NVIDIA GeForce GTX 1080 On Linux: OpenGL, OpenCL, Vulkan Performance

[Michael]

I'm sure Michael would be very happy if AMD sent out free samples of more recent cards to benchmark. He can use what he has, can't keep up with all the rebranding and batch moving to legacy... Feel free to contribute.

True I would be happy, but it's unfortunately not Bridgman's decision who gets cards or not... Obviously I know if Bridgman could, he would be able to get cards for Phoronix, but he's been a tremendous help nevertheless. Hopefully for the RX 480 will be able to get sampled. But yes for the AMD side I was including basically the cards I had available and are basically recent when ignoring all the re-branding / slight changing. There's Fury plus Rx 200-nearly-same-as-300.

[artivision]

True I would be happy, but it's unfortunately not Bridgman's decision who gets cards or not... Obviously I know if Bridgman could, he would be able to get cards for Phoronix, but he's been a tremendous help nevertheless. Hopefully for the RX 480 will be able to get sampled. But yes for the AMD side I was including basically the cards I had available and are basically recent when ignoring all the re-branding / slight changing. There's Fury plus Rx 200-nearly-same-as-300.

It would be nice if Bridgman sends two Gpus (one middle range and one high end) per real generation (GCN-1-2-3-4). Please Mr. Bridgman, we are many that we use RadeonSi (no matter what) now.

[PsynoKhi0]

https://www.youtube.com/watch?v=myDYnofz_JE

TL;DR: this card is only good for wetting your pants running tests on open benches, as long as you leave tweaked/factory OC older cards out of the equation.

[ResponseWriter]

Would like to have seen this benchmark compared with the GPU-PRO driver from AMD in order to get the OpenCL and Vulkan comparisons. It's not like those are impossible to attain. Otherwise a comparison of radeonsi vs nouveau would be nice if we're going to insist on using only open drivers.

Edit: please, avoid saying 'freaking' next time, too, otherwise I find it hard to take you seriously.

[cj.wijtmans]

yeah at this point it would be really nice to see AMD cards with the best performance drivers. 1080 is too pricey... And also you would probably see AMD keeps its performance at higher resolutions while nvidia will drop.

[LeJimster]

Very impressed with the power efficiency with the 16nm chip. Hopefully AMD's cards will have equal or better efficiency. I would love an Nvidia card for the performance, but I'm hoping the improvements on the AMD side, especially with the mesa drivers continues and improves enough to keep me on team red.

[efikkan]

But one question, did you do the benchmarks in an open or closed rig?

Open.

With the ever more aggressive boost and throttling, there can be a significant difference between open and closed rigs. As you can see in this test, the card is not boosting as much in a closed case. This could easily result in a 5-10 % performance difference, or more. I know this will require a different test procedure, but it also makes the benchmarks more realistic.

[bridgman]

It would be nice if Bridgman sends two Gpus (one middle range and one high end) per real generation (GCN-1-2-3-4). Please Mr. Bridgman, we are many that we use RadeonSi (no matter what) now.

As Michael said, I have absolutely nothing to do with choosing where sample cards go other than periodically encouraging our PR folks to send them.

[wizard69]

Some how I was expecting more!

[efikkan]

Some how I was expecting more!

Granted, the hype was exaggerating quite a bit, but it's still a decent improvement over GTX 980.

If you were expecting a new high-end card pushing way beyond GTX 980 Ti though, you'll have to wait until GP102 arrives. It is scheduled for the end of 2016.