The Fury(X) Fiji Owners Thread

Gregster · 2 Jul 2015 at 11:18

AMDMatt said:
Greg, what was the name of the capture card you used on the Fury X? Did it work okay, any issues?

Avermedia LGP Lite. No performance hit but sadly only records 1080P 30 fps

Does a decent job though and very easy to use.

RavenXXX2 · 2 Jul 2015 at 11:20

Gregster said:
I am now encoding PCars from the TX and FuryX. Any other games people want to see?

That's a shame, as I wanted to compare at 1440P

GTA 5. Can you not run AB to show temps, gpu usage and the rest.

Gregster · 2 Jul 2015 at 11:22

RavenXXX2 said:
GTA 5. Can you not run AB to show temps, gpu usage and the rest.

GTA V it is. Leave it with me and I will record that and AB shows up, so no problems seeing what is what. Work at 2pm, so will need to be done tonight as well.

Besty · 2 Jul 2015 at 11:29

we need totally uncompressed screen-grabs and videos to make true comparisons.

elmarko · 2 Jul 2015 at 11:33

Good article here.

Retail AMD Fury X Sound Testing - Pump Whine Investigation

http://www.pcper.com/reviews/Graphics-Cards/Retail-AMD-Fury-X-Sound-Testing-Pump-Whine-Investigation

"Closing Thoughts

There are few things to take away from this testing. First, the currently selling and shipping AMD Radeon R9 Fury X cards do not include any kind of fix for the pump whine or sound levels of the cooler. Though AMD tells me that we just had a miscommunication or misinterpretation of the comments they shared with us prior to our NDA, I think it is at the very least fair to say that AMD was hoping to deflect the issue on launch day. Now that retail cards are out and end users (not just us) are getting their hands on them, I think its obvious that the sound issues are more of a problem than AMD would like to admit.

Second, the sound of the Fury X cards is very different than the sounds you are used to seeing associated with current flagship graphics cards. Even though the NVIDIA GTX 980 Ti reference cooler is louder than the Fury X under a gaming load, the high frequency of the sounds we are hearing on the Fury X make it more annoying and grating. Everyone's tolerance for this kind of thing is going to be slightly different, but all three people in our office agreed that we would rather hear the more "white-noise-like" sounds of air movement at a constant rate than the high pitched squeal that AMD is dealing with."

I'm pretty disappointed with AMD for the lack of communication regarding this, I'm happy to forgive an error - but they could at least hold their hands up, apologise & agree to replace/fix all cards with the issue.

Pretending it's not a problem & potentially causing people to pay restocking fees can't be worth a few extra sales, if anything a direct & honest approach would make people more comfortable purchasing (knowing a replacement/fix is no problem at all).

On a side note -
http://forums.guru3d.com/showthread.php?t=400333

"Originally Posted by Unwinder
That's actually a good thing. Assuming that AMD driver's I2C API is not completely broken and not cause a crash after using it as is, black screen is most likely a result of accessing some I2C device critical to scanning. So at least I2C is not completely locked and some device is actually being accessed."

Implying voltage unlock may indeed be likely to be possible.

Gregster · 2 Jul 2015 at 11:34

Besty said:
we need totally uncompressed screen-grabs and videos to make true comparisons.

Spot on. I am just recording and showing what is what with both maxed out. People are welcome to debate the IQ as much as they like but I am not getting involved in that. I bought the Fury X to boost my channel and said from the off that this was my intention (comparing AMD's top card to Nvidia's top card). I will do the same on the next top cards from both as well.

Gelly54 · 2 Jul 2015 at 11:35

Gregster said:
GTA V it is. Leave it with me and I will record that and AB shows up, so no problems seeing what is what. Work at 2pm, so will need to be done tonight as well.

Vram usage halved and 30c cooler? impressive.
Fps difference is huge though but you gotta admit IQ is better on the Fury...

UTmaniac · 2 Jul 2015 at 11:35

Gregster said:
I am now encoding PCars from the TX and FuryX. Any other games people want to see?

That's a shame, as I wanted to compare at 1440P

Probably too late to ask, but for pCARS, could you test...
Azure Circuit (Monaco GP)
Race date @ 1300
Formula A
20 AI
Player start from back of field

A lap with "thunderstorm" vs a lap with "medium cloud"

Let us know what "visual" and "performance" settings you use, which Windows OS you are using (much less CPU overhead in 10 preview which reduces the cpu-gpu bottleneck for me in a huge way), which AMD driver and if you are adding the "1040" DLL files to the root pCARS folder in Steam (they improve fps for me).

Oh and also what launch commands you are using, I suggest...
-pthreads 4 -dx11mt

-skipcrowds gives a boost for me, but my R7 265 is not exactly strong.

Mauller · 2 Jul 2015 at 11:36

The pump whine was fixed with shipping variants. But some of the faulty version that were sent to reviewers were also sent out to the general populous.

