Quote from: GingerAle on July 26, 2015, 03:46:14 PM

Thanks for posting! This definitely helps inform my next purchase. I've been debating going the AMD route for a new monero rig. So at current release, it seems that for whatever reason the fury performs as well as the 290x.

You're very welcome!

Quote from: Claymore on July 26, 2015, 05:11:36 PM

>As i said before, with those low temps, i know this also could go a lot faster. Energy usage confirmed that.

The bottleneck of CryptoNight PoW algo is that it requires loading small data blocks from random memory addresses, a lot of times. Because addresses are random, memory cache is useless. You see low temps because GPU core does not work all the time, most time it waits data from GPU memory due its latency. And it cannot be optimized because the algo was designed to use memory latency to be asic-resistant. Fury X has 4096 bit memory bandwidth, but it is useless for this algo because big bandwidth is not related to memory latency.

It's useless for most algo's..
But again, everything was at stock settings, i have been playing around with the core engine and was able to push it to 1190 without a problem. After that it started to fail, temperatures were still very low though, if i was on windows i could play with voltage as it's possible with Afterburner.

I'm not experienced enough to optimize any algo to be Fury X friendly

But i know someone who is and will be playing with it in the next weeks.
I'm curious about the results!

Quote from: pallas on July 26, 2015, 05:29:47 PM

Quote from: Claymore on July 26, 2015, 05:11:36 PM

>As i said before, with those low temps, i know this also could go a lot faster. Energy usage confirmed that.

The bottleneck of CryptoNight PoW algo is that it requires loading small data blocks from random memory addresses, a lot of times. Because addresses are random, memory cache is useless. You see low temps because GPU core does not work all the time, most time it waits data from GPU memory due its latency. And it cannot be optimized because the algo was designed to use memory latency to be asic-resistant. Fury X has 4096 bit memory bandwidth, but it is useless for this algo because big bandwidth is not related to memory latency.

same for lyra2re (it looks like fury x performance is worse than 280x), but they are changing the algo (at least for vertcoin) in order to be more gpu-friendly, so that might change.

Lyra2re performance is worse than a 280x you mean? i haven't tested that algo "yet".. so i don't know. I ignored Cryptomining blog results because most of them aren't correct compared to my results.

Regarding Cryptonote, seeing the results with other cards (stock settings) the fury beats them all, not by much, but it's still the fastest it seems

Greetings