Jesse Gross 392a270261 ggml: Avoid cudaMemsetAsync during memory fitting ... We pass invalid pointers when we check the size of the required compute graph before fitting. Some CUDA APIs validate these pointers but we can just skip them during this phase. cudaMemsetAsync is one of these that we weren't skipping but never took the code path that used it. Now that we have enabled op_offload, we can hit it in memory pressured situations.		2025-10-31 15:23:28 -07:00
..
backend	ggml: Avoid cudaMemsetAsync during memory fitting	2025-10-31 15:23:28 -07:00
nn	interleaved mrope (#12807)	2025-10-30 11:29:00 -07:00
backend.go	ggml: Enable op_offload to improve partial offload performance	2025-10-30 13:53:10 -07:00
device.go	cpu: always ensure LibOllamaPath included (#12890)	2025-10-31 14:37:29 -07:00
path.go	cpu: always ensure LibOllamaPath included (#12890)	2025-10-31 14:37:29 -07:00