Anyone who can help me out with this... the 7970s is what I'm most concerned with.

I used to GPU mine a lot back in the day. Even mined 10 BTC with GPUs. Then Scrypt which was nice for a quite a while. Then ScryptN and some machines on X11 where full power was an issue. Around 9 months ago with a bunch of maintenance stacking up and profitability... well... not existent. I got a little burned out. So I shut it all down. I have contemplated selling my 7970s and 5970s (I found 5970s to perform equal to 7970s during Scrypt days so I got quite adept at hoovering up a bunch of them and tweaking them out) for a while but I think I'm ready to crank them up again. I tried your Tahiti binary files wolf0 and am fairly certain that I followed everything correctly. I get about 3.3Mh/s per 7970 during normal operation. When I then copy over the wolf0 bin everything loads to the point of looking for work but then the program crashes and the terminal closes suddenly. The only thing I haven't tried is upgrading the AMD drivers (I'm still running 13.x). Could that be the problem? As per usual I have a built in resistance to upgrading Catalyst when I have an otherwise functional system that has run perfectly for years now

Hi Guys,
was just surfing around looking for the best X11 info. I used to mine BTC with my GPUs, then Scrypt which was REALLY fantastic for a while (multipool!) but got Goxed out of a lot of that (stupid me). Did some X11 and Scrypt N mining for a while. Then just got burned out by low profits and constant maintenance (40ish GPUs in 9 machines). Feeling better now and I think I'll get back in the saddle. Is the SGminer fork posted back in March still the best bet? I am looking at the X11 types because the lower GPU temps, over time, should save wear and tear over BTC, Scrypt and Scrypt-N etc. I actually love GPU mining (even wrote a little article about it http://actualeconomy.com/2014/03/ode-to-gpu-mining/) and hope that it becomes more viable again soon and stays around. Though if 7970s ever go for $400 a pop again I might have to cash out Look forward to your comments! Oh and send any good 7970 configs you have my way. I am running 13.x drivers... should I update? I know that there are new ones now... but I also know that that doesn't mean they are better.

Thanks!
Rich

For the folks who are using multiple PSU and complaining that the 2nd PSU is shutting down I am pretty sure this is not because you are overloading the PSU with the card but that you are sending excessive current to the primary PSU and tripping the internal breaker. I've fried a card by having it run on the 12v MB power of the primary and the 12v aux power of the secondary PSU. If, for example, we have PSU1 running at 12.3volts and PSU 2 is running at 11.9V that difference, will cause current to flow from one PSU to another. There may be other factors with combining of multiple PSU rails that could cause this problem as well. The last time I accidentally set something like this up. The secondary PSU cut off and afterward I noticed that the IEC power cord that goes to the wall was was HOT and limp as a noodle. It then subsequently cooled and was RIGID... Like a stick and you could snap and shatter the wiring that had re-solidified. The PSU was fine, the cable was toast. There is a fix for this though. I would recommend that if you use 2 PSUs you use powered risers connected to the secondary PSU that is providing the Molex connected aux power. On those powered risers sever the 12V connections to the motherboard. Pin 1-3 on side B (I think) and 2-3 on side A. Double check those are the right pin somewhere. I always forget which side is A and B and have to look at a picture or riser I've already modified.

Sure... if you can find one the Sapphire 7950s are some of the best Scrypt miners. The reference model! If you find any for sale new please share.

I am looking for this type of cooler if anyone has any they'd like to sell or if you have a cooler and 5970 I'll buy that too.

Thanks,
rich

I've had several XFX 7970 fans go on me. If yours are just kind of sticky and won't spin I use some teflon spray I got at Lowes and it brought them back to life, like new. Also used a product that is a little more specialty called  de-oxit, which is for electronics cleaning, also did the trick once when my teflon spray wasn't handy. Now a little teflon spray periodically is standard procedure for routine maintenance.

Hi all... Big thanks to the community so far for all the help with my rig and getting CGminer working happily. So I have a rig with 4 7970s. two are ASUS, one is a Gigabyte and the last is an XFX. They have varying clock speeds. In the picture below the two ASUS cards @ 925 MHz are GPUs 0 and 1. I think GPU 2 is the XFX and 3 is the Gigabyte. When bitcoin mining they are perform about the same when clocked close to each other. As you can see everything is weird below. There is a 70 or so Khash/s difference between the identical GPU 0 and 1 and a huge difference between 2 and 3. GPU 4 is the onboard APU giving some token Khash. They all report being at max usage, 98-99% activity. Running for an hour now, no HW errors. Here are my cgminer settings:


As a note the --auto GPU isn't doing much I don't think as the range for GPU 0 and 1 is 925-925 etc. I just stole these setting from someone else

Hi,

I have 4 of these boards and that PCI 1x number 3 does not work on them either. Don't know what's up with that. Furthermore, I think 4 cards per mobo is the max right now. It is all I can get to work in Windows 7. Looks like in the forums folks are saying that is the max the drivers will allow. Have not really see a work around. I have not had any problems at all with the onboard GFX. I have them running on all the boards for a nice little 80k/Hash bump with a 6800. I put the IG bios settings to Force on and PCI NB although on some I force the IG to be the display and that works fine too so I don't know what to tell you there. These boards are pretty solid though. I mined Bitcoins for a while with them running 3 7970 off the board before getting powered risers and beginning litecoin mining. Was very disappointed when I put together my 6 card uber rig with 2 PSUs and it looked so beastly and cool only to have only 4 cards ever usable and to find out on my own that that 3rd 1x slot is a lie. I've since found this report in several places. For future builds I'm going to scale back to older tech I think and buy used cheaper mobos. The 80khash from the APU is a cute little bonus but I can shave off over $100 like I did with a pentuim 4 dual core rig I have that runs great for $55 for the mobo and cpu. DDR2 ram though. Don't know if that will negatively affect LTC mining yet. Let me know if you hear anything from Gigabyte about this. Also let me know if you find anyone that has more than 4 cards running in Windows 7 or 8 or hell I'll even try linux if I can get it working. PS [if you count the APU I do have 5 cards going ]

what could be the problem...? I used a dirt simple cgminer command script to start with. Didn't tweak anything really once it was open except turning off GPU2. Here is the script

cgminer --scrypt -o stratum+tcp://ltc-stratum.kattare.com:3333 -u VLABlD.1 -p x -I 13 -g 2 -w 256

This rig has been mining bitcoin for about 2 months straight with no issues using BitMinter's JAVA platform and it has been mining with the current GPU settings (via afterburner) for about 3 weeks.... no problems, no crashes.

I have 16 GB of RAM in the machine and 4 7970 with 3GB each. No driver crashes to my knowledge while cgminer was running. I'm switching over to LTC because I have about 12 Giga Hashes of GPU miners and the difficulty of BTC is to the point now that the power cost of the GPUs is making it hard to make any money. I guess the ASICs are finally coming online en mass.

Any ideas on what the HW issue could be? I have the stuff back mining Bitcoin for the moment and, again, working flawlessly.

Thanks guys.

I'm a noob so I'm stuck here in the newbie section. Here's an image of my cgminer (top) and my burnside's pool stats (bottom). This also happened with coinotron pool and I was using a different machine with them. Can't get my LTC mining to read over on the pool side. What's happening to all of my hash power? As you can see in the screenshot the machine is doing over 2 MHps (LTC) but burnside says I'm only doing like 60 KHps?? I let it run overnight to see if it was just not a realtime stat or something but I checked just now (6hrs later) and it is all the same. It now reads about 120 KHps at the pool which is about 2100KHps off. What is wrong?