Photo Upscaling Times with Different GPU's

Topaz Photo AI General
1

I’m no expert, but have a basic general knowledge of GPU cores. My question is this:

I have been using a Geekom A7 AMD Ryzen 7940HS 32Gb with a Radeon 780M 12 core integrated GPU for Topaz photo upscaling.

I recently purchased a MacBook Air M3 with the 10 core graphic option and 16Gb. So decided to upscale the same RAW file by a 4X factor on both platforms, to compare processing times. With no other tasks or applications running on either machine.

I expected the Geekom with Radeon to crush the Mac M3. Yet the file took 20+ minutes to create on the Radeon system, vs 4 minutes for the MacBook Air M3!

Would this point to Topaz not correctly or optimally utilising all the 12 Radeon cores? And is there a way to tweak/adjust this?

2

Probably less to do with the GPU and more to do with use of the Neural Engine on the M series chips.

1 Like
3

Thank you, I hadn’t thought of that aspect.

4

Update - I’ve solved the mystery. I generally output files directly to Onedrive. For some reason the new Mac is fine with that, but not the Geekom A7.

When I changed the path in Topaz on the Ryzen to save directly to the local C drive, et voila! Processing time was really speedy and beat the M3 as I would have expected. Hope this helps others.

1 Like
5

Yes. :slightly_smiling_face:

6

Does the Mac not get a speed boost if you save just locally?