llama.cpp
f549b000 - vulkan: Call ggml_vk_buffer_write_2d from ggml_vk_buffer_copy (#16793)

Commit

50 days ago

vulkan: Call ggml_vk_buffer_write_2d from ggml_vk_buffer_copy (#16793) This lets the copy to the destination device use the host-visible vidmem optimization.

References

#16793 - vulkan: Call ggml_vk_buffer_write_2d from ggml_vk_buffer_copy

Author

jeffbolznv

Parents

9a3ea685

llama.cpp f549b000 - vulkan: Call ggml_vk_buffer_write_2d from ggml_vk_buffer_copy (#16793)

llama.cpp
f549b000 - vulkan: Call ggml_vk_buffer_write_2d from ggml_vk_buffer_copy (#16793)