You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I was wondering if there are plans to support 3D convolutions in the next releases. The optimization of these operations can be highly beneficial for 3D image processing, for example in medical applications. When using AMD Genoa and 2D convolutions, the performances of ZenDNN are almost on par with those obtained using one A100 GPU. With 3D convolutions instead, ZenDNN fallbacks to the mklDNN kernel, which performs quite badly on the CPU, and timings are around 20x slower than the CUDA ops.
The text was updated successfully, but these errors were encountered:
I was wondering if there are plans to support 3D convolutions in the next releases. The optimization of these operations can be highly beneficial for 3D image processing, for example in medical applications. When using AMD Genoa and 2D convolutions, the performances of ZenDNN are almost on par with those obtained using one A100 GPU. With 3D convolutions instead, ZenDNN fallbacks to the mklDNN kernel, which performs quite badly on the CPU, and timings are around 20x slower than the CUDA ops.
The text was updated successfully, but these errors were encountered: