Posts by marmot
|
21)
(Message 4204)
Posted 22 Mar 2015 by marmot Post: Well, I took some compressed air and blew out my computer. Everything's running at normal temps again. If it gets too hot during summer, I'll just go from my 4.30Ghz overclock back to the stock 3.5 clocks. That makes a huge difference. Just one puff of dust on a blow out was causing between 4 and 7 degree C temp rise on the Dell 6500. The fans don't seem to be running at much lower RPM's so maybe it's the insulating capability of dust on the heat pipes? Clients never seem to believe me when I say how crucial it is to laptops to keep the fans clean and how much difference it makes in heat. |
|
22)
(Message 4175)
Posted 20 Mar 2015 by marmot Post: It got cold again for a while so the I7 and I5 machines got cranked back up. I put ThrottleStop profiles of 25W, 17W and 13W on the i5. Went ahead and told the i7 client to use 22 to 44% of CPU time slices (depending on how much the fans started running) but to lower the GPU usage the only thing I could think of is MSI Afterburner and dragging the clock down to minimum. How far down do either of you (under)clock and do you undervolt then? Do you shut down certain machines, lower the wattage on all or just pay the electric company? |
|
23)
(Message 4120)
Posted 12 Mar 2015 by marmot Post: These 8 machines did a fine job of heating the bedroom (and much of the rest of the house) but now we are hitting 60 degree days outside (yes, I still like Fahrenheit as it sets the zone of tolerated temperatures for humans on a 0 to 100 scale) and I'm setting the machines to accept no new WU's. I read someone on here was in Norway using his machine for a similar goal. |
|
24)
(Message 3979)
Posted 27 Jan 2015 by marmot Post: The hustle and bustle life of being a scientist! Congrats on the new job. |
|
25)
(Message 3740)
Posted 31 Oct 2014 by marmot Post: Does OpenCL support start cheaper and earlier on the vast majority of ATI offerings? The ATI m7740 is $60 used and has OpenCL 1.0 with double precision and was released in 2009, for example. If so, then this might bring a lot of older hardware back online for the A@H project. My particular case has me steering to the ATI for price reasons. My Dell m6500 has a dying nVidia 2800m, which doesn't have double precision and not supported on A@H, so I'm looking for a replacement. Dell shipped the m6500 with two compatible GPU's to this project. The ATI m7820 has OpenCL 1.0 support and double precision instructions and runs $100 used, while the only Kepler nVidia card (Quadro 5000m) Dell shipped on the m6500 is $400 used. The Dell m6400 ATI m7740 is $60 used and could work in the m6500 slots and was OpenCL 1.0 in 2009 with double precision. I'd love to have ATI support after looking at the pricing options on the GPU's that are assured to work in the m6500. |
|
26)
(Message 3722)
Posted 27 Oct 2014 by marmot Post:
That should have said "1.1 CC card" but can't edit/delete the message. |
|
27)
(Message 3721)
Posted 27 Oct 2014 by marmot Post: Just for info: SETI@home accepts my NVIDIA GeForce 8400 GS (511MB) driver: 266.58 Which of these drivers listed in my last message (http://setiathome.berkeley.edu/forum_thread.php?id=75978&postid=1593087#1593087) should I go with? The machine received a cuda unit but someone claims the driver I installed also caused math errors. 266.58 is not available for Quadro but 266.45 and 267.79 from the same era are available. I also have 259.57, 305.93 and 306.97 d/l so far. A blacklist or whitelist chart would be great!
Oops, I can't edit that message to correct my mistake nor delete the message. I was working without sleep at 5am then. |
|
28)
(Message 3719)
Posted 27 Oct 2014 by marmot Post: Website won't let me edit the last post so another new message... Yeah, I was posting that for BilBg (or anyone else running SETI) since he and I both run SETI packets. It seems pretty clear I was referring to SETI GPU units; sorry for the confusion. If the fact I was attached to Asteroids and it was preventing me from getting SETI GPU packets, happened to be the case (you can see this if you read further up the thread), then it would still involve Asteroids or the BOINC software. Since I finally got a GPU work unit today, this isn't in the troubleshooting decision path anymore. So obviously it's a SETI@Home issue and posting over there about it now. There's many nVidia drivers that have had problems and trying to find a driver blacklist. |
|
29)
(Message 3717)
Posted 27 Oct 2014 by marmot Post: Website won't let me edit the last post so another new message... You can see that, finally, the machine is on an older version of nVidia drivers, and still not getting packets. Even increased the SETI resource slice but no luck. Going to sleep on it and try again later. |
|
30)
(Message 3716)
Posted 27 Oct 2014 by marmot Post: I think the problem is here: http://setiathome.berkeley.edu/forum_thread.php?id=75633 Drivers post 340.52 are causing calculation errors. Going to drop back the driver versions and see if the problem is fixed. |
|
31)
(Message 3715)
Posted 27 Oct 2014 by marmot Post: So there are operations being called that are not backward compatible with CUDA 1.1 or 1.2 or is it limited to 2.0 for speed reasons? If it's option 2 is there a way to force a 1.1 CUDA card to run units? EDIT: Or is it related to the invalid calculations the post nVidia 340.xx drivers are causing on pre-Fermi cards as discussed here? http://setiathome.berkeley.edu/forum_thread.php?id=75633&sort_style=&start=20 |
|
32)
(Message 3714)
Posted 27 Oct 2014 by marmot Post: Oh, the GPU's are all 1.1 compatibility. The GPU packets received last week were from SETI and ran on the Quadro FX 2800m. For some reason the SETI project won't d/l GPU units while I'm attached to Asteroids. I'll try pausing Asteroids and doing only SETI. Update: I was able to get ID: 52516 with the 2800m to run a GPU SETI unit only after I suspended Asteroids project but it is the failing GPU. Suspending the Asteroids project on ID: 125480 with the 1600m (also 1.1 CUDA) refuses to allow a GPU unit starting. Going to delete Asteroids and reset SETI and see if I can get GPU packets back. This machine ran them all the time a few months back. Dragging the column headings out I notice that the 2800m machine is getting cuda23 applications but the 1600m isn't receiving any cuda23. I'll check over there for obsolescence but it's odd since both cards are 1.1 CUDA capable. |
|
33)
(Message 3711)
Posted 25 Oct 2014 by marmot Post: Is the project out of work for GPU? Not getting Seti GPU work either but it's set to 100 resources while Asteroids is 1000 resources. One of the 3 machines has a failing GPU (black screen o death) so set it's prefs to 999999 idle minutes and verified the server side prefs don't reflect that local choice. Verified both GPU healthy machines are set to run GPU always. Event log shows request for GPU work was made and 0 packets returned. I was receiving packets about 5 days ago. |
Previous 20