zlacker

[return to "AMD funded a drop-in CUDA implementation built on ROCm: It's now open-source"]
◧
1. sam_go+16[view] [source] 2024-02-12 14:34:23
>>mfigui+(OP)
I don't really follow this, but isn't it a bad sign for ROCm that, for example, ZLUDA + Blender 4's CUDA back-end delivers better performance than the native Radeon HIP back-end?
◧◩
2. Kepler+ta[view] [source] 2024-02-12 14:59:41
>>sam_go+16
Surely this can be attributed to Blender's HIP code just being suboptimal because nobody really cares about it. By extension nobody cares about it because performance is suboptimal.

It's AMDs job to break that circle.

[go to top]