(edit: temps are actually 73/65/67/61... I'll take it)
Topic: Quad GPU mining rig problems (Read 3395 times)
Got my powered risers in the mail today! So I rebuilt the mining rig with the 2 powered and 2 non-powered risers on all 4 cards ( I wanted all 4 cards on risers for spacing / airflow) and so far so good... All cards are hashing away at about 550kh each, mining 2.2mh total on LTC, and the power mosfets on the mobo are less warm to touch than they were running only 2 cards directly in the 16x slots (I fried the mobo using the 16x slots + 2 non-powered risers the first time...). Fans set to 70%, temps in the low 70s ... It took a couple weeks but I think I finally got it right
(edit: temps are actually 73/65/67/61... I'll take it)
(edit: temps are actually 73/65/67/61... I'll take it)
sounds good. maybe I'll bump it up to 60% and keep them close to 75 for a good balance
below 80 sounds ok, I have the same settings. Best tradeoff between electronics temperature and fan wear (and noise!)
The blocks have got better today, almost completely gone 
whatever, it's not a huge issue as long as it still mines for me
GPU temps are typically in the 70's .... I use manual fan controls and with the extra box fan helping I can usually keep my fans at 55% and keep the GPU's in the 77-79 range. If I turn up the fans to 75% I can keep them down closer to 70. My only concern is wearing out the fans by running them at a constant level 24/7, should I be concerned about that or just leave them running faster for lower temps? I was thinking that as long as I am under 80 I should be ok, but what do you guys think?
whatever, it's not a huge issue as long as it still mines for me GPU temps are typically in the 70's .... I use manual fan controls and with the extra box fan helping I can usually keep my fans at 55% and keep the GPU's in the 77-79 range. If I turn up the fans to 75% I can keep them down closer to 70. My only concern is wearing out the fans by running them at a constant level 24/7, should I be concerned about that or just leave them running faster for lower temps? I was thinking that as long as I am under 80 I should be ok, but what do you guys think?
I ran quite a few 5870 for a couple of years with the fans at 70% 24/7 and never had any problems.
I personally would make 75c-77c your target if you plan on keeping them on 24/7. The closer you get to 80 the higher the chance over time a card will crash.
The blocks have got better today, almost completely gone whatever, it's not a huge issue as long as it still mines for me
GPU temps are typically in the 70's .... I use manual fan controls and with the extra box fan helping I can usually keep my fans at 55% and keep the GPU's in the 77-79 range. If I turn up the fans to 75% I can keep them down closer to 70. My only concern is wearing out the fans by running them at a constant level 24/7, should I be concerned about that or just leave them running faster for lower temps? I was thinking that as long as I am under 80 I should be ok, but what do you guys think?
whatever, it's not a huge issue as long as it still mines for me GPU temps are typically in the 70's .... I use manual fan controls and with the extra box fan helping I can usually keep my fans at 55% and keep the GPU's in the 77-79 range. If I turn up the fans to 75% I can keep them down closer to 70. My only concern is wearing out the fans by running them at a constant level 24/7, should I be concerned about that or just leave them running faster for lower temps? I was thinking that as long as I am under 80 I should be ok, but what do you guys think?
What are your GPU temps? 4 card rigs can overheat very quickly
On a side note I have started getting weird blocks that flicker all over my screen, not sure what the issue is there but I can still use the computer and it still hashes so I am not too concerned.... after all I only use the display to make sure the miner is still running anyway
Yes, I also get the blocks. Guess the mining takes up almost all the resources, and that interferes with the graphic cards actual tasks. But doesn't cause stability issues here either. Also no blocks on the display when using teamviewer.
I don't know, that might have been it. I have seen it happen a few times since when I restart the miner / computer, but if i stop / start the miner a couple times it usually stops by itself and keeps hashing along.... I am using an hdmi output though
On a side note I have started getting weird blocks that flicker all over my screen, not sure what the issue is there but I can still use the computer and it still hashes so I am not too concerned.... after all I only use the display to make sure the miner is still running anyway
On a side note I have started getting weird blocks that flicker all over my screen, not sure what the issue is there but I can still use the computer and it still hashes so I am not too concerned.... after all I only use the display to make sure the miner is still running anyway
ROFL.... ok, this is really weird, and I'm certainly not complaining.... but it seems to have fixed itself 
I really didn't do anything, rebooted a couple times trying to get at my bios settings, which I was unable to do because the display is hdmi and it doesn't give me output until I am actually in windows. I have guiminer set to open on login and auto start mining and I noticed my problem card started at 250 khash instead of the usual 70 khash while I was having this issue.... watched it for a few minutes and it climbed right up to the 500 mark alongside the 3 others, I am mining LTC at 2.1 mhash now. Dunno how / why, but I'll take it ;
I really didn't do anything, rebooted a couple times trying to get at my bios settings, which I was unable to do because the display is hdmi and it doesn't give me output until I am actually in windows. I have guiminer set to open on login and auto start mining and I noticed my problem card started at 250 khash instead of the usual 70 khash while I was having this issue.... watched it for a few minutes and it climbed right up to the 500 mark alongside the 3 others, I am mining LTC at 2.1 mhash now. Dunno how / why, but I'll take it ;
So was the slow card the one with the HDMI connected? Had this issue as well. Gets healed by unplugging the HDMI and replugging it. Issue does not show up any more now that is use Teamviewer, and no HDMI connection at all present.
