github/AMD-SHARK-Studio
1
1
Fork 0
mirror of https://github.com/nod-ai/AMD-SHARK-Studio.git synced 2026-04-03 03:00:17 -04:00
Code Issues Packages Projects Releases Wiki Activity

[SD] Change default VMA large heap block size for windows perf. (#715)

Browse Source 
Windows perform can boost from 2.67s/image to 2.4523s/image.
While Linux stays the same.
This commit is contained in:
Stanley Winata
2022-12-24 01:40:58 +07:00
committed by GitHub
parent fee4ba3746
commit 136021424c
2 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
shark/examples/shark_inference/stable_diffusion/stable_args.py
View File

@@ -137,7 +137,7 @@ p.add_argument(
p.add_argument(
"--vulkan_large_heap_block_size",
default="2147483648",
default="4147483648",
help="flag for setting VMA preferredLargeHeapBlockSize for vulkan device, default is 4G",
)

2
web/models/stable_diffusion/stable_args.py
View File

@@ -137,7 +137,7 @@ p.add_argument(
p.add_argument(
"--vulkan_large_heap_block_size",
default="2147483648",
default="4147483648",
help="flag for setting VMA preferredLargeHeapBlockSize for vulkan device, default is 4G",
)