Posts: 413
Threads: 2
Joined: Dec 2015
(09-23-2018, 02:36 PM)malakudi Wrote: Can you please run a --benchmark-all test? Can you also report running gpu frequency during this test?
https://gist.github.com/epixoip/23068f4b...e7751522f2
Benchmarks for everything are available there, running clocks are a bit variable on the release day drivers, instant peaks were up to 1980MHz but after a little while of running, you will see them fall to be about 1815-1840MHz as things heat up and everything levels out. There is PLENTY of room for overclocking from playing with it a bit, but that always runs the risk of instability so for testing I don't have any offsets for the clock speed, it's boosting on it's own.
Posts: 12
Threads: 2
Joined: Dec 2017
(09-23-2018, 02:50 PM)Chick3nman Wrote: (09-23-2018, 02:36 PM)malakudi Wrote: Can you please run a --benchmark-all test? Can you also report running gpu frequency during this test?
https://gist.github.com/epixoip/23068f4b...e7751522f2
Benchmarks for everything are available there, running clocks are a bit variable on the release day drivers, instant peaks were up to 1980MHz but after a little while of running, you will see them fall to be about 1815-1840MHz as things heat up and everything levels out. There is PLENTY of room for overclocking from playing with it a bit, but that always runs the risk of instability so for testing I don't have any offsets for the clock speed, it's boosting on it's own.
Great, thank you. FYI, Linux driver does exist, it is 410.57 and available in nvidia site. Maybe it wasn't available three days ago.
Posts: 413
Threads: 2
Joined: Dec 2015
(09-23-2018, 03:02 PM)malakudi Wrote: Great, thank you. FYI, Linux driver does exist, it is 410.57 and available in nvidia site. Maybe it wasn't available three days ago.
It was indeed not available from the Nvidia site on launch day, not sure if that was intentional by Nvidia or accidental but I can see it now, thank you!
Posts: 12
Threads: 2
Joined: Dec 2017
And since we collect performance details to compare with latest hashcat 4.2.1, here is my run on GTX 1070 Ti.
https://gist.github.com/malakudi/ac7ede9...173a118eb8
Posts: 381
Threads: 1
Joined: Aug 2014
Thanks malakudi. Which driver did you use? FE-version? Stock clock?
Posts: 12
Threads: 2
Joined: Dec 2017
09-23-2018, 05:03 PM
(This post was last modified: 09-23-2018, 05:06 PM by malakudi.)
It is a Gigabyte Windforce, clocks are stock - no preapplied OC allowed from nvidia in 1070 ti and I didn't overclock for this test - although I usually run at +180 MHz.
edit: driver is 390.87 in Linux.
Posts: 194
Threads: 7
Joined: Jul 2015
Flomac, I'd be weary using the benchmarks I provided on 411.63. They are anywhere between 5 and 10% slower than my benchmark on 398.82, depending on the algorithm.
Posts: 381
Threads: 1
Joined: Aug 2014
Same here with the GTX980. NVidia 397.xx fine, 411.62 slow. I don't wanna sound paranoid, but it's odd that with the new driver and cards the older ones suddenly get slower.
I'm just preparing the comparison with the old driver results.
Posts: 194
Threads: 7
Joined: Jul 2015
Paranoid as in they're intentionally nerfing older cards? Hmmmm, would be the first that I've personally seen by Nvidia but certainly not a first in the field. I'll see about testing with my GTX 980 later as well. It's currently in my Ubuntu 18 box.
Posts: 10
Threads: 0
Joined: Aug 2015
Sounds like Apple wrote the drivers.
