AMD Fiji Thread

elmarko · 22 Jun 2015 at 09:03

Personally noise is only a decider between equally performing cards. If one pulls ahead in performance & max stable overclock I'll go with the noise.

I will pay a small premium for silent & cool, but only usually as a byproduct of higher expected overclocking potential.

Mtom · 22 Jun 2015 at 09:13

Rossi~ said:
In what sense? Price or sales?
Do you have figures to show this?

In quality and effort...290x with no work done just slapped on 780 cooler with no vrm or vram cooler, heatpipes dont meet with the gpu..then same crappy cooler on the matrix platinum....
A lot of the not so great motherboards sporting the ROG sticker now, with a price premium...

I'm not saying asus don't make good stuff too, but lately they have a lot of crappy pieces as well.

On the other hand Gigabyte really upped their game from the not so good WF3 cooler to the new ones, a pushing out cards like the G1.

cactusweasel · 22 Jun 2015 at 09:14

Gigabyet have stepped up with Z170 also, the new boards from them look amazing, RoG is same-old-same-old.

Mtom · 22 Jun 2015 at 09:18

Silent_Scone said:
I think there is a high and likely chance this stuff degrades fast, we discussed this in another thread. Might well see it unlocked at some point though.

Why should it degrades fast? I don't see why would it degrade by running it at say 60C instead of 40C.

Gibbo · 22 Jun 2015 at 09:18

flopper said:
HBM has enough bandwidth so OC dont yeild results one can expect as its new tech.
Individually cards will have different OC reaches as usual and Gibbo kinda should know that by now.
Problem for nvidia is they are behind both with software as the new standard Dx12 is Mantle AMD developed tech along with the new HBM memory. No way out for them as Fury wins however they try.

Of course I know it, but I can only share the results I get with the card given. I do believe however AMD are right, overclocking the HBM is simply not required as there is more than enough bandwidth.

I would like to hit 1200-1300MHz core with some voltage though as performance then could be very interesting.

Gibbo · 22 Jun 2015 at 09:20

Hi there

I did a test last night, I closed up my case but left the radiator just in the bottom of the case exhausting its heat inside the case, this is a really bad idea by the way, but I wanted to see how hot the card would get.

After running Heaven 4.0 all night, the maximum temperature was 58c and the fan speed never exceeded 16%, starting point percentage for the fan is 15%, so it was quieter than the case fans and I was able to sleep with no issues at all.

Here is a picture, its in a Corsair case with the smoked side panels, so its not easiest to see and ignore the SSD's just randomly hanging in the bottom of case:

Gibbo · 22 Jun 2015 at 09:23

elmarko said:
The cooling seems fine in the Fury X, amazing from what Gibbo is saying - it's just how much we can push out of that cooling with unlocked voltage is the question.

The only thing that concerns me is, why if this is the case isn't the voltage & therefore the core clock set higher by default if it is possible & if not - why not?. I've had a fair few gigabyte windforce cards in the past & all of them have overclocked amazingly well, run very quiet & had no issues with - motherboards on the other hand

They have TDP to hit, at stock it is 275W I believe which is pretty good.

If they increase voltage and clock speed then 275W is out the window, just like a 980Ti at 1.250-1.300mv with 1500/8000 is consuming well over 400W.

Also if manufacturers released cards with huge overclocks but then customers could gain very little more they would all moan.

You can never please the customer.

ALXAndy · 22 Jun 2015 at 09:23

What sort of overclock have you managed sans the voltage control?

elmarko · 22 Jun 2015 at 09:24

That's crazy cool, how cool does it run at full usage at 100% & how noisy is it?.

Also do you have the total power usage at max overclock at the moment (let us know how much 'headroom' we have), this should give us a very rough idea of overclocking potential if unlocked.

Gibbo · 22 Jun 2015 at 09:25

ALXAndy said:
What sort of overclock have you managed sans the voltage control?

I am not going to report my overclocks without voltage yet as stating Mhz will probably get my hands slapped.

But mine is not doing 1200MHz on stock voltage.

Gibbo · 22 Jun 2015 at 09:27

elmarko said:
That's crazy cool, how cool down it run at full usage at 100% & how noisy is it?.

Also do you have the total power usage at max overclock, this should give us a very rough idea of overclocking potential if unlocked.

At 100% fan it is a strong whooshing noise but at 100% fan the under-load temperatures never exceed 10c above ambient and normally a lot less. So if say your case temp is around 26c, with 100% fan you will be under 35c loaded.

