As mentioned yesterday, in a bored moment at the weekend, I experimented with running 4 x 295s and 2 x VMs on an old 9850 in a closed case just to see if I could disprove the issues referred to in the CPC mag. For a PSU and case I used a 1250W Enermax and an Antec 900 2, so credit to Christopher N. Lew for pointing out that the new case has 8 expansion slots. I am a big fan of the original Antec 900 but didn't realise the revised model had an additional slot All went well with the GPUs running at or under 80 degrees C as long as I didn't overclock the shaders and try and run the VMs at the same time. I could only do one or the other. I kept a close eye on things and this state of affairs went on for about 24 hours until a string of 511 pointers and a hot day combined to drive GPU temps to above 85 degrees. Even running in open case mode the temps continued to rise and so I reluctantly concluded that I was just not going to defeat the high ambient temps. At this point I ran for the hills, screaming and waving my arms in the air i.e. lost my nerve and removed a card. The rig is now running closed again with 3 x 295s. GPU temps are at around 65 degrees running a combi of 768 and 1888 pointers. I probably should have used a cooler running CPU but it would seem that either watercooling or a/c would be required to run successfully in the conditions I experienced. Oh well, at least it was fun
Rabble rabble rabble stock cooler rabble rabble I was actually looking to do something similar, but with 285's instead...hmmm... What about this or this with a few more fans whacked on?
Good experiment mate, did the 1250w PSU power all 4 295's ok? Just wondering as I have a 1250w CoolerMaster and wasn't sure if it will power 4 cards or not...
My hat's off to you for attempting this, well done! I can't imagine how hot that thing must run with even a modest overclock on it. I could almost cook an egg on the roof of my quad card Antec 300 rig, and that's only overclocked 8800's
Stupid question time - don't suppose you felt the power supply casing while it was running the 4x GTX295's. How hot was it?
Wibble. The power supply wasn't a problem. As mentioned on Jack's thread it performed admirably throughout, with the rig reaching a peak of 1140w from the wall with 2 VMs running. This was running mostly power hungry 511 pointers but with no overclock. Can't really comment on the Coolermaster equivalent and wouldn't want to say anything derogatory about it anyway considering your username Thanks Uni. Glad to hear your kit is up, running and chimping again I was monitoring everything very closely and so yes I did Jack and it was OK to touch but quite hot. There again, so was the entire room. I deliberately tried the experiment with minimal room ventilation to simulate what the conditions would've like if I was at work. I wasn't too worried because I've never had a Enermax, Corsair or a Be-quiet (the only PSUs I ever use) take out itself or any other hardware. Each time I've push them too hard they've cut off and then recovered without any damage. Of course, I may just have been lucky so far
That speaks to me! In a way, that says a lot. Certainly compaired to that Thermaltake 1500W which was described as being too hot to touch its case > 1100W. That's pretty much the PSU list engrained into my head. But I have this "compatibility" niggle with Enermax. Some boards didn't like the "Coolguard" technology on the Infiniti. Kept the +12V line active on the board after switch-off to run the fans for a couple of mins, but next re-start it would cause a "failed overclock" on the mb. And recently I've read of problems again with the latest Revolution - specifically, not working with the EVGA X58 Classified mb. Not that I have any real problem with Enermax, I've been a fan, but I don't like reading stuff like that. A power supply is a power supply. It should just work. Compatability, blah .... (Too simple a view, I know.)
Yep, I had at least 4 failures of Enermax Infiniti 720W's in different PC's due to Coolguard faults - all worked fine for about 6 months of 24/7 load, and then they died. Currently using Corsair HX1000's - no problems (so far) .....
I have been considering what PSU's to use in the farm when I start building the next couple of rigs for it, and have been thinking Corsair HX1000's or FSP Everest Pro 1200's. I want the rigs to be upgradeable (card wise) without needing to change the PSU. Must post a reply to my farm thread about this tomorrow (well, later today )
Well, last Tuesday, the rig with just 3 x GTX295s started shutting itself down every couple of hours even when temps were quite low. I swapped out the Enermax for a spare Corsair HX1000 and things have been fine ever since. I guess it's about time I had my first power supply failure but I'm wondering whether I've just been unlucky or whether there's something wrong with the new Enermax. It's twin, in an almost identical rig, has been running fine since I bought it two weeks ago. Ho hum, RMA time. Having read JackOfAlls dreadful P6T7 experience, I should consider myself lucky it didn't take anything with it when it failed.
Oh dear, don't say that. I got a 1250W Revolution on Friday. Scan had it in the 'Today Only' on Thursday for £205 inc VAT. Seemed like a good idea at the time!
The twin rig with the other 1250w Enermax continues to run fine. I'm hoping the first was just bad luck or else we're both in trouble.
