Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 7 Sep 2020 10:14:01 +0200
From: Lukas Odzioba <lukas.odzioba@...il.com>
To: john-users@...ts.openwall.com
Subject: Re: NVIDIA GeForce RTX 3000 Series Performance

pon., 7 wrz 2020 o 08:57 Anatoly Pugachev <matorola@...il.com> napisaƂ(a):
> some of the cuda performance stats released
> https://videocardz.com/newz/nvidia-geforce-rtx-3080-168-of-rtx-2080-performance-in-cuda-and-opencl-benchmarks

It is hard to guesstimate from it. During hashing we care about
integer performance and a lot of those numbers are based on floating
point operations, which find the most use in 3d graphics - that's why
you can find a theoretical number of "flops" of a given GPU quite
easily, but not a number of integer operations per second.

BTW 3090 doesn't seem to make sense because for the same price you can
go double 3080, or triple 3070.
And as most of the Internet speculates 3000 should be pretty decent
because it jumps from 12nm to 8nm and includes architectural changes
at the same time.
By the end of September, there should be some benchmarks available to
answer your question without speculating.
Personally I'd wait for BigNavi, competition is usually good for your wallet.

Thanks,
Lukas

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.