DiabloD3 (OP)
Legendary
Offline
Activity: 1162
Merit: 1000
DiabloMiner author
|
|
September 13, 2012, 11:41:05 PM |
|
".. or immediately if connection is refused" ---you mean for example "...cant connect to server.." ? some times some pool have "...cant connect to server.." 20 sec in my test (( 100 minutes was chosen because it is optimal as it is the average length of downtime spent on unplanned maintenance inside of a data center. ---- so i can or not reduce 100min? for me is very important (( After what time diablo go back to first pool ? Pools that don't use passwords allow anything in the password field.--- i must use "-p first pas," or "-p first pas" -with comma or not? will you support "asic mining" in future? I and many people hope , that you will support. Thx for realy good product! DiabloMiner won't switch if the pool is still up. As long as your pool never goes down, you don't have an issue. -p first,xxx works for most people DiabloMiner will never support fpga or asic mining unless someone makes raw USB access not suck in Java. If you want that, use cgminer, it's essentially DiabloMiner ported to C.
|
|
|
|
Elmojo
|
|
September 26, 2012, 01:41:56 PM |
|
Hey folks, need a little help here. I'm running Diablo on my 7970 and it's been working great, up until recently. For background, my card is water cooled, and runs at 1187/850, yielding about 715mH/s on average. It's been super stable, until... About a week ago, my miner suddenly stopped. I tried restarting it, but it just crashed every time. I then noticed that Afterburner and GPU-Z were no longer reporting the clocks, they both just said 0.0000. At this point, I spent several hours trying to reinstall my GPU drivers, which was a nightmare in itself. I was FINALLY able to get the AMD drivers installed by stripping the old drivers out, sweeping the remnants with Driver Sweeper, and reinstalling the Beta 12.9 drivers. I tried unstalling 12.8 WHQL, but they wouldn't run any of my miners, just kept giving errors. At this point, I can mine, but Diablo crashes much more frequently than before. The big problem is that when it does crash, I can't restart it. When I try, it says something about "no OpenCL devices found". I have to actually reboot the system, at which point it will go back to mining, for a while. Not sure where to go with this, but it's getting awfully frustrating and is costing me money in lost revenue. Anyone got any ideas where to start? Thanks in advance, Elmojo
|
|
|
|
Elmojo
|
|
September 26, 2012, 02:01:19 PM |
|
Elmojo,
If your card has two BIOSes, switch to non-writable one, boot see if GPU-Z can read it. Backup that BIOS, switch to read/write BIOS, flash it with backup. See if this stabilizes your system. I've had similar issues, I flashed all my cards with BIOS images from techpowerup.com
I had to switch the BIOS way back when I first got the card. I'm not sure which one it's set on now. At the time, there were no BIOS images for that card, but I'll look and see if there's anything new. Do I want a particular BIOS or just any of the ones listed that match my card model and brand? Which tool is best for flashing the HD7970?
|
|
|
|
DiabloD3 (OP)
Legendary
Offline
Activity: 1162
Merit: 1000
DiabloMiner author
|
|
September 26, 2012, 03:15:51 PM |
|
Hey folks, need a little help here. I'm running Diablo on my 7970 and it's been working great, up until recently. For background, my card is water cooled, and runs at 1187/850, yielding about 715mH/s on average. It's been super stable, until... About a week ago, my miner suddenly stopped. I tried restarting it, but it just crashed every time. I then noticed that Afterburner and GPU-Z were no longer reporting the clocks, they both just said 0.0000. At this point, I spent several hours trying to reinstall my GPU drivers, which was a nightmare in itself. I was FINALLY able to get the AMD drivers installed by stripping the old drivers out, sweeping the remnants with Driver Sweeper, and reinstalling the Beta 12.9 drivers. I tried unstalling 12.8 WHQL, but they wouldn't run any of my miners, just kept giving errors. At this point, I can mine, but Diablo crashes much more frequently than before. The big problem is that when it does crash, I can't restart it. When I try, it says something about "no OpenCL devices found". I have to actually reboot the system, at which point it will go back to mining, for a while. Not sure where to go with this, but it's getting awfully frustrating and is costing me money in lost revenue. Anyone got any ideas where to start? Thanks in advance, Elmojo
Wow, I wonder if your VRMs are overheating. My 7970 sometimes overheats to the point that the machine hangs on boot at 1125 on really hot days, so maybe thats what happened with you.
