2 ms·
Actually GPU latencies and transfer costs dominate most simple image filtering operations. A GPU might be fast once the data is in GPU local RAM, but if just tr
by vardump 11y ago
Actually GPU latencies and transfer costs dominate most simple image filtering operations. A GPU might be fast once the data is in GPU local RAM, but if just transferring the data back and forth takes 3x the time running the filter locally, what's the point?
You're better off performing it on a CPU, of course SIMD optimized.
Besides, web servers probably won't have GPUs anyways.
- mozumder 11y agoIntel is building entire Xeon lines with embedded GPUs for serving web imagery. And these Intel Xeon's with embedded GPUs don't need to transfer data to the coprocessor, since they operate off of system memory, so transfer latencies are a non-issue.