Benchmarks for hashcat v4.00 on AMD and NVIDIA cards

[epixoip]

Nikos, you were already banned once, why are you returning to spout the same AMD fanboy nonsense that got you banned in the first place?

when I first got my 2 vega 64's I posted a benchmark that looked really good, just like the one atom posted, however, in practice the card was actually about 40% slower than the benchmark speed.

I have stated this over and over and over. These cards rely extremely heavily on firmware to throttle the GPU down to keep power and heat within range. Benchmarks are bursty so the firmware doesn't have much of an opportunity to throttle, but when you start using the card for real workloads, PowerTune rears its ugly head.

Consumer Volta cards will be released in March 2018, this was confirmed in September. The Tesla V100 shows just how much a beast the Volta line is.

Strongly doubt we'll see Navi in 2018, but even if we do, I'm sure it won't even meet a fraction of expectations.

[Flomac]

Maybe it's time for a new kind of benchmark. A dozen hashes that can be cracked with a wordlist and some rule set that powers out a GPU at 100%. In the commen formats MD5, SHA-1 and maybe NTLMv2 or somthing equal. The benchmark could offer a runtime of ~15min for a GTX1080Ti.

atom, if you're interested, I could create some hashes that would fit those specs.

[Nikos]

Nikos, you were already banned once, why are you returning to spout the same AMD fanboy nonsense that got you banned in the first place?

Hello.

I don't know really get you regarding "AMD fanboy nonsense" and "already banned once" but is that you that you were insisting on GTX 1080 being faster than VEGA in hashcat ?

I remember reading an Nvidia hardcore fanboy writing posts like yours a few months ago.

[epixoip]

Sigh. The GTX 1080 is faster than Vega with Hashcat. The performance is exactly as I said it would be months before the cards were even released. I said it will likely benchmark better because benchmarks are bursty, but it would fall apart in real-world performance because of PowerTune. This is precisely what users are seeing: real world performance that is 40% of benchmark speeds, nowhere near the speed of the GTX 1080, let alone the 1080 Ti.

As someone who sells GPUs for password cracking for a living, don't you think I would be selling AMD GPUs if they were better? There's a reason I abandoned AMD and ran far, far away.

[Flomac]

I remember reading an Nvidia hardcore fanboy writing posts like yours a few months ago.

There are no fanboys here, only people with experience