ASUS R9 290X DirectCU II OC Overclocking Review @ [H]

FrgMstr

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,596
ASUS R9 290X DirectCU II OC Overclocking Review - We will take the ASUS R9 290X DC2 OC custom AMD R9 290X based video card and for the first time see how well the 290X can overclock. We will also for the first time compare it to an overclocked GeForce GTX 780 Ti video card head-to-head and see who wins when overclocking is accounted for.
 
780ti is a beast. It'll be interesting to see how BF4 shapes up once the Mantle patch is released.
 
Good review.

Did overclocking the 780 Ti allow for any settings changes in the games?
That wasn't addressed in this article or any of your 780Ti articles.

It'd be nice if you could get Asus to send you the DirectCU II OC version of the 780Ti for review.
That'd be a nice apples-to-apples comparison.
 
gathagan said:
Good review.

Did overclocking the 780 Ti allow for any settings changes in the games?
That wasn't addressed in this article or any of your 780Ti articles.

It'd be nice if you could get Asus to send you the DirectCU II OC version of the 780Ti for review.
That'd be a nice apples-to-apples comparison.
Click to expand...

Battlefield 4 - It was closer to being playable at 4X, but we'd still settle on 2X AA to maintain the high performance gamers demand. At 4X AA it would drop just a bit too low during heavy smoke/fire/explosions still. So no change.

Crysis 3 - The next available increase in image quality is to enable "Very High" system spec. Unfortunately that is still utterly too demanding at 1600p, even with the overclock. So no change.

Far Cry 3 - This one was playable at a higher setting than shown. You can play at 4X AA at Ultra settings with the overclock. Note though that 4X AA is also playable with the card at stock settings - http://www.hardocp.com/article/2013/11/07/nvidia_geforce_gtx_780_ti_video_card_review/4 So in the end, no real change, just higher performance. 8X MSAA is just out of range.

Tomb Raider - There is no where else to go in Tomb Raider but to enable SSAA, and that is entirely too demanding. You'd need SLI to achieve that. So no change.
 
Interesting...

But really, both those cards are overpriced.
 
I love when these reviews come out, if only to watch the fallout on other websites **coughOCNcough*** where all the angry posters claim nobody but them knows how to overclock.
 
Nice write up, very honest.
Did you record any VRM temps during standard clocks and overclocks?
 
I don't think it is fair testing these cards at different fan speeds....the only way to fairly test them is with both with 100% fans regardless of noise.

You say there was no throttling as you kept temps below 94c....that is debatable, power tune alters settings in fractions of a second and if you looked at this microscopically you may well find that throttling is occurring.

Secondly your selection of a lower fan speed for the AMD card may well be affecting the maximum overclock you are achieving....I would say this is undoubtedly happening as you are hitting 90c....higher silicon temps mean greater leakage which means more voltage is required to be stable which means more heat...you have basically surpassed the ability of the cooling solutions effectiveness.....its a vicious circle which you have deliberately brought about in the AMD card not only by using less than 100% fan but also using a slower fan speed than the Nvidia card you are comparing it to.....how can you ever justify that when making a "fair" comparison.
 
technogiant said:
I don't think it is fair testing these cards at different fan speeds....the only way to fairly test them is with both with 100% fans regardless of noise.
Click to expand...
They do realistic testing of how the cards will be used.

You say there was no throttling as you kept temps below 94c....that is debatable, power tune alters settings in fractions of a second and if you looked at this microscopically you may well find that throttling is occurring.
Click to expand...
If it doesnt register any throttling, there is very little to worry about.
When it registers zero throttling as the average, are you saying that it is clocking higher than the maximum set clock to balance the average?

Secondly your selection of a lower fan speed for the AMD card may well be affecting the maximum overclock you are achieving....I would say this is undoubtedly happening as you are hitting 90c....higher silicon temps mean greater leakage which means more voltage is required to be stable which means more heat...you have basically surpassed the ability of the cooling solutions effectiveness.....its a vicious circle which you have deliberately brought about in the AMD card not only by using less than 100% fan but also using a slower fan speed than the Nvidia card you are comparing it to.....how can you ever justify that when making a "fair" comparison.
Click to expand...
It would help if you understood these cards.
The design temperature for these cards is 95C, this max temp can be configured. It was at maximum allowed during their testing.
The power target was also at maximum. They demonstrated that this allows them to hit 600W from the wall at over 1.4V.
Which of those 2 would have been doing the throttling during the gaming tests?

