Hi guys
With the latest bench, (vray-benchmark-5.00.03), the CUDA results for the 3090 comes in at roughly 2080 to 2311
However, the A6000 with double ram and more cuda, tensor and RT cores, comes in at only 1668
That seems way off to me…
Why do you think that is @daniel.taskoff ?
Hi, Morne,
you’re right, that there’s a big difference comparing the maximum scores achieved by those configurations, however, comparing the averages for the 3090, and the A6000, the gap is almost closed (1570 vs 1505), which is very consistent with all the other benchmarks out there. I guess the RTX 3090 was clocked higher than the A6000, causing the larger gap for the maximum scores.
Although A6000 is said to perform better in memory-hungry situations (which is not the case with the benchmark scenes), its memory is slower than that of the RTX 3090, which, I think, explains the results.
Best regards,
Daniel
Ah, I didnt notice the “Avg” dropdown in the results. Makes much more sense now
Thanks