I experience half the FPS of one card with a crossfire R9 290 setup.

[RCoon]

500Mhz is enough to conduct the experiment. The OP has nothing to lose at this point.

CPU OC for experimental purposes yes, but there is zero use in overclocking his RAM

 

[Mussels]

CPU OC for experimental purposes yes, but there is zero use in overclocking his RAM

both go up together with BCLK OCing

 

[JaredzzC]

I'm not exactly comfortable with overclocking my RAM (or if it's even possible with my one), so I'd prefer not to.
I don't mind overclocking my CPU, but I doubt that will help much. It already turbo boosts from 3.3-3.6 when it needs to, so I don't think I would get much higher than 3.6, but I can always try.

 

[Mussels]

I'm not exactly comfortable with overclocking my RAM (or if it's even possible with my one), so I'd prefer not to.
I don't mind overclocking my CPU, but I doubt that will help much. It already turbo boosts from 3.3-3.6 when it needs to, so I don't think I would get much higher than 3.6, but I can always try.

turn the ram down a notch, then raise the bclk.

i managed 103.5 on my last board, 107 on this one (same CPU and ram) - but a 5% boost is a 5% boost

 

[yogurt_21]

well, on the bright side if it is the cpu, an ivy bridge upgrade should fix the issue, you might just have to run with 1 card while the other waits for you to save enough to get a 3570k or if you can swing it a 3770k. Annoying it makes such a difference but I guess 3.0 card + 3.0 slot through a 2.0 lane on the cpu = trouble. Didn't newtechie just recommend someone put tape on the card to try and force it to 8x would something like that work here?

 

[JaredzzC]

well, on the bright side if it is the cpu, an ivy bridge upgrade should fix the issue, you might just have to run with 1 card while the other waits for you to save enough to get a 3570k or if you can swing it a 3770k. Annoying it makes such a difference but I guess 3.0 card + 3.0 slot through a 2.0 lane on the cpu = trouble. Didn't newtechie just recommend someone put tape on the card to try and force it to 8x would something like that work here?

Yea, I have my sister's birthday coming up as well as other stuff I need to buy, so I will be spending a few hundred all together for that, so I guess i'll have to wait until Christmas for my new CPU :/ At least I now know what the problem is so it can be fixed : )

The cards already run @ x8/x8 already(as far as i'm aware) so there's no need to force it to do so.
As far as I'm aware, the R9 290 can't even fully utilise 3.0 bandwidth, but I guess it is being restricted by 2.0

 

[ThE_MaD_ShOt]

Well I can say I never had that issue with Bf3 with my crossfired setup. haven't ran crysis3. I am running a oc'd fx8350 so maybe that would be the difference for me.

 

[bubbleawsome]

That makes sense. IIRC a 2.0x16 bus was nearly saturated by one 7990. Two cards running without the CF connector since they have the new XDMA stuff or whatever it is called could easily surpass 2.0 at 8x.

 

[Naito]

I'm not all too familiar with Sandy Bridge, but can't the multiplier on the non-k CPUs be raised by 4 bins or so (similar to Ivy Bridge)? If correct, why not raise the ratio and get 3.8GHz? without wasting time with bCLK tweaks.

 

[JaredzzC]

I'm not all too familiar with Sandy Bridge, but can't the multiplier on the non-k CPUs be raised by 4 bins or so (similar to Ivy Bridge)? If correct, why not raise the ratio and get 3.8GHz? without wasting time with bCLK tweaks.

I think I overclocked it quite a while back ( a few years back now ), but then it kept resetting itself back to 3.3 when I booted the PC. Maybe bc I clocked it too high? (or maybe I'm just talking nonsense... can't remember if it just doesn't let you go above a certain frequency or not).
I don't mind overclocking it, but as I said before - I doubt it would make enough difference to make crossfire work properly for those games/other things.

 

[Naito]

I think I overclocked it quite a while back ( a few years back now ), but then it kept resetting itself back to 3.3 when I booted the PC.

I'd say almost any would be able to cope with a clock increase of 400Mhz (4x on ratio), hence why the option is there. Once you start touching the bCLK, it may lead to instability. Having said that, you might not have been lucky in the "silicon lottery" and may need a voltage (read up on what needs to be changed) increase or two to support the boost.

 

[AsRock]

the 1x1.1 is a power saving feature for idle or low clock states. Literally, I see nothing wrong in GPU-Z or in the way the PC is set up. Visually, there is absolutely nothing wrong, and software isn't reporting issues either.

It may be worth having Afterburner open, and check the GPU Usage figures for each GPU core during a render test or something. Then we can see if each GPU is at 99% usage during the render, despite only providing half the performance, or if each GPU is being bottlenecked down to 50% during the test.

EDIT: By the way, there is a VRM_1 issue with XFX 290 DD cards. After reaching a certain temperature, DD 290/290X's will severely throttle performance. They literally crap themselves because the VRM cooler doesn't even exist. May be worth noting the temps. Maybe airflow is bad.

High vrm doe's not throttle speeds. High GPU temps throttle speed.

And VRM is normally higher than VRM2 although with good air flow should be a none issue.

 

[Mussels]

I'm not all too familiar with Sandy Bridge, but can't the multiplier on the non-k CPUs be raised by 4 bins or so (similar to Ivy Bridge)? If correct, why not raise the ratio and get 3.8GHz? without wasting time with bCLK tweaks.

some of them can (mine went from 3.1 to 3.6) but not all boards do it stable.

in my case the board lets me choose up to 38x, but would only really work at 36x.

 

[bubbleawsome]

Where does BCLK oc get weird. Isn't it around 106Mhz? (SATA corruption and stuff)

 

[Mussels]

Where does BCLK oc get weird. Isn't it around 106Mhz? (SATA corruption and stuff)

depends entirely on the board, my last one couldnt do 104 without SATA crapping out

 

[RCoon]

High vrm doe's not throttle speeds. High GPU temps throttle speed.

And VRM is normally higher than VRM2 although with good air flow should be a none issue.

It does, you're welcome to check the hard forum reports on throttling caused by the DD coolers.

 

[Aquinus]

depends entirely on the board, my last one couldnt do 104 without SATA crapping out

It depends on the chipset, but on average 5Mhz is about the best you're typically going to do on the BCLK. It doesn't typically like BCLK overclocks because DMI becomes unstable as it is tied directly to the BCLK (much like the PCI-E root compelx,) so a variety of things can start acting up if you push it too far. Bumping the PCH voltage may give you an extra Mhz or two, but I can't say for certain.

 

[Dent1]

JaredzzC,

Run a CPU benchmark like Sysmarks or WPrime, and compare the results to other users with the same CPU.

If your score deviates from theirs then you know that your CPU is underperforming, which will explain the bad CF performance.

This would then narrow it down to software, OS, configuration, overheating etc.

 

[AsRock]

It does, you're welcome to check the hard forum reports on throttling caused by the DD coolers.

I could check all i wanted it's not what i see with mine i had the VRM as high as 105c there was no down clocking happening how ever when the GPU gets around 95c then the down clocking happens. I have tried it i have all so talked to one of the XFX engineers about it as it was a issue.

I have noticed this though using GPU-Z with TPU OSD as it tells me the temps and clocks while in game and seen no proof in what your claiming.

I solved my issue by putting my system back in my case and tweaking the fans and taking the cooler completely of the card which the black plate which is on them needs care to take of as it can feel like it glued on but once of you can make sure that all contact is good.

A warning to the OP if he refits the cooler depending were he is from could void the warranty, example any were other than the USA and Canada i believe it be void as they put the stickers on all the cards not just were it's not allowed.

 

[Naito]

some of them can (mine went from 3.1 to 3.6) but not all boards do it stable.

in my case the board lets me choose up to 38x, but would only really work at 36x.

Oh ahk, wasn't sure. Would have thought the of the last few generations to be fairly stable (although, I know Nehalem was knife edge at times) by now.

 

[RCoon]

I could check all i wanted it's not what i see with mine i had the VRM as high as 105c there was no down clocking happening how ever when the GPU gets around 95c then the down clocking happens. I have tried it i have all so talked to one of the XFX engineers about it as it was a issue.

I have noticed this though using GPU-Z with TPU OSD as it tells me the temps and clocks while in game and seen no proof in what your claiming.

I solved my issue by putting my system back in my case and tweaking the fans and taking the cooler completely of the card which the black plate which is on them needs care to take of as it can feel like it glued on but once of you can make sure that all contact is good.

A warning to the OP if he refits the cooler depending were he is from could void the warranty, example any were other than the USA and Canada i believe it be void as they put the stickers on all the cards not just were it's not allowed.

VRMs at 128 degrees causes throttling on the first gen of 290 DD coolers. XFX were made aware of it and there is hard proof the VRMs were not getting cooled properly, they admitted and said they would resolve it. Claiming you do not have a problem does not mean everyone else doesn't. You can say yours works fine, I am fully aware that a lot of the first gen didn't and it did cause throttling.

If you'd like to discuss this specific issue further you are more than welcome to PM me. I have ample factual information I can provide.

 

[AsRock]

VRMs at 128 degrees causes throttling on the first gen of 290 DD coolers. XFX were made aware of it and there is hard proof the VRMs were not getting cooled properly, they admitted and said they would resolve it. Claiming you do not have a problem does not mean everyone else doesn't. You can say yours works fine, I am fully aware that a lot of the first gen didn't and it did cause throttling.

If you'd like to discuss this specific issue further you are more than welcome to PM me. I have ample factual information I can provide.

Well ok,, how ever if your vrm's are getting that hot the pads on the card need sorting out or you have really bad air flow.. It was 29c room temp here a few weeks ago and i was running the card hard and still did not hit those temps.

I will say that the cooler design is a bit of a pain as it sits close to the black plate that cools the vrm which makes it a pain to just stick extra heatsinks to the black plate.

And XFX will recommend you to send the card in to them if it's hitting any higher than 95c.

I don't know which version i have never checked on it but to guess i would say a later version but i only say that due to the different packaging i have seen of others as mine advertises the newer bios for win8.

All so mines their top tire card too.

 

[newconroer]

To be clear, we are telling him that an i2500 is bottlenecked by a standard Crossfire setup?

The performance loss he is describing can not be down to the i2500. If it was, then every 'gaming' bundle out there that you could buy from legitimate companies two or three years ago for the Sandy Bridge, with a SLI or Crossfire setup, would have been pointless!

 

[GhostRyder]

To be clear, we are telling him that an i2500 is bottlenecked by a standard Crossfire setup?

The performance loss he is describing can not be down to the i2500. If it was, then every 'gaming' bundle out there that you could buy from legitimate companies two or three years ago for the Sandy Bridge, with a SLI or Crossfire setup, would have been pointless!

2 or 3 years ago being the key word there. Were talking about recent games and much higher power GPU setups and resolutions. The Sandy-Bridge architecture is a few years old, but maintains good performance when overclocked very far as it is excellent for that. The problem here is the i5 2500, its base and boost clocks are not extremely high mixed with the fact that its on an old architecture (Which would mean about a 20% increase give or take on new chips clock to clock) and its clocks are pretty low. An i5 2500k clocked at 4.5ghz would still be beyond excellent for gaming and a single GPU on the base clocks would be fine, however multi-gpu setups require a bit more CPU power in this day and age and that chip is just showing a bit of age.

Not a bad chip, he can follow what the others have stated and bump the clocks a bit to see how much that helps. However I do not know how far that will take him in the end.

 

[kinggavin]

I HAVE 290 CROSSFIRE AND FX 8350 AND 1300W PSU
NUMBER 1- 850W IS NOT ENOUGH YOU NEED 1000W MINIMUM FOR CROSSFIRE 290
NUMBER 2 TWO GPUS ON AN INTEL I5 IS A BAD IDEA YOU NEED UPGRADE TO AN INTEL I7 OR FX 8350