If you read the review, you would see the report on how the coolers are not effective enough for higher voltage.
Why are you repeating this?
 
technogiant said:
I don't think it is fair testing these cards at different fan speeds....the only way to fairly test them is with both with 100% fans regardless of noise.

You say there was no throttling as you kept temps below 94c....that is debatable, power tune alters settings in fractions of a second and if you looked at this microscopically you may well find that throttling is occurring.

Secondly your selection of a lower fan speed for the AMD card may well be affecting the maximum overclock you are achieving....I would say this is undoubtedly happening as you are hitting 90c....higher silicon temps mean greater leakage which means more voltage is required to be stable which means more heat...you have basically surpassed the ability of the cooling solutions effectiveness.....its a vicious circle which you have deliberately brought about in the AMD card not only by using less than 100% fan but also using a slower fan speed than the Nvidia card you are comparing it to.....how can you ever justify that when making a "fair" comparison.
Click to expand...

So you're claiming that comparing a reference 780Ti with a single the fan @3405RPM to a 290X DCII with two bigger and more efficient fans @2929RPM is unfair... towards the Radeon? :confused:

The review is really well done. This is the kind of balancing I'm personally after: sustainable max OC on air. The particular chip in the 290X is below average for sure, but 1163MHz for a 780Ti ain't stellar either.

The real and universal truth which the article shows nicely is this: Hawaii is a power eating monster. It's a good chip, but it is pushed to the edge at stock and going even further throws all balance to the wind.

IMHO, Hawaii is either a high-end value card for the non-OC crowd (when at or near MSRP) or the high res multi card watercooling monster for the OC crowd. Anything in between and GK110 usually comes out on top.
 
Interesting indeed.

If I am understanding correctly the stock cooled 780Ti is beating the custom cooled 290X by a fair margin in BF4 & C4 when both are overclocked and after the "warm up period" when the 780ti is throttling back to 941MHz.

Once the 780Ti has decent custom cooling it is going to trounce the 290X even more as it will be quieter and cooler (so will throttle back less if at all). Other 290X custom coolers are not likely to improve much on the Asus customer cooler.

So on air it is a bit of a no brainer 780Ti all the way (especially given the pricing, in the UK for instance Palit780Ti £510 vs. 290X MSI Gaming £470).

The next review should be to evaluate both GPU's under water.
 
So...

I built my machine (in sig) in mid-December. I was trying to hold out for the non-reference 290X cards but I'm glad I stopped waiting and went with the 780Ti setup now (this review basically reinforces that opinion).

I had feeling there wasn't that much more in the 290X because of the temps/power issues and it seems this is true. Even with a heavy OC it seems the 290X is only yielding a couple of extra frames and OC vs. OC the Ti still wins out.

Out of interest, could someone tell me what the memory bandwidth is on the 780Ti at the 7.6GHz memory clock? How does that compare to the bandwidth of the 290X at 5.67GHz?
 
Yeah, they are close to the edge.
My 290 running as a 290x with an Xtreme III cooler and +100mV on the core (I cant set explicit voltages with the bios I currently use) gets 1150MHz, somewhere around 75C, I cant remember.
(default voltage reads as 1.18 to 1.2V under load)
I havent pushed it further because it aces everything I've played at the moment at stock 290x speeds.
 
Syphon Filter said:
So...

I built my machine (in sig) in mid-December. I was trying to hold out for the non-reference 290X cards but I'm glad I stopped waiting and went with the 780Ti setup now (this review basically reinforces that opinion).

I had feeling there wasn't that much more in the 290X because of the temps/power issues and it seems this is true. Even with a heavy OC it seems the 290X is only yielding a couple of extra frames and OC vs. OC the Ti still wins out.