|
|
|
|
Elmojo
|
|
September 26, 2012, 03:21:35 PM |
|
Wow, I wonder if your VRMs are overheating. My 7970 sometimes overheats to the point that the machine hangs on boot at 1125 on really hot days, so maybe thats what happened with you.
I guess it's possible, but if it was going to overheat, wouldn't it have done it during the summer? My office stays a bit warm (78F+/-) but is pretty constant throughout the year. AB and GPU-Z report full burn temps of around 74c, but maybe they aren't monitoring the VRMs?
|
|
|
|
Joshwaa
|
|
September 26, 2012, 03:23:44 PM |
|
Also on the 7970's in AB you can control your mem volts. I have mine all the way down and my mem stays relatively cool. Just a thought.
|
|
|
|
DiabloD3 (OP)
Legendary
Offline
Activity: 1162
Merit: 1000
DiabloMiner author
|
|
September 26, 2012, 03:23:50 PM |
|
Wow, I wonder if your VRMs are overheating. My 7970 sometimes overheats to the point that the machine hangs on boot at 1125 on really hot days, so maybe thats what happened with you.
I guess it's possible, but if it was going to overheat, wouldn't it have done it during the summer? My office stays a bit warm (78F+/-) but is pretty constant throughout the year. AB and GPU-Z report full burn temps of around 74c, but maybe they aren't monitoring the VRMs? Unless your card has a really retarded heatsink setup, 74c is fine, I wasn't having issues until I was >80c for 8 hours at a whack.
|
|
|
|
Elmojo
|
|
September 26, 2012, 03:27:17 PM |
|
Also on the 7970's in AB you can control your mem volts. I have mine all the way down and my mem stays relatively cool. Just a thought.
I don't see anything about mem volts in AB. In fact, AB has never really worked as advertised for me. With unofficial overclocking in mode 2, I can get the core clock to where I want it, and the mem clock down to about 650 before it gets unstable, but I never have been able to get anything related to voltage working. @DiabloD3: The water block I have seems to be working fine. I wonder if the lack of fans on the VRMs is affecting temps? Shouldn't the water block cool the VRMs also?
|
|
|
|
DiabloD3 (OP)
Legendary
Offline
Activity: 1162
Merit: 1000
DiabloMiner author
|
|
September 26, 2012, 03:30:45 PM |
|
Also on the 7970's in AB you can control your mem volts. I have mine all the way down and my mem stays relatively cool. Just a thought.
I don't see anything about mem volts in AB. In fact, AB has never really worked as advertised for me. With unofficial overclocking in mode 2, I can get the core clock to where I want it, and the mem clock down to about 650 before it gets unstable, but I never have been able to get anything related to voltage working. @DiabloD3: The water block I have seems to be working fine. I wonder if the lack of fans on the VRMs is affecting temps? Shouldn't the water block cool the VRMs also? I think consumer/prosumer/enthusiast watercooling is pointless. That said, there is absolutely no reason you shouldn't have a full block on there that covers everything: gpu, memory, and VRMs. If there are any chips exposed at all and they dont have normal air cooled heatsinks with a fan blowing on them, you have a problem.
|
|
|
|
Elmojo
|
|
September 26, 2012, 03:38:22 PM |
|
I think consumer/prosumer/enthusiast watercooling is pointless. That said, there is absolutely no reason you shouldn't have a full block on there that covers everything: gpu, memory, and VRMs. If there are any chips exposed at all and they dont have normal air cooled heatsinks with a fan blowing on them, you have a problem.
Not pointless at all if you work in your office 8-10 hrs per day and don't want a turbine screaming in your ear all day! My rig is heavily overclocked, but is quieter than my ceiling fan on low, so I can work or play in comfort. The water block I have is full coverage, so I doubt that's the problem. Is there any way to measure the temps of the VRMs?
|
|
|
|
Joshwaa
|
|
September 26, 2012, 03:59:25 PM |
|
GPU-z should show VRM temps. In afterburner press the arrow to the right of volts should pop up the adjustment bar.
|
|
|
|
Elmojo
|
|
September 26, 2012, 04:13:05 PM |
|
GPU-z should show VRM temps. In afterburner press the arrow to the right of volts should pop up the adjustment bar.
Ok, I'll check out both of those options when I get home, thanks!
|
|
|
|
crazyates
Legendary
Offline
Activity: 952
Merit: 1000
|
|
September 26, 2012, 04:29:41 PM |
|
Wow, I wonder if your VRMs are overheating. My 7970 sometimes overheats to the point that the machine hangs on boot at 1125 on really hot days, so maybe thats what happened with you.
I guess it's possible, but if it was going to overheat, wouldn't it have done it during the summer? My office stays a bit warm (78F+/-) but is pretty constant throughout the year. AB and GPU-Z report full burn temps of around 74c, but maybe they aren't monitoring the VRMs? VDDC (VRM) temps are located on the sensors tab in GPU-z, usually toward the bottom. Also: https://bitcointalk.org/index.php?topic=102167.msg1138807#msg1138807
|
|
|
|
DiabloD3 (OP)
Legendary
Offline
Activity: 1162
Merit: 1000
DiabloMiner author
|
|
September 27, 2012, 02:21:31 AM |
|
Jesus. Those VRMs AMD uses are usually safe up to 120c for 24/7 usage (the same way GPUs are safe up to 85c for 24/7 usage), hes really trashing them. Also 90% fan? More like dead fan.
|
|
|
|
crazyates
Legendary
Offline
Activity: 952
Merit: 1000
|
|
September 27, 2012, 02:31:16 AM |
|
Jesus. Those VRMs AMD uses are usually safe up to 120c for 24/7 usage (the same way GPUs are safe up to 85c for 24/7 usage), hes really trashing them. Also 90% fan? More like dead fan. That's my thread, and in my last post, I mention that I finally got undervolting in MSI AB to work, and now my fan is much lower, with much lower VRM temps.
|
|
|
|
Elmojo
|
|
September 27, 2012, 03:55:20 AM |
|
That's my thread, and in my last post, I mention that I finally got undervolting in MSI AB to work, and now my fan is much lower, with much lower VRM temps.
How did you get the voltage working in AB? I haven't been able to get that figured out yet.
|
|
|
|
crazyates
Legendary
Offline
Activity: 952
Merit: 1000
|
|
September 27, 2012, 04:48:10 AM |
|
That's my thread, and in my last post, I mention that I finally got undervolting in MSI AB to work, and now my fan is much lower, with much lower VRM temps.
How did you get the voltage working in AB? I haven't been able to get that figured out yet. It was the "Force Constant Voltage" setting. Before I checked it, changing the slider did nothing.
|
|
|
|
Elmojo
|
|
September 27, 2012, 12:06:18 PM |
|
It was the "Force Constant Voltage" setting. Before I checked it, changing the slider did nothing. Cool, I'll give that a try, thanks!
|
|
|
|
Elmojo
|
|
September 27, 2012, 12:16:10 PM |
|
VDDC (VRM) temps are located on the sensors tab in GPU-z, usually toward the bottom.
I monitored the VRM temps with GPU-Z overnight. At full burn, VRM1 is at 70c and VRM2 is at 58c. The GPU temp is currently at 61c. I think it was as high as 71c last night while I was playing a game on the other card. Those temps all sound fine to me, what do you guys think?
|
|
|
|
Joshwaa
|
|
September 27, 2012, 03:41:27 PM |
|
Those temps are fine.
|
|
|
|
|