Author

Topic: Crossfire mining? (Read 30281 times)

hero member
Activity: 504
Merit: 502
June 14, 2011, 11:17:53 AM
#20
Hi i'm  planning on doing a Crossfire setup of 2 5850s but my motherboard is 16x 4x on the PCI-E Lanes. How should I go about setting that up? Will I get around 700 MH/s at least? Suggestions are greatly appreciated.

You need a decent card brand, first. I was only able to get 700+ MH/s on ASUS cards because nothing else would OC high enough. I tried some Visiontek cards, but I was simply unable to get past 600 MH/s because the cards are low-quality.

I was wondering about PCI-e bus lane multiplier reductions, too, but it turns out it isn't an issue in mining.

To get 710 MH/s, I used the poclbm miner with flags "-v -w128." Overclocking's dead-simple in Afterburner. I was able to increase my core clock speed (and this is where brand's very important) up to 912 MHz on one card, 925 MHz on the other I'm using. I also downclocked my mem speed to 319 MHz (see here). Your max OC will vary depending on how lucky you get with the quality of your particular card.

Im doing 410mhash per hd5850 at the following, 995/180 , the difference between memory @ 300 and @ 180 is about 10mhash per card, so give that a try rather than your memory @ 319.

I used RBE to biosmod the core-mem so no software required.
donator
Activity: 1218
Merit: 1015
May 13, 2011, 10:38:49 AM
#19
Hi i'm  planning on doing a Crossfire setup of 2 5850s but my motherboard is 16x 4x on the PCI-E Lanes. How should I go about setting that up? Will I get around 700 MH/s at least? Suggestions are greatly appreciated.

You need a decent card brand, first. I was only able to get 700+ MH/s on ASUS cards because nothing else would OC high enough. I tried some Visiontek cards, but I was simply unable to get past 600 MH/s because the cards are low-quality.

I was wondering about PCI-e bus lane multiplier reductions, too, but it turns out it isn't an issue in mining.

To get 710 MH/s, I used the poclbm miner with flags "-v -w128." Overclocking's dead-simple in Afterburner. I was able to increase my core clock speed (and this is where brand's very important) up to 912 MHz on one card, 925 MHz on the other I'm using. I also downclocked my mem speed to 319 MHz (see here). Your max OC will vary depending on how lucky you get with the quality of your particular card.
full member
Activity: 154
Merit: 100
May 13, 2011, 01:36:35 AM
#18
Hi i'm  planning on doing a Crossfire setup of 2 5850s but my motherboard is 16x 4x on the PCI-E Lanes. How should I go about setting that up? Will I get around 700 MH/s at least? Suggestions are greatly appreciated.
sr. member
Activity: 714
Merit: 250
May 12, 2011, 11:06:45 PM
#17
No differemce with my two 5870s. Both my gpus reach 99%, I just set their aggression high.
hero member
Activity: 846
Merit: 1000
The One and Only
May 12, 2011, 10:23:52 PM
#16
With crossfire on usually you will not reach 99% usage on every card.

Not always true. I've got crossfire enabled on my 5870x2's and I get 99% usage on all four GPU's even with crossfire enabled. (I added the 5870x2 listings on the wiki)
qed
full member
Activity: 196
Merit: 100
May 12, 2011, 05:42:45 PM
#15
With crossfire on usually you will not reach 99% usage on every card.
donator
Activity: 1218
Merit: 1015
May 12, 2011, 03:12:20 PM
#14
That's weird, considering I see the exact opposite opposite results when I lower -f. If i set it at 1 the computer becomes hard to use due to graphical lag, but I get an extra ~10% Mh/s.

What if you we're to run the cards independently with -f set to 1?

355-357 MH/s each, a very, very slight improvement. A bottleneck is created somewhere when -f parameter is set very low while CFX is enabled. I'm not knowledgeable enough to know what's going on, just experimental data to suggest. Others experiencing significant slowdowns on CFX should try -f 50 or higher, or just remove the argument altogether and see if it improves.

With -f 1 with CFX enabled, one GPU is @ 96-99% load while the other is low and extremely erratic, sometimes only 2 MH/s, then next poll it's at 250 MH/s. With -f 50, or without the argument altogether, load on both cards is 99%.
newbie
Activity: 23
Merit: 0
May 12, 2011, 01:15:30 PM
#13
I'm running 2 5850s in CrossfireX and have the top-rated card performance listed on the mining hardware comparison wiki page (710 MH/s combined). Both run at 99% load. Fwiw. Win7 x64

I would attribute your higher output entirely to your ~200mhz core overclock. What do they put out individually? Have you tried not running them in crossfire for comparison?

Roughly 355 MH/s each with CFX disabled. The GPUs can become very inefficient if the "-f" parameter is set very low (<40). I think this is what's happening in most peoples' cases when they're running CFX.

That's weird, considering I see the exact opposite opposite results when I lower -f. If i set it at 1 the computer becomes hard to use due to graphical lag, but I get an extra ~10% Mh/s.

What if you we're to run the cards independently with -f set to 1?
donator
Activity: 1218
Merit: 1015
May 12, 2011, 12:51:15 PM
#12
I'm running 2 5850s in CrossfireX and have the top-rated card performance listed on the mining hardware comparison wiki page (710 MH/s combined). Both run at 99% load. Fwiw. Win7 x64

I would attribute your higher output entirely to your ~200mhz core overclock. What do they put out individually? Have you tried not running them in crossfire for comparison?

Roughly 355 MH/s each with CFX disabled. The GPUs can become very inefficient if the "-f" parameter is set very low (<40). I think this is what's happening in most peoples' cases when they're running CFX.
newbie
Activity: 23
Merit: 0
May 12, 2011, 12:39:22 PM
#11
I'm running 2 5850s in CrossfireX and have the top-rated card performance listed on the mining hardware comparison wiki page (710 MH/s combined). Both run at 99% load. Fwiw. Win7 x64

I would attribute your higher output entirely to your ~200mhz core overclock. What do they put out individually? Have you tried not running them in crossfire for comparison?
donator
Activity: 1218
Merit: 1015
May 12, 2011, 09:52:55 AM
#10
I'm running 2 5850s in CrossfireX and have the top-rated card performance listed on the mining hardware comparison wiki page (710 MH/s combined). Both run at 99% load. Fwiw. Win7 x64
hero member
Activity: 531
Merit: 505
May 12, 2011, 02:13:26 AM
#9
I have dummy plug in the "slave" card.
member
Activity: 92
Merit: 10
May 12, 2011, 02:07:46 AM
#8
Dummy plug the secondary card so the system doesn't dumb it down, because it doesn't detect display?
hero member
Activity: 531
Merit: 505
May 12, 2011, 12:37:32 AM
#7
The "slave" card also uses only 90% of GPU, while master runs at 99% all time. Damn!!! I would bet this did not happen until I tried that damn CrossFire thing!! Does someone with 2x5870 experience this too (in Windows)?

Regarding the Linux - Is there a dedicated mining-optimized distribution? Like, boot from LiveCD and run it?
full member
Activity: 154
Merit: 100
May 12, 2011, 12:30:10 AM
#6
Guys, how to get rid of that "ATI Crossfire could not be enabled" message after Windows boot up?

Install Linux.  Wink

Or don't install Catalyst Control Center...
member
Activity: 68
Merit: 10
May 12, 2011, 12:27:17 AM
#5
Guys, how to get rid of that "ATI Crossfire could not be enabled" message after Windows boot up?

Install Linux.  Wink
hero member
Activity: 531
Merit: 505
May 12, 2011, 12:22:04 AM
#4
Guys, how to get rid of that "ATI Crossfire could not be enabled" message after Windows boot up?

I tried to install the CrossFire strip and realized, as stated before, that CrossFiring hampers performance of OpenCL. So, I removed the strip. But now, the message appears, drivers are still trying to CrossFire the cards !!!!

Also, in some graphics card log, I can see that one is Master, other is Slave.

And, finally, that "slave" one is slower a bit .. even when there is no CrossFire strip installed.

How to get into state before I tried the CrossFire?
newbie
Activity: 23
Merit: 0
May 11, 2011, 11:40:23 PM
#3
i am fairly certain if you run your cards with crossfire enabled you will see a decrease in performance
member
Activity: 63
Merit: 10
May 11, 2011, 11:26:58 PM
#2
i have test in 2 system
i7 2x5850 crossfire 870@500 =330mh
amd 4600 1x5850 870@500 = 330mh too no diffen
full member
Activity: 168
Merit: 100
May 11, 2011, 10:57:26 PM
#1
I think some people on the forums mine with corssfire set up?

Currently I'm mining with no crossfire bridge.  If I enable crossfire will efficiency decrease?  How does that work?

I would like to be able to use crossfire when I game.

Thanks!
Jump to: