Permabanned
- Joined
- 9 Sep 2010
- Posts
- 363
Oh, it will also be worth checking to see if there are any drivers updates for your motherboards chipsets!
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
HD5850 and HD5830 Crossfire
- Thread starter mattgal
- Start date
Associate
- Joined
- 15 Jun 2007
- Posts
- 1,201
- Location
- Liverpool
The 5850 is by far the faster card. Who cares about core clock when you have many more SP's and ROP's?
Caporegime
- Joined
- 18 Sep 2009
- Posts
- 30,522
- Location
- Dormanstown.
Crossfire doesn't down clock.
The cards run at there own dependent speeds..
The cards run at there own dependent speeds..
well if we are going in to this really the 5830 can do more frames per second as its faster core clock makes it render faster.
But the 5850 can process more pixels per clock then the 5830. this is how it performs better.
in terms of faster you may have a point there but it can lead to confusing situations when the device your measuring on speeds had a chip that runs at a speed.
back on topic i know this may sound mad but put the 5830 in the top slot. if there is any down clocking then this will force the 5850 to OC to the 5830 speeds.
but looking on google many people confirm that this is a fault and there appears to be no work around.
Man of Honour
- Joined
- 16 May 2005
- Posts
- 31,297
- Location
- Manchester
You're completely ignoring the fact that when people refer to a card being 'faster' they are talking about the overall graphics performance. There's no lack of clarity from anyone else here - the 'faster' card is the 5850.
This kind of thing is a common mistake made by people that don't really know their way around computer hardware - assuming the biggest number of mhz/ghz is always best.
hey i know my stuf. this is the first time i have ever heard people say a card is faster than another card. i am very used to a card Performing better than another one. i accept that this is a difference with this forum.
Soldato
- Joined
- 23 Apr 2009
- Posts
- 11,973
- Location
- Cheshire
I'm sorry m8 - when people say "faster" in terms of graphics cards they mean that the card produces a higher framerate in games. In other words the actual performance is better.
If you say "faster" without qualification, people will not assume that you mean core frequency - but assume overall performance (in games).
To put this in other terms, you have Car A and Car B. Card B's engine runs at a higher RPM than Car A - however the acceleration and top speed of Card A are higher (due to other factors like engine design and gearing). Therefore one would say that Car A is "faster" due to its better performance, the RPM of Car B wouldn't factor into this.
cmndr andi thank you you put it in good perspective. lesson learnt.
this how ever, if we are talking about core speed, is wrong as the 5830 has a higher core clock thus it would have to slow down to the 5850.
now i am fairly sure crossfire wont disable parts of the core to match the card? but if it did then the 5850 would have parts of its core disabled to match.
this how ever, if we are talking about core speed, is wrong as the 5830 has a higher core clock thus it would have to slow down to the 5850.
now i am fairly sure crossfire wont disable parts of the core to match the card? but if it did then the 5850 would have parts of its core disabled to match.
Soldato
- Joined
- 23 Apr 2009
- Posts
- 11,973
- Location
- Cheshire
Crossfire works by alternate frame rendering. I'm not sure if the actual clockspeed of any of the cards is changed - but the faster (better performing) card has to wait for the next frame to be rendered by the slower card. Hence, in the dual card arrangement the faster card can only perform at the level of the slower card - as for a proportion of the time it is "waiting" for the slower card to complete its task.
For the OP I strongly suspect that it is a driver issue with a very new game. Often there are numerous bugs, especially for making multi-GPU setups work correctly. I would expect a patch and/or new drivers imminently.
For the OP I strongly suspect that it is a driver issue with a very new game. Often there are numerous bugs, especially for making multi-GPU setups work correctly. I would expect a patch and/or new drivers imminently.
Man of Honour
- Joined
- 16 May 2005
- Posts
- 31,297
- Location
- Manchester
This isn't a cute little way of talking that is limited to this forum - it's standard everywhere on the internet. Try reading any hardware review sites or any other forums.
Permabanned
- Joined
- 9 Sep 2010
- Posts
- 363
Try the different catalyst drivers available if that doesn't make a difference I don't know.
You could try installing the 5850 first, overclock it to the same core and memory speeds as the 5830 and then install the 5830.
Although in all honesty don't think you should be getting so much trouble!
Normally when you xfire two cards that are not the same like with your 5850 and 5830 you should see the same results as if you were xfireing with two of the lowest spec card. So you should see similiar results to 2x5830 in xfire, which should be more than your 5850 by itself!
Can you try the cards in a friends mobo/comp to see if they work fine?
You could try installing the 5850 first, overclock it to the same core and memory speeds as the 5830 and then install the 5830.
Although in all honesty don't think you should be getting so much trouble!
Normally when you xfire two cards that are not the same like with your 5850 and 5830 you should see the same results as if you were xfireing with two of the lowest spec card. So you should see similiar results to 2x5830 in xfire, which should be more than your 5850 by itself!
Can you try the cards in a friends mobo/comp to see if they work fine?
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
Which motherboard slots do you have it in, a quick check seems to suggest the mobo you have has 2 8x slots and 1 4x slot, the 4x slot might be slowing it down even further.
The more obvious question is, have you tried the 5830 on its own? Maybe its just faulty, xfire is working really VERY well these days, even on 58xx cards you shouldn't get crashing nor a massive speed drop. The VERY worst thing I ever had was single card performance, and that in the past 2 years has been so rare and far between its basically none existant, the worst games for scaling these days are, Civilisation, Metro 2033, and you're still talking 60% scaling or so even on 58xx cards(Metro scales similarly badly on SLI).
So I'd guess, faulty card, faulty xfire bridge is certainly possible, driver combination. Load up something like gpu-z, couple of instances set it to log in background and have each instance monitor a different gpu, see if they both get to anywhere near 100% load in gaming, same thing use task manager to check if the cpu is just not loading for some reason in games, see if something else is eating cpu time.
Windows updates, disable antivirus for a bit, maybe turn off network connection to be safe first, then install drivers again(the odd anti virus program or update can get upset at drivers and block bits being installed, rare, but happens).
Make sure in gpu-z aswell of the clock speeds, maybe one isn't clocking up in games, which will absolutely trash performance of both cards, as someones said, you can have a 5970 (or 6990, or a 580gtx) in one slot, if the other card is running at 250mhz clocks, then the fast card still has to wait between every frame for the slow card and then it only does one frame.
It can still also be power, though I'm not familiar with your power supply, if its separated into rails you can(more often if you use a lot of adaptors for pci-e power connectors) can put too much load on one of the PSU rails, thats less of an issue these days though.
The more obvious question is, have you tried the 5830 on its own? Maybe its just faulty, xfire is working really VERY well these days, even on 58xx cards you shouldn't get crashing nor a massive speed drop. The VERY worst thing I ever had was single card performance, and that in the past 2 years has been so rare and far between its basically none existant, the worst games for scaling these days are, Civilisation, Metro 2033, and you're still talking 60% scaling or so even on 58xx cards(Metro scales similarly badly on SLI).
So I'd guess, faulty card, faulty xfire bridge is certainly possible, driver combination. Load up something like gpu-z, couple of instances set it to log in background and have each instance monitor a different gpu, see if they both get to anywhere near 100% load in gaming, same thing use task manager to check if the cpu is just not loading for some reason in games, see if something else is eating cpu time.
Windows updates, disable antivirus for a bit, maybe turn off network connection to be safe first, then install drivers again(the odd anti virus program or update can get upset at drivers and block bits being installed, rare, but happens).
Make sure in gpu-z aswell of the clock speeds, maybe one isn't clocking up in games, which will absolutely trash performance of both cards, as someones said, you can have a 5970 (or 6990, or a 580gtx) in one slot, if the other card is running at 250mhz clocks, then the fast card still has to wait between every frame for the slow card and then it only does one frame.
It can still also be power, though I'm not familiar with your power supply, if its separated into rails you can(more often if you use a lot of adaptors for pci-e power connectors) can put too much load on one of the PSU rails, thats less of an issue these days though.
i tried the 5830 on its own and it doesn't perform THAT bad (3100 points in kombastor and 44fps in NFS:HP)
the psu i have has a single 12v rail with 68Amps
when CF is enabled, i use msi afterburn (for custom fan profiles) and selected GPU usage for both GPUs to show.
they usually average at 60s but they fluctuate a lot, sometimes gpu1 is 0 and gpu2 is 30%. most of the time this happens in loading screens.
also, there seems to be an issue when choosing the car.. fps drop to a solid 7.4/7.5 fps though then viewing the car (and choosing what color) it goes up to 50s again.
the psu i have has a single 12v rail with 68Amps
when CF is enabled, i use msi afterburn (for custom fan profiles) and selected GPU usage for both GPUs to show.
they usually average at 60s but they fluctuate a lot, sometimes gpu1 is 0 and gpu2 is 30%. most of the time this happens in loading screens.
also, there seems to be an issue when choosing the car.. fps drop to a solid 7.4/7.5 fps though then viewing the car (and choosing what color) it goes up to 50s again.
