I haven't heard of this issue before. There's no particular reason for any GPU to behave differently than the others. A bit more detail might help me or other posters figure it out. What cards are you running? What pool and what settings? Are you using OpenCL or CUDA? I have a problem with GUI miner on one of my dedicated rigs
I have four identical cards running, all with identical settings
GPU1 and GPU2 are having connection problems intermittently while GPU3 and 4 are solid.
What would cause this?
I'm having the same issue.
I'm reciving "Connection Problems" error as well.
The thing is it will work on launch then stop and give the error than maybe pick pick up again in 20sec and work for lie 5 sec and the give the error again.
Its like on and off.
It was working great since I got it till today. No clue what is going on.
INFO:
1 GPU (5870) with SDK 2.1
Just got the latest version of the GUI and same stuff is happaning although it went from 360 Mhashes/sec to 402!
OH MY GOD ME TOO! http://bitcointalk.org/index.php?topic=7542.0This problem is driving me crazy! PLEASE TELL ME IF YOU FIGURE OUT HOW TO FIX THIS! |
|
|
|
Most excellent. I just bridged 1 and 2, enabled crossfire on those two, then plugged in 4 and I'm now mining on all 4 cards. Only 2 odd things: - MSI Afterburner won't start due to "Unable to initialize driver wrapper" so I can't overclock as much (I can only go to 840)
- The hash rates of the bridged GPUs is very unstable. they bounce between 250 Mhash/s and 315 Mhash/s
That said, I'm extremely happy to not have a $250 card laying at the bottom of my case doing nothing.  Thanks! Jere Yeah, well are you getting any weird connection issues? What mining client you using? |
|
|
|
I've seen you post in other threads saying that you couldn't get 4 cards to work at all. Is this the same setup? Is GPU-Z working for you now?
What did you do to fix that?
Jere
I've been having a lot of problems getting 4 GPUs working I got all 4 to run by putting GPU's 1 and 2 in crossfire mode with a bridge. Suggestion by another forum post, and it worked. poclbm worked I thought I was having this problem only on one computer, but I got the "connection problem" on my other dedicated miner that was identical, but it had been mining stable for 48 hours (although no GPU driver crash. just GUI miner for now)... I don't get why this is happening. I've never seen the error before and its causing havoc lol |
|
|
|
Try removing one of your cards entirely and see if it is stable like that.
I've had problems trying to run 4 6950's and pulling one card stabilized it completely. I'm wondering if you are seeing a similar problem.
Jere
Nope. First two GPU's again, stopped mining. |
|
|
|
Ok I think this might be a pool problem
I recently bumped up production by a lot and am having problems with my miners doing this:
"Problems communicating with bitcoin RPC"
I have 4 GPUs mining on the same worker name.
Should I be splitting this up? I'm getting very angry with all the small pauses. Is this DNS related or anything? Firewall is disabled.
No, you won't hit the worker limit with only that. If this ever happens with you, you'll see a message in your miner recommending to use no more that 20 LP clients on one worker account. As I understand, you have this problem on one of your mining rigs, but everything works fine on another ? It just started on my second dedicated machine. I'm wondering if I'm having a transmission issue? |
|
|
|
|
Ok I think this might be a pool problem
I recently bumped up production by a lot and am having problems with my miners doing this:
"Problems communicating with bitcoin RPC"
I have 4 GPUs mining on the same worker name.
Should I be splitting this up? I'm getting very angry with all the small pauses. Is this DNS related or anything? Firewall is disabled. Should I be using VPN?
|
|
|
|
Try removing one of your cards entirely and see if it is stable like that.
I've had problems trying to run 4 6950's and pulling one card stabilized it completely. I'm wondering if you are seeing a similar problem.
Jere
Doing this now. Pulled one card, seeing how it goes. |
|
|
|
|
Preface info:
MSI 890FXA-GD70 - motherboard with room for four double-wide GPU
Athlon II X2
2GB DDR3
Caviar black HDD
Four Radeon 6870
PSU: Corsair HX1000 and Ultra X3 1000W (testing both)
Windows 7 64bit / Windows 7 32 bit (both tested)
Catalyst 11.4
Clean install of windows, installed drivers, and mining software
Note: i have another computer identical to this. I bought enough to build two identical machines
Problem:
To get all 4 GPUs working, I have to Crossfire GPU1 and GPU2, GPU3/4 are dummy plugged
I can start mining on all 4 GPUs and it works like a charm at 1.1Gh/s just like the other rig. This one after ~10 minutes or so, GPU1 and GPU2 will have a problem
"Problem connecting to bitcoin RPC" is what I get under GUIminer
poclbm , the first two clients just lock up. The last two clients keep going
diablominer: mines on all 4 GPUs for about 10 minutes then drivers crash, then runs at half speed
rpcminer-openCL: I can only get it working on two GPUs
I swapped cards into different order, and its still the cards in PCIE 1 and PCIE 2 that fail first
I swapped around the PSUs, reinstalled windows, re-did drivers. I'm thinking this might be a motherboard problem now, but I don't know what to test.. What do you guys suggest?
|
|
|
|
Have you tried moving your dummy plugs around?
No I have not... I'll try |
|
|
|
If you can't use all 4 gpu's it's probably because you need to do it on windoes 64. It does it right for me for all my max overclocked nvidia fermi cards x 4.
...I AM on 64bit. Read the first post. |
|
|
|
|
What are the performance arguements with this miner? I loaded up the rpcminer-opencl on Radeon 6870's and its only giving 200Mh/s where as poclbm did ~260
Also, how do i dictate which GPU the device runs on so I can run multiple instantces. Right now it only runs on the first GPU out of four
|
|
|
|
|
I have a problem with GUI miner on one of my dedicated rigs
I have four identical cards running, all with identical settings
GPU1 and GPU2 are having connection problems intermittently while GPU3 and 4 are solid.
What would cause this?
|
|
|
|
Really all the overclocking forums are like this. If you see cheesy images or numerous lines of variously-sized and colored text in the signatures of the users, you can bet the community is clueless and absolute trash.
I can somewhat agree to this. |
|
|
|
I'm keeping hush until I've made a decent amount of money (to payoff small investment) so that I can say its at least worth it for some and not look like an idiot. Also about buying it for use on other stuff, lot of us travel often  |
|
|
|
How exactly does the "whole thing go to crap"?
Does your OS bluescreen or something?
When trying linux, the machine won't boot. All I get is a black screen. In Windows, the following happens: - GPU-Z reports the OpenCL bug
- MSI Afterburner won't start due to "Unable to initialize driver wrapper"
- rpcminer-opencl crashes immediately
Jere EXACT thing here I load up linux with 4 cards installed and I get a black and purple flashing screen.. probably the same thing however, given the stupid shit that linux does sometimes |
|
|
|
I have built a mining rig with 4 Radeon 6950's and I'm having no end of problems.
The first problem was heat. Those puppies just sit so close together that there was no room for air to move and it sounded like a jet aircraft before I even started mining. So I took out two gpus and ran it on air cooling until the water cooling equipment showed up.
This was my first water cooling system so it took me a while to setup. The forums at Overclock.net were extremely helpful there. But after a day or so, it's all physically setup.
After installing Windows 7 (32bit) without issue, everything seemed extremely stable and quiet. Gotta love quiet.
I installed the guiminer and it wouldn't even start. Neither would MSI Afterburner. GPU-Z would start but with a dialog saying that there was a "Bug in the OpenCL driver. OpenCL checks will be skipped".
I spent the next 4 days (I'm on vacation) trying various drivers and even a variety of Linux distributions with no luck.
Right now, I am only running 3 cards and everything is fine. But if I install that 4th card then the whole world goes to crap.
My question is: Has anyone gotten 4 Radeon 69XX cards to run together harmoniously?
My contribution to the community is this:
1) XFX Radeon HD 6950 1 Gig cards (HD-695X-ZNFC) are not reference cards for water cooling purposes. They are however laid out exactly like the 6870 reference cards so you can get a full coverage block for a reference 6870 and it fits perfectly.
2) If you have cards that aren't plugged into monitors and they show up as "disabled", what you need is a dummy plug. It is, in essence, a DVI to VGA adapter with 3 resistors plugged into it. Google "DVI Dummy Plug" and you'll find lots of instructions.
3) The three cards I have running are overclocked to 920Mhz and produce around 340Mhash/s each. The hottest card runs at about 60C.
Thanks,
Jere
How exactly does the "whole thing go to crap"? Does your OS bluescreen or something? Poclbm, guiminer, phoenix miner, none of the miners will launch I conform 4 GPU works on windows, its just 4 different cards dont work (if 4 pcie slots full=fail) I have a Radeon 5970 and two 6870's running in one rig right now churning out over 1.2 Gh/s.. Four GPU works. its just when I use 4 cards/PCIe slots, everything breaks |
|
|
|
You sir are a god damned idiot.
Although I agree that the PSU is more than powerful enough, this kind of reply doesn't exactly encourage people to try to help you. Sorry, its a little frustrating when I'm topping 10 hours into working on this software problem and I cant even get Linux to install the ATI/Radeon display driver correctly on a different card. I've been trying to figure this out since ~330pm, its just past two am now. |
|
|
|
|
How do I change the PW's after running the script
|
|
|
|
|
Show Posts
