llama.cpp
ef93e98d - vulkan: Fix Windows performance regression on Intel GPU BF16 workloads for Xe2 and newer (#22461)

Commit

1 day ago

vulkan: Fix Windows performance regression on Intel GPU BF16 workloads for Xe2 and newer (#22461) * refactor * Use l_warptile only when coopamt is available for BF16

References

#22461 - vulkan: Fix Windows performance regression on Intel GPU BF16 workloads for Xe2 and newer

Author

rillomas

Parents

706fbd8a

llama.cpp ef93e98d - vulkan: Fix Windows performance regression on Intel GPU BF16 workloads for Xe2 and newer (#22461)

llama.cpp
ef93e98d - vulkan: Fix Windows performance regression on Intel GPU BF16 workloads for Xe2 and newer (#22461)