6990 Underwhelming

[Jobot]

Hello all. My problem is that I just purchased the MSI HD 6990 (supposedly a huge upgrade from the 5870 1gb), and it seems like it's performing about the same as my 5870 did on some games, and it seems worse on others. I'm sure my system is bottlenecking it, somehow, but I don't know what to update or alter in order to get it running up to par.

I'm not even asking much of it, only running it 1920x1080, and was looking to add more monitors to the mix, but if it's giving such a lackluster performance at 1080, then I'd hate to see it running on multiple displays. Anyway, my system is setup as follows:

Processor: i7 920 @ 3.6
Motherboard: Asus P6T
RAM: 12gb 1600 Triple-Channel Corsair Vengeance
PSU: Corsair TX 850w
Graphics: MSI HD6990

Thanks in advance.

 

[cjl]

It should definitely be substantially faster than the 5870. What kind of tests are you running, and what are the results?

 

[asantesoul]

damn...sounds like it could be a bottleneck..but not too sure..since your playing at such a low resolution some games will require more of the cpu power...go into the task manager for some games and check the cpu's utilization..and check the gpu's usage as well..its possible that at that resolution the video card isnt being using much of its potential

 

[Jobot]

It should definitely be substantially faster than the 5870. What kind of tests are you running, and what are the results?

Only benchmarking that I've really done to it is with Unigine. Maxed out all the settings, extreme tessellation, 16x Anisotropy, 8x AA, DirectX 11, at 1920x1080. Score came out to be 1161, 46.1 FPS (60.7 max; 19.3 min). I did some tests on lower settings, but they still seemed underwhelming.

The games I've been playing are Dragon Age 2, Rift, and DC Universe Online. Dragon Age 2 runs decent enough, but stutters occasionally, for no apparent reason. Rift gets 60fps vertical sync, when I'm holding still, but fluctuates a lot when I'm moving around the world, even if it's not in a high pop area. DC Universe disappoints me, because it stutters a lot, and this card should be able to destroy that game. DC is a particular case where my 5870 ran it smoother, with the same video settings.

I was thinking my 850w psu may be a part of it, seeing as I OCed my cpu, and am running 2-3 SATA hard drives at any given time, along with my 6990. I wouldn't think that my i7 would be affecting it so drastically. I'm at a loss

 

[phendric]

It should definitely be substantially faster than the 5870.

What he said.

I'm running a system similar to yours, except with a dual 5870 in crossfire, and I only have a 750W PSU that's working great. My first thought wouldn't be that your system is bottlenecking the card.

I can, however, think of 4 potential problems:

■Your screen resolution isn't nearly high enough to see the advantages of the 6990 over the 5870. At 1920x1080, 1 GB VRAM is probably large enough for the frame buffer without paging out to system RAM, so 4 GB doesn't gain you anything. What benchmark results do you get with Unigine under the 5870?
■You either don't have updated drivers OR you need to wait for AMD to release a couple of updates before you'll see the higher performance - their early graphics drivers sometimes have serious problems.
■You have a faulty card?
■I know that if the card gets too hot, it automatically underclocks itself, and you lose performance. Could this be happening to you? I know that the 6990 blows just as much hot air into the case as it does out - do you have a case with good air flow?

P

 

[asantesoul]

As i mentioned before...I suspect its the resolution your playing at...if it was the psu (which is feasible)...you would get more severe issues...the fact that your playing at that resolution with that powerful 4gb vram monster may be the reason your frames are suffering..and yeah, the drivers could be an issue as well...

have you tried updating them to the beta drivers..im not too familiar with ati drivers but there should be a recent one out...

 

[m0ngy]

Hello all. My problem is that I just purchased the MSI HD 6990 (supposedly a huge upgrade from the 5870 1gb), and it seems like it's performing about the same as my 5870 did on some games, and it seems worse on others.

Hey man, I have almost exactly the same system as yours (920@3.6, P6Td, 6Gb RAM, 6990, 1200wPSU, 24" 1080p) and have just upgraded from the 5870 to the 6990... I'm also a bit underwhelmed.

In BC2 I'm getting more frames generally, but when the action's thick (when it really counts) I'm only getting 10-20 fps more than with the 5870.

In Shogun 2 the 6990's performance is either on par or worse than the 5870, I'm not entirely sure, but it ain't great, this I know.

I'm a bit dissapointed really, not sure what the hell is going on.

 

[asantesoul]

increase the settings in BC2 to the MAX and if you can play on a higher resolution do that as well..if not, your resolution is bottlenecking the card..disable the crossfire on the 6990 in ccc and check if ur frames are better on the same res...update drivers as well

 

[adam_x_brookes]

wait untill games come out to really test the card and show the HD5870 up. There is no doubt though that if you had eyefinity that the HD6990 would definatly shine through over the HD5870. <--- It's a great card and with it I can max out most games at 1980 x 1080

 

[asantesoul]

OP as long as your using resolutions like that, your monitor will be bottlenecking your cards lol...you need to play at higher resolutions to utilize more power..disable the crossfire in the ccc and your frames will increase because 1 card will handle the resolution a bit better

 

[bystander]

OP as long as your using resolutions like that, your monitor will be bottlenecking your cards lol...you need to play at higher resolutions to utilize more power..disable the crossfire in the ccc and your frames will increase because 1 card will handle the resolution a bit better

1 card should not improve performance over 2, unless he's playing a game that's not supported by CF or it's bugging.

You are misunderstanding what a bottleneck is. Lowering the performance of one area of a setup to remove a bottleneck, does not improve performance, but it might not make any negative impact.

 

[rez_evil]

hi im sorry for intruding in this tread but im just wanting to ask a stupiud question what is bottlenecking mean .....

 

[asantesoul]

1 card should not improve performance over 2, unless he's playing a game that's not supported by CF or it's bugging.

You are misunderstanding what a bottleneck is. Lowering the performance of one area of a setup to remove a bottleneck, does not improve performance, but it might not make any negative impact.

umm..yeah..i do know what a bottleneck is burdy ....hypothetically speaking, if an individual decides to use an i5760 @ stock with 2 GTX 580's...they processor in this case causes a "bottleneck" because the processor is unable to adequately provide the necessary power/calculations necessary to run the two mentioned video cards...

I'm suggesting to the OP to disable the crossfire and use a single video core (for some games on the 1080p monitor he's using) in order to resolve or remedy the problem...again, I've never used radeon video cards, and am not familiar with the drivers...however, using a powerful video card like the 6990 on a monitor with a native 1080p resolution would make some sense that it may yield unsatisfactory results... too much power isn't always a good thing... i simply stated the monitor as a bottleneck and wrote "lol" as a joke..although I suspect using the card with that monitor and some games wont always deliver great results...

To think that you can just pop in a 6990 and max everything without any issues is a bit naive...sometimes, it's more complicated than that

 

[asantesoul]

A bottleneck occurs when a component or more than one component of a computer setup hinders the overall performance of another...for instance, if I attempt to use a powerful video card like the GTX 590 with an AMD phenom II X4 955(quad) processor @ the cpu's stock settings, there would more than likely be a bottleneck because the processor cannot efficiently provide the "juice" necessary to allow the GTX 590 to perform at it's best...however, if the phenom cpu is overclocked by about 600-800mhz (3.8-4.0ghz) you an reduce (probably not eliminate) the bottleneck and allow the gtx 590 to perform much better

 

[bystander]

There is problem with your idea about a bottleneck. One that is pretty important to understand. Using a high end card when the CPU or monitor cannot take advantage of it does not slow down performance. It does not hinder performance in any way.

The bottleneck will simply prevent the high end part from being used to it's full potential, it does not slow performance down beyond what the monitor and CPU are already limiting it to.

This example does have a 2nd variable, and that's crossfire vs non-crossfire, so there can be some software issues as well. Barring software issues, using one GPU core instead of two, will not improve performance.

 

[rez_evil]

A bottleneck occurs when a component or more than one component of a computer setup hinders the overall performance of another...for instance, if I attempt to use a powerful video card like the GTX 590 with an AMD phenom II X4 955(quad) processor @ the cpu's stock settings, there would more than likely be a bottleneck because the processor cannot efficiently provide the "juice" necessary to allow the GTX 590 to perform at it's best...however, if the phenom cpu is overclocked by about 600-800mhz (3.8-4.0ghz) you an reduce (probably not eliminate) the bottleneck and allow the gtx 590 to perform much better

ohhh wow thank you then i should not worry about mine im running intel I7 950....with dual ASUS EAH6850 DirectCU/2DIS/1GD5 12gb ram CORSAIR H%) HYDRO Liquid cooling

 

[asantesoul]

There is problem with your idea about a bottleneck. One that is pretty important to understand. Using a high end card when the CPU or monitor cannot take advantage of it does not slow down performance. It does not hinder performance in any way.

The bottleneck will simply prevent the high end part from being used to it's full potential, it does not slow performance down beyond what the monitor and CPU are already limiting it to.

This example does have a 2nd variable, and that's crossfire vs non-crossfire, so there can be some software issues as well. Barring software issues, using one GPU core instead of two, will not improve performance.

I stand by my statements and explanations, and fail to see where your going with yours...really, what your saying in contrast to what I've said bares no significant difference..the main idea here is performance isn't up to snuff because of software, the monitor, or a cpu bottleneck. Show me your evidence and then I can humbly accept your rationale without a rebuttal. Of course, I'm human and I make mistakes, but keep in mind that i have made suggestions based on the info provided...simple as that homaay

 

[asantesoul]

ohhh wow thank you then i should not worry about mine im running intel I7 950....with dual ASUS EAH6850 DirectCU/2DIS/1GD5 12gb ram CORSAIR H%) HYDRO Liquid cooling

no..should be good..and honestly...if there was ever any kind of bottleneck..our brains and eyes will never be able to determine it without a benchmark..unless it's a much more video card setup up and slower cpu overall.

 

[bystander]

I stand by my statements and explanations, and fail to see where your going with yours...really, what your saying in contrast to what I've said bares no significant difference..the main idea here is performance isn't up to snuff because of software, the monitor, or a cpu bottleneck. Show me your evidence and then I can humbly accept your rationale without a rebuttal. Of course, I'm human and I make mistakes, but keep in mind that i have made suggestions based on the info provided...simple as that homaay

The part that is different is you seem to be under the impression that weakening the strongest point of a system will some how improve performance. I.E. you said he'd gain performance by using only 1 of the GPU's.

Slowing down the part of a system that is not able to be fully taken advantage of, does not speed up the rest of the system.

Your definition of what a bottleneck was not wrong, but your advise to alleviate the problem was.