[Benchmark] GeForce RTX 5090
#8
(02-11-2025, 02:08 AM)Chick3nman Wrote: I've posted full benchmarks for the 5090 here: https://gist.github.com/Chick3nman/09bac...c1e1363d5c

I see "Your device driver installation is probably broken. See also: https://hashcat.net/faq/wrongdriver" multiple times in that benchmark, e.g. in TrueCrypt and VeraCrypt modes. And the values are a lot higher than for 4090, e.g. Mode 6232 192.2 kH/s (4090) vs. 539.4 kH/s (5090).
Does something went wrong? Or are there modules working correctly and are the number real?

I know that Blackwell has all INT32/FP32 cores while in Ada half the cores are fixed FP32. May these be the only modes benefiting from the architectural change, because all other modes are already limiting by other factors? For example server cards have more cores than the consumer variants but they (server cards) have lower power limit and thus lower clocks and a lower performance in most cases (in my experience).
Reply


Messages In This Thread
[Benchmark] GeForce RTX 5090 - by afd - 01-28-2025, 09:29 AM
RE: [Benchmark] GeForce RTX 5090 - by Chick3nman - 01-28-2025, 11:10 PM
RE: [Benchmark] GeForce RTX 5090 - by malakudi - 02-04-2025, 10:49 AM
RE: [Benchmark] GeForce RTX 5090 - by Chick3nman - 02-04-2025, 09:45 PM
RE: [Benchmark] GeForce RTX 5090 - by Chick3nman - 02-11-2025, 02:08 AM
RE: [Benchmark] GeForce RTX 5090 - by BotPass - 02-13-2025, 03:08 PM
RE: [Benchmark] GeForce RTX 5090 - by wgmmmx - 02-12-2025, 01:05 AM
RE: [Benchmark] GeForce RTX 5090 - by Chick3nman - 02-12-2025, 09:17 PM