Out of interest, could someone tell me what the memory bandwidth is on the 780Ti at the 7.6GHz memory clock? How does that compare to the bandwidth of the 290X at 5.67GHz?
Click to expand...

From the stock 780Ti numbers: 7.6GHz/7GHz * 336 GB/sec = 365 GB/sec.

From the stock 290X numbers: 5.67GHz/5GHz * 320 GB/sec = 363 GB/sec.
 
Its sad to see 290/290X hold up by their drivers, no performance driver yet from launch. I'm sure the cards would be neck to neck otherwise.

Thanks Brent and Kyle for the great review, and for including vrm temps, that's great, please keep that methodology in your tests!
 
psyside said:
Its sad to see 290/290X hold up by their drivers, no performance driver yet from launch. I'm sure the cards would be neck to neck otherwise.

Thanks Brent and Kyle for the great review, and for including vrm temps, that's great, please keep that methodology in your tests!
Click to expand...

Missed the vrm temps, thanks for pointing that out.
 
psyside said:
Its sad to see 290/290X hold up by their drivers, no performance driver yet from launch. I'm sure the cards would be neck to neck otherwise.

Thanks Brent and Kyle for the great review, and for including vrm temps, that's great, please keep that methodology in your tests!
Click to expand...

That's an interesting point. The 290X is a new architecture isn't it? One assumes there will be some (significant?) gains to be had from driver updates whereas the 780Ti is quite mature now.
 
Syphon Filter said:
That's an interesting point. The 290X is a new architecture isn't it? One assumes there will be some (significant?) gains to be had from driver updates whereas the 780Ti is quite mature now.
Click to expand...

GCN is GCN. As an architecture, it's more mature than Kepler.
 
This is GCN 2.0 though.

Are you suggesting there will be minimal gains to be had from driver updates?
 
Doesent matter if its similar or not, point being, no performance driver is out yet one way or another.

The 780 Ti had like 2/3 if i'm not mistaken, one of them which brought like 15% + increase in BF4.
 
Nenu said:
They do realistic testing of how the cards will be used.

If it doesnt register any throttling, there is very little to worry about.
When it registers zero throttling as the average, are you saying that it is clocking higher than the maximum set clock to balance the average?
Click to expand...

I did say it was debatable and you may well be right there.....but using monitoring software that samples values at a set rate when the underlying values being measured are more rapidly varying can lead to erroneous results......but granted should tend to average so long as the sample size is statistically large enough.

Nenu said:
It would help if you understood these cards.
The design temperature for these cards is 95C, this max temp can be configured. It was at maximum allowed during their testing.
The power target was also at maximum. They demonstrated that this allows them to hit 600W from the wall at over 1.4V.
Which of those 2 would have been doing the throttling during the gaming tests?

If you read the review, you would see the report on how the coolers are not effective enough for higher voltage.
Why are you repeating this?
Click to expand...

I believe I do understand these cards, but its more important to understand the the relationship between silicon temperature and power draw.

For a given set voltage the power draw of a chip increases almost logarithmically with temperature due to increased silicon leakage. Increased power draw leads to even higher temps and the requirement for even more voltage to stabilize the chip.

It makes no difference that the chip has been manufactured to be-able to work at these temps without breaking.....these effects will still occur.

So given the logarithmic slope of the power draw curve its very easy to manufacture a situation where by thermals will fly out of control.....by doing something like using a lower % fan speed on one card when compared to another....Oh yeah just like has happened in this review.....did the author explain why he used a lower % fan speed on the AMD card compared to the Nvidia card?......there by manufacturing the results they wished to publish.

For what its worth when tinkering with effectiveness of the cooling solution as has happened here is removed from the equation by water cooling then the 290X out performs 780ti
 
What would be the point in reviewing the cards at fan-speeds that no (sane) person would ever use in the real world unless they wanted their ears to bleed out?
 
technogiant said:
For what its worth when tinkering with effectiveness of the cooling solution as has happened here is removed from the equation by water cooling then the 290X out performs 780ti
Click to expand...