If you look a page or two back you can find how to tell the difference. Someone posted a picture showing which are the newer/fixed model of pump.

40Four · 2 Jul 2015 at 11:37

Mauller said:
The video is perfectly fine to pull up rendering issues as both halves were recorded at the same quality and with the same capture device.

Here is what i am talking about which you keep ignoring or just don not see.

It is a texture filtering issue on the TX. Either there is an issue with BF4 and nvidia drivers or Nvidia are purposely fiddling LOD to boost performance.

I'm not ignoring anything, my only argument is that the material provided makes it impossible to judge one way or the other In this case my preference is for the TX image (in this one particular instance) yours is for the FX i'm not saying you are wrong and i am right or vice versa. My personal opinion is that when all the variables are the same then they will both end up pretty even.

elmarko · 2 Jul 2015 at 11:38

Mauller said:
The pump whine was fixed with shipping variants. But some of the faulty version that were sent to reviewers were also sent out to the general populous.

If you look a page or two back you can find how to tell the difference. Someone posted a picture showing which are the newer/fixed model of pump.

Which needs communicating & the product lines in question recalled for fixing, not leaving the consumer to play Fury X lottery & hoping they don't get a turkey.

Twirlywoos · 2 Jul 2015 at 11:42

Fury X gets a revised pump (Again sorry if already posted)

http://www.eteknix.com/amd-revises-r9-fury-x-water-pump/?utm_source=dlvr.it&utm_medium=twitter

40Four · 2 Jul 2015 at 11:50

elmarko said:
Which needs communicating & the product lines in question recalled for fixing, not leaving the consumer to play Fury X lottery & hoping they don't get a turkey.

Has it been confirmed whether or not retailers are taking returns. I seem to recall someone earlier in the thread had a problem with returning his due to the whine. I hope it gets sorted eventually as i agree you shouldn't have to play a lottery.

Dygaza · 2 Jul 2015 at 11:52

Greg, have gpu usage showing in pcars please.

elmarko · 2 Jul 2015 at 11:53

40Four said:
Has it been confirmed whether or not retailers are taking returns. I seem to recall someone earlier in the thread had a problem with returning his due to the whine. I hope it gets sorted eventually as i agree you shouldn't have to play a lottery.

Mine is being collected today by a competitor.

I'll find out if I have any issues with the return over the new few days, I'm also finding out if they can confirm replacements are with revised pumps (if not I'll go for a refund).

Gregster · 2 Jul 2015 at 11:56

Dygaza said:
Greg, have gpu usage showing in pcars please.

Yer, I have done

Oddly, the VRAM usage on the Fury is half of what it is on the TX.... Any idea's? Memory compression perhaps?

africanos23 · 2 Jul 2015 at 12:00

Gregster said:
Yer, I have done

Oddly, the VRAM usage on the Fury is half of what it is on the TX.... Any idea's? Memory compression perhaps?

Nope, just utilizing a bigger frame buffer with the TX because it can.

LoadsaMoney · 2 Jul 2015 at 12:02

Pmsl, IQ-gate

Stanners · 2 Jul 2015 at 12:02

Mauller said:
The pump whine was fixed with shipping variants. But some of the faulty version that were sent to reviewers were also sent out to the general populous.

If you look a page or two back you can find how to tell the difference. Someone posted a picture showing which are the newer/fixed model of pump.

Sounds like more tall tales. As if this is true why are AMD not recalling the affected units. Or do they not care about customers?

humbug · 2 Jul 2015 at 12:05

Meladath said:
Really annoys me when people say nvidia cards overclock "so much better" than AMD cards, then quote the base clock to final overclocked boost clock comparison and say like "oh my god 50% overclock" Lmfao... idiots... you need to compare out of the box observed boost clock (not just what it is advertised as) with fully overclocked boost clock to see how much you are really overclocking. Hawaii isn't an amazing overclocker by any means, maxwell is better yeah, but its hardly an insanely large amount of difference, usually the MHz gain is around the same, which funnily enough, is an advantage for AMD, since each Hz gives more performance in hawaii than maxwell.

With voltage unlocking I am hoping this will be able to get pretty high. I mean I can get 1200 out of my r9 290 with 120mv, GCN 1.2 overclocks better than 1.1, so I am hoping 1300 with 200mv. 24% overclock isn't bad imo, very similar to reference 980Ti overclocking (although it does need more volts to get there... 200mv compared to 50mv).

This.

Take a maxwell card's boosting out of the box to around 1200Mhz + overclock to 1450Mhz is about 20%

A GTX 980 and R9 290X on the latest Drivers are about even @ 1440P

The average overclock on a 290X is 1150Mhz, thats 15%.

The difference is a 290X running at 1150Mhz is a 290X running at 1150Mhz, no more no less unless you have the Powerplay settings at 0 or are strangled by the CPU.

Nvidia cards boost up-to 1450Mhz they don't run at that speed at a constant like the AMD cards.