HWZone Forum - Sapphire R9 Fury X pics and benchmarks!

[therealnickdanger]

http://forums.hardwarezone.com.sg/h...iew]-sapphire-amd-r9-fury-x-rise-5087633.html

 

[Head1985]

Only 30% faster than 290x i dont get it.
fury X have + 46% more SP and more memory bandwidth.It should be like 45-50% faster.
Its ROP limited?

64Rops was mistake :ninja:

 

[Greenlepricon]

Within 1fps of Anandtechs 980ti bench. Not too shabby!

 

[Stuka87]

Not sure we should trust these benchmarks without him comparing it to another GPU in he same exact system.

 

[Hitman928]

Not sure we should trust these benchmarks without him comparing it to another GPU in he same exact system.

Yep, compare his results to HWC and you get a different story:

 

[Head1985]

He also have 980Ti and he uses game benchmark
FuryX is on same level as 980TI
http://forums.hardwarezone.com.sg/94717850-post94.html

Anandtech results
http://www.anandtech.com/show/9306/the-nvidia-geforce-gtx-980-ti-review/6

Furyx is 30% faster than 290x

 

[Hitman928]

He also have 980Ti and he uses game benchmark
FuryX is on same level as 980TI
http://forums.hardwarezone.com.sg/94717850-post94.html

Edit: Looking at the thread, I don't think he has a 980Ti, I think that result is the Fury X with FXAA turned off. He then uses that result in comparison to the 980 Ti from the link he quotes which shows them performing the same. We'll still need to wait for a more direct comparison, but good stuff initially.

 

[Subyman]

Later in the thread he OCs and can't manage much more over stock. The speculation is it is TDP limited, so maybe a bios mod would help. Was hoping for great OC headroom with that AIO.

 

[Jaydip]

I just compared the results to 980Ti, they are damn close, 1-2 fps difference between them.Is Fury-X is the highest tier or do we have another one above it? if it is the best at this moment then various 980Ti non reference will be faster than Fury-X.I am interested between a oc vs oc showdown between them, also even if it is tied the 980Ti Ultra can still take the performance crown.I think when they went for wc they should have pushed the clocks a little bit higher.I am waiting for the final review results before buying it though.

 

[Hitman928]

With a 6.7% oc he was able to get 6.6% more performance in SoM. From what I understand, the Fury uses a new voltage boost system that is more card dependent to save on power. Hopefully voltages/power adjustments will be able to be made, otherwise it doesn't seem to get much past stock.

 

[Face2Face]

Later in the thread he OCs and can't manage much more over stock. The speculation is it is TDP limited, so maybe a bios mod would help. Was hoping for great OC headroom with that AIO.

That's with no voltage adjustment though.

 

[Udgnim]

I just compared the results to 980Ti, they are damn close, 1-2 fps difference between them.Is Fury-X is the highest tier or do we have another one above it? if it is the best at this moment then various 980Ti non reference will be faster than Fury-X.I am interested between a oc vs oc showdown between them, also even if it is tied the 980Ti Ultra can still take the performance crown.I think when they went for wc they should have pushed the clocks a little bit higher.I am waiting for the final review results before buying it though.

1050 clock is what fits with AMD's 275W TDP target

only thing higher for AMD is going to be their dual GPU card, but performance advantage will be hit & miss due to crossfire profile reliance / drivers

assuming stock Fury X performance is similar to stock GTX 980Ti, the biggest selling point of Fury X is that it's going to have competitive stock performance while running cooler & producing less noise. I think 980Ti is going to have better OC headroom than Fury X.

can also try to BS that Fury X is still immature and future driver updates will eke out increased performance

 

[Jaydip]

1050 clock is what fits with AMD's 275W TDP target

only thing higher for AMD is going to be their dual GPU card, but performance advantage will be hit & miss due to crossfire profile reliance / drivers

assuming stock Fury X performance is similar to stock GTX 980Ti, the biggest selling point of Fury X is that it's going to have competitive stock performance while running cooler & producing less noise. I think 980Ti is going to have better OC headroom than Fury X.

can also try to BS that Fury X is still immature and future driver updates will eke out increased performance

Valid points but since it has two 8+8 pins they should have set it a little higher.I want to be impressed with this card but matching a 980Ti performance at this moment is really not mind blowing imo.

 

[exar333]

1050 clock is what fits with AMD's 275W TDP target

only thing higher for AMD is going to be their dual GPU card, but performance advantage will be hit & miss due to crossfire profile reliance / drivers

assuming stock Fury X performance is similar to stock GTX 980Ti, the biggest selling point of Fury X is that it's going to have competitive stock performance while running cooler & producing less noise. I think 980Ti is going to have better OC headroom than Fury X.

can also try to BS that Fury X is still immature and future driver updates will eke out increased performance

This doesn't give us much more to go on though. Hoping reviews tomorrow help us understand 4K performance, CF scaling and OC headroom. The last 2 are the most important to me, but the 4GB mem will play into the 4K performance the most. HBM is another wild-card here too...

 

[flopper]

take my money amd just take it

 

[DownTheSky]

AMD cards don't OC well because AMD's transistors are tighter packed and less optimized for speed. Also the cards are clocked closer to their limits.

If r9 290X and Fiji were nVidia releases, the clocks would be something like 800mhz with boost up to 1000.

 

[Stuka87]

AMD cards don't OC well because AMD's transistors are tighter packed and less optimized for speed.

Source?

 

[LTC8K6]

Source?

"They"

They said it. :biggrin:

 

[Despoiler]

Source?

"They"

They said it. :biggrin:

Don't even need to source it. You heard it here first people. Fiji = Hawaii.

 

[utahraptor]

The reviewer's guide itself only showed a 100 MHz overclock being obtainable. Signs point to the HBM memory frequency being locked.

 

[LTC8K6]

The reviewer's guide itself only showed a 100 MHz overclock being obtainable. Signs point to the HBM memory frequency being locked.

They were probably not releasing accurate OC info, I'd think.

 

[Stuka87]

The reviewer's guide itself only showed a 100 MHz overclock being obtainable. Signs point to the HBM memory frequency being locked.

GPU frequency has nothing to do with the HBM frequency. And everything points to this person not having the proper tooling.

 

[iSeeker]

http://techfrag.com/2015/06/23/amd-radeon-r9-fury-x-official-benchmark-results-released/
Fury X with 4.7%-5.7% OC performance?
Doesn't seem a lot.

 

[Subyman]

http://techfrag.com/2015/06/23/amd-radeon-r9-fury-x-official-benchmark-results-released/
Fury X with 4.7%-5.7% OC performance?
Doesn't seem a lot.

Old chart. That's been going around for the last few days.

 

[iSeeker]

I know, but I though the graph showing OC at +100MHz was just a start.
I read from new user benches that +100MHz could be the limit?
http://forums.hardwarezone.com.sg/hardware-clinic-2/%5Bgpu-review%5D-sapphire-amd-r9-fury-x-rise-5087633-10.html