Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

ggml: add env var GGML_OP_OFFLOAD_MIN_BATCH (#18535)

Browse Source 
* ggml: add env var GGML_OP_OFFLOAD_MIN_BATCH
* makes the min_batch_size for triggering op offload configurable via env var, defaulting to the prior hardcoded value of 32

* ggml: read GGML_OP_OFFLOAD_MIN_BATCH once and store to dev ctx

* cann: forward declaration of device context struct

* cann: move offload op check after device context declaration

* cuda: fix whitespace

Co-authored-by: Aman Gupta <amangupta052@gmail.com>

---------

Co-authored-by: Aman Gupta <amangupta052@gmail.com>
This commit is contained in:
Doctor Shotgun
2026-01-08 11:03:21 +02:00
committed by GitHub
co-authored by Aman Gupta
parent 9c142e3a2a
commit 9a5724dee2
7 changed files with 45 additions and 38 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-metal/ggml-metal-device.m
+2
View File
@@ -782,6 +782,8 @@ ggml_metal_device_t ggml_metal_device_init(void) {
dev->props.supports_gpu_family_apple7 = [dev->mtl_device supportsFamily:MTLGPUFamilyApple7];
dev->props.op_offload_min_batch_size = getenv("GGML_OP_OFFLOAD_MIN_BATCH") ? atoi(getenv("GGML_OP_OFFLOAD_MIN_BATCH")) : 32;
dev->props.max_buffer_size = dev->mtl_device.maxBufferLength;
dev->props.max_working_set_size = dev->mtl_device.recommendedMaxWorkingSetSize;
dev->props.max_theadgroup_memory_size = dev->mtl_device.maxThreadgroupMemoryLength;