But 100% fan is not really required on this kind of cooler, optimal fan speed for this cooler is around 40-50% which is audible but not crazy loud, I find the fan becomes audible over case fans at 25-30% but the fan never really needs to exceed 15% for regular gaming at stock which is as quiet as graphics cards get.

elmarko · 22 Jun 2015 at 09:30

Hmm, interesting.

Btw, Gibbo - your posts are already across a few websites (other overclocking ones) wouldn't want to get your wrists slapped by asking too much

.

I guess the decider for me will be, if they unlock the voltage & if the card can handle more voltage. From the looks of it it seems to be under-clocked/powered to be running that cool..

Gibbo · 22 Jun 2015 at 09:33

elmarko said:
Hmm, interesting.

Btw, Gibbo - your posts are already across a few websites (other overclocking ones) wouldn't want to get your wrists slapped by asking too much .

I guess the decider for me will be, if they unlock the voltage & if the card can handle more voltage. From the looks of it it seems to be under-clocked/powered to be running that cool..

If AMD are pushing the voltages as low as possible to remain cool and low TDP, then it will depend if the technology responds well to voltage.

In theory it will do and as there is at a least a good 20-30c left on the table before throttling occurs (75c) then I would suspect even just a little extra voltage will have quite a positive impact to clock speeds.

Mtom · 22 Jun 2015 at 09:33

Gibbo said:
Of course I know it, but I can only share the results I get with the card given. I do believe however AMD are right, overclocking the HBM is simply not required as there is more than enough bandwidth.

I would like to hit 1200-1300MHz core with some voltage though as performance then could be very interesting.

And is there a possibility for that by the conference call with the AMD engineers?

Superpatsy · 22 Jun 2015 at 09:35

Gibbo said:
Hi there

I did a test last night, I closed up my case but left the radiator just in the bottom of the case exhausting its heat inside the case, this is a really bad idea by the way, but I wanted to see how hot the card would get.

After running Heaven 4.0 all night, the maximum temperature was 58c and the fan speed never exceeded 16%, starting point percentage for the fan is 15%, so it was quieter than the case fans and I was able to sleep with no issues at all.

Here is a picture, its in a Corsair case with the smoked side panels, so its not easiest to see and ignore the SSD's just randomly hanging in the bottom of case

What resolution was that?

flopper · 22 Jun 2015 at 09:37

2 days to go
we need a countdown

mabbutts · 22 Jun 2015 at 09:40

This is making me so indecisive!

Do I cancel my weekends Zotac order and wait for the Fury x.... or just leave it and potentially regret my purchase!

Decisions decisions!

ubersonic · 22 Jun 2015 at 09:40

Gibbo said:
I did a test last night, I closed up my case but left the radiator just in the bottom of the case exhausting its heat inside the case, this is a really bad idea by the way, but I wanted to see how hot the card would get.

Sorry that pics a bit dark, do you mean you had the rad on a bottom fan mount drawing air in from outside or that it was just loose in the case?

elmarko · 22 Jun 2015 at 09:43

It's almost it's a shame it's an new chip, as we can't extrapolate based on the stock voltage what kind of limits are on the chip for increasing. Sadly from the sounds of it the benchmarks we will see in a few days won't give the full picture.

That is, unless AMD confirm the situation regarding voltage unlocking. Based on the presentation saying the card will be an 'overclockers dream' then I'd hope they mean more than 200mhz on the core max (when we have cards pushing 4/500 on the core at the moment in direct competition).

I'd guess.

Stock 1000/1050 = Matching stock TI (beating some games/losing others).
1200 over-clock = Matching/beating most overclocked 980ti - losing to G1.
1300+ over-clock= potentially beating 980ti G1.

A nice 1450 w/ additional volts would be the kind of performance required to really give Nvidia a black eye, I'd very much like this card to wipe the floor to shake things up a bit & keep Nvidia on the game.

Silent_Scone · 22 Jun 2015 at 09:45

Mtom said:
Why should it degrades fast? I don't see why would it degrade by running it at say 60C instead of 40C.

Hi,

Temperature is only one factor. HBM is very low voltage. It's current that kills and memory ic as compacted as this is likely to not handle voltage well, may even be dead spots in stablity by doing this.

Competitor rules

**AMD Fiji Thread**

AMD Fiji Thread