980Ti faster than Fury X in DX12 API OVERHEAD TEST 1.2

[StereoPixel]

Fury X = ~14 554 928
980Ti = 17 451 744 (DX12 draw calls per second)
http://www.3dmark.com/3dm/8173627

 

[alcoholbob]

290X also scored higher than a Titan X in that test so I wouldn't consider it indicative of GPU performance.

 

[ShintaiDK]

Its just frontend and a useless test. Same reason why an OCed GTX960 can beat a GTX980 in it.

 

[frozentundra123456]

I agree, please stop with the worthless API overhead tests.

 

[DustinBrowder]

These simple basic tests are not representative of real performance. I don't even think we'll see decent DX12 gains for at least a year.

Entire engines have to be completely written from the ground up for DX12 to take maximum advantage and then games built on those engine/s need to be optimized to run smooth on a wide range of hardware, including hardware that doesn't support DX12.

 

[bystander36]

That's like 100's of times more calls a second than needed, so it's not going to matter much beyond a certain point.

 

[Silverforce11]

That's like 100's of times more calls a second than needed, so it's not going to matter much beyond a certain point.

Maybe now, but if they keep pushing realism, hairs, grass, dust particles, would make for ridiculous jumps in draw calls.

Diminishing returns but what else are we gonna use the GPU grunt for?

 

[selni]

Its just frontend and a useless test. Same reason why an OCed GTX960 can beat a GTX980 in it.

This, please understand what this test actually is - it's measuring the maximum rate of do almost nothing draw calls. It's just a synthetic benchmark for showing cpu+api overhead or command processor limits.

In no way does it represent a meaningful performance metric for comparing anything but API overhead on the same card as the number of draw calls that actually do something you can use will be far lower and unrelated to this.

 

[Nhirlathothep]

next test we re going to compare their fan color and light opacity

 

[bystander36]

Maybe now, but if they keep pushing realism, hairs, grass, dust particles, would make for ridiculous jumps in draw calls.

Diminishing returns but what else are we gonna use the GPU grunt for?

That's like well over a 1000 times what we use now. For a while at least, it's 100's of times more than what we use now. By the time we need that much, hardware will be very different than now.

 

[railven]

That's like well over a 1000 times what we use now. For a while at least, it's 100's of times more than what we use now. By the time we need that much, hardware will be very different than now.

Makes me think of this:

https://www.youtube.com/watch?v=nUBVvk-JUwQ

 

[Kenmitch]

Makes me think of this:

https://www.youtube.com/watch?v=nUBVvk-JUwQ

Dang....Who would of thought AMD could last that long....No nvidia logos plastered everywhere!

 

[railven]

Dang....Who would of thought AMD could last that long....No nvidia logos plastered everywhere!

AMD? I don't even think Sony would last that long. haha.

 

[wege12]

What are the odds that the Fury X will one day outperform a 980ti due to future software/driver optimizations?

 

[Glo.]

What are the odds that the Fury X will one day outperform a 980ti due to future software/driver optimizations?

In some games it already does.
http://www.computerbase.de/2015-07/...ndigkeit/#diagramm-ryse-son-of-rome-1920-1080

We have seen this situation already. R9 290X vs 780 Ti. In 2014 Geforce was faster. Right now in most of the cases(excluding Gameworks) we see that 290X is faster.

Also in their test of driver overhead it plainly shows that Fury X it does much more draw calls than GTX 980 Ti.

 

[RobertR1]

What are the odds that the Fury X will one day outperform a 980ti due to future software/driver optimizations?

The question will become if/when that day does come are either card still relevant for their intended purpose?

Most people who buy high end GPU's tend to like to playing newer games at high settings and will generally upgrade their GPU's often to keep up.

Banking on future performance should be derived by your length of intended ownership. Even then there is no guarantee. With Furyx and 980ti being around the same price, it'd be very tough to recommend a furyx to anyone, objectively speaking.

 

[railven]

The question will become if/when that day does come are either card still relevant for their intended purpose?

Most people who buy high end GPU's tend to like to playing newer games at high settings and will generally upgrade their GPU's often to keep up.

Banking on future performance should be derived by your length of intended ownership. Even then there is no guarantee. With Furyx and 980ti being around the same price, it'd be very tough to recommend a furyx to anyone, objectively speaking.

I already got my 14nm money ready! Question is, who gets it next round?

GF gets my 980 Ti with water cooler Kit and I get a shiny new GPU to go with my shiny new rig.

(Unless we pick a house and the new mortgage wrecks havoc on my plans ).

 

[chimaxi83]

I couldn't find this game on Steam. I'm pretty sure I spelled it right, I searched for "DX12 API OVERHEAD TEST 1.2". Is it in beta or something??

 

[ShintaiDK]

Its part of 3DMark.

 

[Sabrewings]

Its part of 3DMark.

The paid version, mind you.

 

[soccerballtux]

The question will become if/when that day does come are either card still relevant for their intended purpose?

Most people who buy high end GPU's tend to like to playing newer games at high settings and will generally upgrade their GPU's often to keep up.

Banking on future performance should be derived by your length of intended ownership. Even then there is no guarantee. With Furyx and 980ti being around the same price, it'd be very tough to recommend a furyx to anyone, objectively speaking.

most people grow up.

I still play the same games e.g. tribes ascend. when ut4 comes out I'll get a top of the line AMD because I'll be using it brand new

 

[tential]

What are the odds that the Fury X will one day outperform a 980ti due to future software/driver optimizations?

Pretty good. The question is, what type of high end buyer are you. If yo uare buying a $650 graphics card to hold onto it for 5+ years, you can go back to one of RS's many posts about how that is a bad idea.

If you're a high end buyer who likely has a high turnover rate of cards? Then it won't matter. By the time Fury X outperforms the 980Ti, majority of owners of both cards will have moved to node shrink cards and sold off their cards to second hand buyers (or keep them in their collection or whatever).

I don't see my friends with SLI Titan X worrying about if the Fury X will outperform it one day. They're waiting to purchase Pascal already.

At the high end, people usually want a card that will perform for now, not for 3 years down the road when by then they've sold off the card and moved on to a card that is far faster....

 

[Sabrewings]

They're waiting to purchase Pascal already.

I'm right there with them. By the time a Fury X is consistently faster than my 980 Ti, it'll just be a glimmer in the rear view mirror.

 

[alcoholbob]

Fury X will be faster than 980Ti the day 980Ti is a equal to a $150 20fps card.

 

[tential]

I'm right there with them. By the time a Fury X is consistently faster than my 980 Ti, it'll just be a glimmer in the rear view mirror.

I honestly used to think it was a pretty bad idea to purchase a high end card and then sell it and repeat. But now that I think of it, I've paid yearly for things like xbox live and cable TV. As long as you can sell it off... you're really just enjoying the best performance all the time.

I need to start selling my cards while they're still valuable lol.

Edit: I'm curious if both above posters who have "luxury" (by my standards anyway) gpus, if you don't mind me asking do you sell your GPUs or do you keep them? I still have old GPUs lying around and CPUs. I've never thrown out or sold any lol...