Asus R9 290x DCII (Guru3D Review)

[stahlhart]

If I continue to see bitcoin mining derailing discussions here, it will be infracted as thread crapping. Discuss bitcoin mining in the bitcoin mining thread.
-- stahlhart

 

[blackened23]

What the... Did anyone catch this? The 290X DC II is throttling in uber mode or "performance" mode as Asus calls it. Several of the games tested showed silent mode with clocks in the 700s, while the Asus DC II 290X throttled even in performance mode? Or is my browser translation failing....

http://www.computerbase.de/artikel/grafikkarten/2013/asus-radeon-r9-290x-directcu-ii-oc-im-test/3/

EDIT: Same at canucks, apparently:

All tests were done after a 15 minute "warm up" period": (computerbase.de)

Asus DirectCU II OC R9 290X "Perf" / UBER MODE 1050 1050 (Grid 2)
1020-1050 (Battlefield 4)
1016-1050 (Metro: LL)
962-1017 (Crysis 3)
978-1017 (Bioshock)

Apparently this card is throttling in uber mode? These tests were done after a 15 minute warmup which presumably other websites have not done. VRM issue maybe? Just underscores why websites need to test Hawaii based cards after a significant warm up period to show real performance...first and second runs perform great, and apparently even with this cooler it is throttling. I would not expect any throttling on aftermarket cards. Not in uber mode. This shouldn't be happening. I don't even think reference cards are throttling in uber mode....pretty sure they're not, even though they're LOUD at 55%.

Could be the VRM cooling, 3 heatpipes, or maybe a defective card. But am I off base in thinking that throttling should never happen in uber / performance mode. I think computerbase attributed it to the cooler having 3 heatpipes or something. But AFAIK the design is the exact same as the GTX 780 DC II, and that card doesn't have issues..

 

[Silverforce11]

Hawaii is designed to throttle when load is not 100% constantly, if a scene has a dip, it will downclock appropriately until its maxed load again. This is irrespective of temps.

 

[blackened23]

Oh please. Do you really think Crysis 3 and BF4 at ultra settings isn't at 95%+ load more or less all the time. Come on.

They aren't testing Darksiders here. They aren't testing DEAD SPACE. This is crysis 3, BF4, Bioshock INFINITE, Metro: LL at ULTRA SETTINGS. These games are always 95%+. And if you look at the canucks graph, the throttling doesn't occur until after 7-8 minutes or so. So 7 minutes it doesn't throttle. Then it does throttle even in uber mode.

If your theory were even remotely correct, which it isn't, it would have uneven clockspeeds throughout the game. But it doesn't happen until 7 minutes in. Computerbase found the same thing. No throttling until after warmup....

 

[Silverforce11]

Oh please. Do you really think Crysis 3 and BF4 at ultra settings isn't at 95%+ load more or less all the time. Come on.

They aren't testing Darksiders here. They aren't testing DEAD SPACE. This is crysis 3, BF4, Bioshock INFINITE, Metro: LL at ULTRA SETTINGS. These games are always 95%+.

95% is not 99.999%

 

[blackened23]

That is perhaps the worst counter argument ever because you're basically wrong. These games don't dip in GPU usage because they're on ultra. I would expect this in a non demanding game like...dead space. Darksiders. DOTA 2. Not crysis 3 on ultra settings.

In the meantime, if there's any reasonable explanation for this, i'm curious. I'm chalking it up to Asus failing on the 290X DC II design. Probably the VRM cooling like communism mentioned.

 

[Silverforce11]

So you KNOW FOR A FACT or just spouting bs again?

Oooh look, its throttling!!

A graph without actual game context!!

http://www.pcper.com/reviews/Graphi...Graphics-Card-Review/Cooling-and-Clock-Benefi
"Well look at that, the ASUS R9 290X is quite the champ! Throughout our 25 minute session of Metro: Last Light it stays at the 1050 MHz frequency pretty much the whole time with only the occasional dip to something in the 600-700s when the level reloads."

 

[blackened23]

Are you going to get upset and make it personal again? You keep doing this. These games are always pegged at 99% for me on a single GPU when on ultra settings. I think you'll find that matches the experience of everyone else as well.

Being that you're wrong in your explanation, perhaps someone else can chime in on the real situation with the throttling here. If your theory were remotely correct, it would not take 10-15 minutes for the throttling to occur. Yet the DC II 290X doesn't throttle until it is warmed up for 10-15 minutes. For canucks, the throttling did not occur until the card was warmed up at 7 minutes. Maybe you missed that part.

 

[Imouto]

If you check their temp graphs with the perf mode reaching 94ºC you should know already that they did something really wrong. First thing I noticed is that they used the EVGA Precision tool. Second I noticed is that they said the throttling was caused by thermals when neither Guru3D, PCPer and HardwareCanucks cards surpassed the 78ºC mark.

But no, let's speculate about the card throttling. Stop this, please.

 

[blackened23]

Oh, so the card throttled in uber mode for 2 websites but not for PCPer. Wonder if PCPer has good AC in their building.

That doesn't explain why the card throttled for two different websites after a warm in period.

 

[Silverforce11]

Are you going to get upset and make it personal again? You keep doing this. These games are always pegged at 99% for me on a single GPU when on ultra settings. I think you'll find that matches the experience of everyone else as well.

Being that you're wrong in your explanation, perhaps someone else can chime in on the real situation with the throttling here. If your theory were remotely correct, it would not take 10-15 minutes for the throttling to occur. Yet the DC II 290X doesn't throttle until it is warmed up for 10-15 minutes. For canucks, the throttling did not occur until the card was warmed up at 7 minutes. Maybe you missed that part.

Now you're just being on your crusade again, the graph you quote from HC doesn't have throttling on the ASUS card on performance mode, its only throttling on silent mode. Look again carefully before you spout your BS claims.

Hint, its the RED line.

 

[blackened23]

You're correct, my mistake. By the way. STOP MAKING IT PERSONAL. If you have an issue take it to PM. Cuz. If i'm on a crusade you're really on a crusade. It's rude of you to even say that because....HELLO POT KETTLE... I'm not making it personal so you just need to stop making it personal. Not sure what your problem is. If I make an error in a post I have no problems admitting it.

 

[Silverforce11]

If you also looked at your computerbase.de review properly: http://www.computerbase.de/artikel/grafikkarten/2013/asus-radeon-r9-290x-directcu-ii-oc-im-test/2/

They have an aberrant sample, performance mode obviously not working and temps reach 94C. Their perf mode and silent mode have the same fan speed cap! If there's anything interesting, its this, an obvious bad case of QC from ASUS to send reviewers a dodgy sample.

 

[Imouto]

Yes, you have. Something really wrong happened with CB review since it's the only one out of 4 reviews reaching 94ºC in performance mode.

CB made some kind of huge mistake. In the PCPer review the card doesn't get past 1600 rpm while the CB one went past 2300 rpm.

But yeah, you don't have a personal problem with these cards. Everyone is against you and all.

 

[BallaTheFeared]

If I continue to see bitcoin mining derailing discussions here, it will be infracted as thread crapping. Discuss bitcoin mining in the bitcoin mining thread.
-- stahlhart