You state the obvious then draw the wrong conclusion from it.
I'm using a better cooler than the 780ti on my card running as a 290x and I cant beat an overclocked 780ti.

did the author explain why he used a lower % fan speed on the AMD card compared to the Nvidia card?
Click to expand...
Yes he did :rolleyes:
 
^^.....I see no reason or logic in comparing two cards when you use unequal settings.....if you are not going to use 100% fans on both...which you should in a performance test....this is a performance test not an acoustic test isn't it?....then at the very least use the same fan speeds on both cards, and no the author did not explain why he used a lower fan speed on the 290x than on the 780ti.....only why he didn't use 100% fan.

Running the 290x at 90c and the 780ti at 75c is ridiculous....the lower temp will obviously allow for a higher stable overclock....as I stated, whether you accept it or not, if adequate cooling is not an issue 290x outperforms 780ti
 
Why is running the 290x at 90c ridiculous? Lol. The gpu has been designed to run at that temperature. 90c does not hinder the card's performance at all. Did you even read the review or did you just skim it, saw a conclusion you did not like, and come here to complain?
 
technogiant said:
^^.....I see no reason or logic in comparing two cards when you use unequal settings.....if you are not going to use 100% fans on both...which you should in a performance test....this is a performance test not an acoustic test isn't it?....then at the very least use the same fan speeds on both cards, and no the author did not explain why he used a lower fan speed on the 290x than on the 780ti.....only why he didn't use 100% fan.

Running the 290x at 90c and the 780ti at 75c is ridiculous....the lower temp will obviously allow for a higher stable overclock....as I stated, whether you accept it or not, if adequate cooling is not an issue 290x outperforms 780ti
Click to expand...

Consider that the heatsink and fan designs are different.
Therefore they produce different sounds and different levels of sound at the same rpm.
You cannot equate one fan speeds usability with the other when they are both too loud way before maximum speed.
The design temperatures of the cards are similarly not comparable, the 780ti would be damaged at 95C continuous.
If you arent happy with the way it is reviewed here, there are plenty of other websites.
 
Nenu said:
Consider that the heatsink and fan designs are different.
Therefore they produce different sounds and different levels of sound at the same rpm.
You cannot equate one fan speeds usability with the other when they are both too loud way before maximum speed.
The design temperatures of the cards are similarly not comparable, the 780ti would be damaged at 95C continuous.
If you arent happy with the way it is reviewed here, there are plenty of other websites.
Click to expand...

And given what you say about acoustics there is no measurement of that here?....that would at least be one equal factor which you could hang your hat on and say that is fair.

But no there is no measurement of that to show they are equal and in fact no measurement of what level the author has decided is "reasonable".....it just happens conveniently to be a level where the 290x is running at 90c....so sorry...this is far too subjective and qualitative rather than quantitative to be considered a fair review....but I'll hang around this site just to point out the flaws and bias that some of you seem so happy to swallow..
 
So, the goal being to run both cards at fan levels that keep their respective temperatures below their respective throttle points, at sound levels that are tolerable is bias?
 
Syphon Filter said:
This is GCN 2.0 though.

Are you suggesting there will be minimal gains to be had from driver updates?
Click to expand...

Not minimal, but nothing like what we had with 7970 and 7950. I'd say there probably is a bit more left in the tank for Hawaii than there is for Kepler, but only because AMD tends to be slower in getting there with drivers. The oddball in all of this is Mantle, of course. As I see it, AMD is outsourcing some of their driver development to game developers with Mantle. That is an additional bonus for the red team to go with the straight up technical advantage of Mantle vs. DX.
 
Croak said:
So, the goal being to run both cards at fan levels that keep their respective temperatures below their respective throttle points, at sound levels that are tolerable is bias?
Click to expand...

Exactly. This is a well balanced review to allow readers to draw conculsions.
 
technogiant said:
And given what you say about acoustics there is no measurement of that here?....that would at least be one equal factor which you could hang your hat on and say that is fair.

