.

Also add a debug message to better track USM system allocations.

Signed-off-by: Francois Dugast <francois.dugast@intel.com>

2026-07-30 16:31:40 +03:00

..

cmake

ggml : Parallelize quant LUT init (llama/23595)

2026-05-25 12:26:07 +03:00

include

ggml : add a set of functions for checking contiguity of inner tensor dimensions (llama/25650)

2026-07-30 16:31:39 +03:00

src

sycl: Increase minimum buffer size for USM system allocations (llama/25525)

2026-07-30 16:31:40 +03:00

.gitignore

…

CMakeLists.txt

ggml-et: Initial ET backend (llama/24179)

2026-07-30 16:31:36 +03:00