Does Radeon R9 290X Behave Any Differently In A Closed Case?

[Mousemonkey]

the heat is on! remember when some one fry a egg on geforce now is time to AMD

IIRC the egg didn't cook very well, unless you consider raw and runny whites as "cooked".

 

[FormatC]

IIRC the egg didn't cook very well, unless you consider raw and runny whites as "cooked".

With a Geforce:

With a Radeon:

And finally a hot dog:

The complete kitchen story:
http://www.tomshardware.de/HD6990-GTX590-Corsair-Obsidian-800D,testberichte-240761-9.html

 

[Mousemonkey]

IIRC the egg didn't cook very well, unless you consider raw and runny whites as "cooked".

With a Geforce:

With a Radeon:

And finally a hot dog:

The complete kitchen story:
http://www.tomshardware.de/HD6990-GTX590-Corsair-Obsidian-800D,testberichte-240761-9.html

Is there an English version of that?

 

[FormatC]

Nope, sorry. Try Google translator please, this is an older story from me...
(But funny. See se heating benchmark to warm up one of my rooms)

 

[Mousemonkey]

In this video that was posted at the time the egg didn't cook.

[flash=420,315]<iframe width="640" height="360" src="//www.youtube.com/embed/ASu3Xw6JM1w?feature=player_detailpage" frameborder="0" allowfullscreen></iframe>[/flash]

 

[FormatC]

You need more than 60°C (and a lot of time) to make a good egg (and a small and very thin bowl with a with smooth underbody).
But this works only with this older cards. The HD 6990 with OC BIOS has a power consumption up to 450 Watts, the oc'ed GTX 590 was a lot slower and lost the race

 

[Mousemonkey]

I noticed in the pictures that you looked as though you were using the back of the card to do your cooking, as that is always going to be hotter that the front of the HSF I wonder how may other cards that would work on?

 

[FormatC]

This works only on the backside. I've tested a lot of cards and only this two cards were working well

 

[Mousemonkey]

This works only on the backside. I've tested a lot of cards and only this two cards were working well

Is cooking an egg part of the normal benchmarking suite then? :lol:

 

[FormatC]

This is too energy expensive... I have to pay here 0.28 EUR per kW/h

 

[Haravikk]

Actually, I seem to remember reading somewhere that for cooling purposes you want airflow to be turbulent, which is something you can do best inside a case. However, the problem with that is that many quiet fans produce "smooth" airflow in order to reduce noise, so while they might produce respectable airflow values, the quality of airflow may not be sufficient to cool effectively.So yeah, any GPU should be capable of running *better* in a case, but it will depend greatly on the cooling efficiency of your case and cooling systems.

 

[FormatC]

If a card can hold the maximum clock rate in quiet mode, why I have to use the so-called performance mode to kill my ears? This mode was created by AMD due the bad performance of the reference cooler. The board partners of AMD are absolutely free to play with this or to design something other. I have benchmarked the default mode, not a special quiete mode and that means: the card as is.

Gigabyte calls the default mode "advantage" mode and the other mode is nearly unusable. Nervous fan controller, a lot of ups and downs. May be this is for OC but the voltages are completely the same. You get a higher power target and a lot of noise, that's all.

 

[Cilverjolt]

i would still like a test in performance mode

 

[FormatC]

i would still like a test in performance mode

If the cards are running below the temperature target the so-called "performance"- or uber-mode is not faster. This relic from AMDs reference card is obsolete for this better cooling solutions. All custom cards are more or less well-balanced. I see no reason to produce a lot of noise only to play a game 0.5 fps faster (if...)

 

[somebodyspecial]

So even 290x OC cards still come up 10% short of 780ti OC OOBE, and you can still OC the 780ti OC. Nothing changed, and we now know there is no magical aftermarket solution that will make a 290x beat a 780ti OC even out of the box (never mind what you can OC that card to also). I didn't think anything would change anyway but after a few months of waiting this is now over IMHO. NV chips are just better this time. I hope AMD does better at 20nm. You don't make money when it takes 8 months to catch the other guy (with 3 SMX's off) and even then all the enemy does is release the REAL version to dominate again. Ouch.

 

[Mousemonkey]

So even 290x OC cards still come up 10% short of 780ti OC OOBE, and you can still OC the 780ti OC. Nothing changed, and we now know there is no magical aftermarket solution that will make a 290x beat a 780ti OC even out of the box (never mind what you can OC that card to also). I didn't think anything would change anyway but after a few months of waiting this is now over IMHO. NV chips are just better this time. I hope AMD does better at 20nm. You don't make money when it takes 8 months to catch the other guy (with 3 SMX's off) and even then all the enemy does is release the REAL version to dominate again. Ouch.

http://www.fudzilla.com/home/item/25712-nvidia-not-impressed-with-hd-7900-series

 

[kevin1981]

PowerTune/PowerPlay... These cards sound like they share some of the issues seen with the 7870 LE's. People needed to manually bump the voltage to get them to perform as advertised or fiddle with card bios flashing. I had to do it with my Powercolor.

 

[all stalked out]

What I'm coming away from this article with is that the Asus DCII 290x is not as good as I thought it was and the Sapphire tri-x looks better than it's cheap look credits it with, I owned more Sapphire cards from the 7000 series than any other brand and really liked them but took an immediate dislike to the look of these R9 models. I'm glad I never acted on my first impressions.

 

[Cyne17]

Can you do a test of these cards in crossfire close to each other? I have an issue on temps for my mATX maximus VI gene. Temp reaches 89C for windforce result in a downclock to 727MHZ. For single card I have no issue at all.

 

[happycamperjack]

Litecoin mining has really gave me a whole new perspective on graphic card heatsink design. Running 3 7970(different brand and heatsinks) in a closed case is incredibly hard to keep all the cards running cool..... unless you have reference design! Yes, AMD's reference design is actually the best for a closed case packed with graphic cards. Why? It's design to force airflow out of the case instead of dissipating them inside the case. As long as you have a positive airflow in the case, AMD's reference design performed incredibly well in crossfire as I overclock the card(sandwich between two other 7970 as they would crush if I put them in middle) from 925 to 1035 and memory from 1350 to 1750 while keeping the fan running at 70% and 75c. And we are talking about running at this setting for over a month now without it crushing (other 2 7970 do crush sometimes)In conclusion, if you are considering crossfiring your AMD cards in a closed case, GET REFERENCE HEATSINK!

 

[nikoli707]

hmmm, i thought mining rigs want to keep cpu and gpu temps around 72-75c if they can for efficiency, even blocking vents with towels to do so.

 

[Jared88]

Why compare 3 cards built by the same company? ASUS and Gigabyte = Same. Just so you know. Of course it does make write ups easier I suppose.

 

[FormatC]

Asus isn't Gigabyte. Write this in Taiwan and they will kill you

This are two completely different companies with own HQs, R&D and factories. I was many times in asia and visited a lot of HQs and factories by myself. And for your interest: they are also using different OEMs for their coolers. ASRock and Asus is another story.