Ivan Avdeev 755b2d59a9 vk: fixup BLAS preallocation on Linux/amdgpu ... Fixing pool allocator to properly signal allocation failure uncovered an existing issue where we were lacking enough memory for dynamic model BLASes on Linux/amdgpu. Erroneously the same memory region was used for >1 BLAS. Surprisingly this hasn't led to any noticeable issues so far. Increasing accels buffer size fixes the issue.		2023-09-14 13:57:59 -04:00
..
gl	engine: client: implement v_dark in engine, instead of making it ref-specific	2023-04-28 16:52:22 +03:00
soft	ref: add size argument to COM_FileBase calls	2023-04-26 05:20:45 +03:00
vk	vk: fixup BLAS preallocation on Linux/amdgpu	2023-09-14 13:57:59 -04:00