But no there is no measurement of that to show they are equal and in fact no measurement of what level the author has decided is "reasonable".....it just happens conveniently to be a level where the 290x is running at 90c....so sorry...this is far too subjective and qualitative rather than quantitative to be considered a fair review....but I'll hang around this site just to point out the flaws and bias that some of you seem so happy to swallow..
Click to expand...

Seriously, it will help you to actually read the review.
90C is a combination of clock speed, voltage and fan speed.
They could easily have had lower temperature by reducing the voltage and clock speed, but they wanted to find the max clock speed that was stable while not going to max temp and not being too loud.

The whole point of writing up how they review is so that you understand how they review.
I suggest reading up how they review.
If this isnt your cup of tea, fair enough. Go somewhere that is.
If you continue like this, I will report your posts.
 
Somewhat poor clocking chip we have here. I think the value is in the 290 really. Given MRSP prices, the 290 is much better value, considering it's only 3-4% slower clock for clock. Trix cooler is superior to Asus cooler, so 290 Trix is where it's at IMO.
 
Nenu said:
Seriously, it will help you to actually read the review.
90C is a combination of clock speed, voltage and fan speed.
They could easily have had lower temperature by reducing the voltage and clock speed, but they wanted to find the max clock speed that was stable while not going to max temp and not being too loud.

The whole point of writing up how they review is so that you understand how they review.
I suggest reading up how they review.
If this isnt your cup of tea, fair enough. Go somewhere that is.
If you continue like this, I will report your posts.
Click to expand...


I completely understand the logic of this review, as always its well done.
I also understand this guy's logic as well.
Threatening him with the report button is pretty immature.
We should all be allowed our respective opinions as long as they are civil.
Its pretty obvious to me, using two 290Xs, they need waterblocks if you want the highest performance.
 
I understand Technogiant's point. I also agree that if the fan is too loud then the premise of a real world test would come under scrutiny. So maybe they could have tossed in a "fun" 100% fan run, but not have it count towards the conclusion. But if nobody will run their cards at those settings then what is the point?

To people musing over the fact that the Asus AMD card doesn't overclock as well as other cards in other reviews, that's just the manufacturing process variance. I have a Sapphire 7950 that does 1100 / 1575 all day with the stock cooler. I have a PowerColor that chokes at 1000 / 1275. Since I run them in Crossfirex I'm stuck at 900 / 1275 as the PowerColor will crash at 1000 sometimes. It's just the luck of the draw.

I have a FX-8120 that won't go above 4.2GHz. I was chided for being a noob since all of them can get to 4.6GHz some years ago. Found out later on that some people couldn't get above 4.0GHz. My FX-9370 using the same motherboard that the FX-8120 was in does 5.0GHz without doing a thing but raising the FSB. I don't even have to tweak the voltage to run at 5.0 all day, everyday. Guys with FX-8350's, which are the same exact chip, sometimes can't get above 4.6GHz. It's all just manufacturing variance. And it should be shown in reviews to show people what the worst case scenario is before they run out to buy something that may or may not suit their expectations.

Great review as always! Thanks guys!
 
magoo said:
I completely understand the logic of this review, as always its well done.
I also understand this guy's logic as well.
Threatening him with the report button is pretty immature.
We should all be allowed our respective opinions as long as they are civil.
Click to expand...
He went beyond civil by saying this.
technogiant said:
but I'll hang around this site just to point out the flaws and bias that some of you seem so happy to swallow..
Click to expand...



cageymaru said:
To people musing over the fact that the Asus AMD card doesn't overclock as well as other cards in other reviews, that's just the manufacturing process variance.
Click to expand...
Brent made a comment about this in the review.
He predicted that the maximum overclock will be less than on other sites because [H] have spent a long time using the card and have made sure it is rock stable all the time.
Whereas most reviews will do a comparatively quick overclock test.
Your point may apply as well, but we probably wont ever know.
 
Question :

Can the 290(x) chip possibly be paired up with faster memory? With the ultra-wide bus this could make for a drastic speed increase, I'm curious if there may be a technical reason (TDP?), or is it just cost? (At which point it opens up further possibilities for AIBs)
 
You must log in or register to reply here.
Back
Top