ROFL.... ok, this is really weird, and I'm certainly not complaining.... but it seems to have fixed itself
I really didn't do anything, rebooted a couple times trying to get at my bios settings, which I was unable to do because the display is hdmi and it doesn't give me output until I am actually in windows. I have guiminer set to open on login and auto start mining and I noticed my problem card started at 250 khash instead of the usual 70 khash while I was having this issue.... watched it for a few minutes and it climbed right up to the 500 mark alongside the 3 others, I am mining LTC at 2.1 mhash now. Dunno how / why, but I'll take it
I really didn't do anything, rebooted a couple times trying to get at my bios settings, which I was unable to do because the display is hdmi and it doesn't give me output until I am actually in windows. I have guiminer set to open on login and auto start mining and I noticed my problem card started at 250 khash instead of the usual 70 khash while I was having this issue.... watched it for a few minutes and it climbed right up to the 500 mark alongside the 3 others, I am mining LTC at 2.1 mhash now. Dunno how / why, but I'll take it
OK so in my cheapness, I have been looking and not able to find any references to needing more RAM. Can anyone point me in the right direction with a link?
tacotime's consolidated ltc mining guide thread says
tacotime's consolidated ltc mining guide thread says
Quote from: tacotime
RAM REQUIREMENTS ON-MOTHERBOARD
The equivalent amount of system RAM as for the sum of all the vRAM used by the GPUs is required when mining with reaper. For instance, 3x 5970s mining with thread_concurrency values of 8000 each would require 3x 500MB = 1500MB system ram (plus additional RAM for the OS). The memory requirements for 7xxx cards are also higher because of larger thread_concurrency values, for instance a thread_concurrency of 24000 means 1.5GB system RAM per card is required.
So it would seem that I have my required 6gb for the 4 cards and 2gb left over for windows right now. Also, if I run the miner on only my [0-0] card (the slow one) it still won't mine at a decent rate like the others even if it is the only one going. So can this really be a issue of not enough RAM? The equivalent amount of system RAM as for the sum of all the vRAM used by the GPUs is required when mining with reaper. For instance, 3x 5970s mining with thread_concurrency values of 8000 each would require 3x 500MB = 1500MB system ram (plus additional RAM for the OS). The memory requirements for 7xxx cards are also higher because of larger thread_concurrency values, for instance a thread_concurrency of 24000 means 1.5GB system RAM per card is required.
So if I have 12gb of RAM on the cards, I need 12-16gb of system RAM too?
I am getting the same issue mining btc as well, is that the same solution? (edit, same issue, but not as bad mining btc, the slow card is going about 250 instead of the 500 it should be... I guess I need more RAM!)
I am getting the same issue mining btc as well, is that the same solution? (edit, same issue, but not as bad mining btc, the slow card is going about 250 instead of the 500 it should be... I guess I need more RAM!)
You are still scrimping on RAM. The ratio of card memory to system memory is critical. Look at the Tacotime milk crate rig, 3 cards, 8GB of RAM. This system RAM to GPU Total Ram is critical when mining LTC or any scrypt based Crypto Currency.
So I just built a mining rig, mostly for ltc and maybe a little btc while I still can...
I have an AMD Sempron on a asrock 970 extreme mobo with 4x Sapphire 7950 cards, win8x64, newest drivers (played A LOT with different driver versions) 8gb system ram, 1250w seasonic gold plus psu.
The first time I fired it up, it ran great. Each card did over 500khash/s and it was stable for about an hour before I stopped it. Then I started having problems with one card not hashing after I did a reboot and I have been racking my brain all day, trying different drivers, etc...
I started with 4gb ram and just went out and got another 4gb chip because I read that the issue could be not enough ram, and that hasn't seemed to help.
I get output from guiminer that looks like
Otherwise it all looks normal / good. The dead card is technically hashing / accepting shares, but going at like 70khash while the other 3 are going around 500khash. I do not think it is a hardware issue (shouldn't have bought that extra ram ;p ) because like I said, it did hash with all 4 cards going the proper speed until I shut it down the first time. But I didn't change any drivers / software between the first shutdown and the first problems either. Any ideas??
I have an AMD Sempron on a asrock 970 extreme mobo with 4x Sapphire 7950 cards, win8x64, newest drivers (played A LOT with different driver versions) 8gb system ram, 1250w seasonic gold plus psu.
The first time I fired it up, it ran great. Each card did over 500khash/s and it was stable for about an hour before I stopped it. Then I started having problems with one card not hashing after I did a reboot and I have been racking my brain all day, trying different drivers, etc...
I started with 4gb ram and just went out and got another 4gb chip because I read that the issue could be not enough ram, and that hasn't seemed to help.
I get output from guiminer that looks like
Quote from: guiminer
“Default”: (1s):0.000 (avg):0.000h/s | Q:1 A:0 R:0 HW:0 E:0% U:0.0/m
Otherwise it all looks normal / good. The dead card is technically hashing / accepting shares, but going at like 70khash while the other 3 are going around 500khash. I do not think it is a hardware issue (shouldn't have bought that extra ram ;p ) because like I said, it did hash with all 4 cards going the proper speed until I shut it down the first time. But I didn't change any drivers / software between the first shutdown and the first problems either. Any ideas??
Jump to: