Message boards :
Number crunching :
Gigabyte GA-EP45-UD3P Ver 1.6 MOBO with EVGA GTX-750TI SC...
Message board moderation
Previous · 1 · 2
Author | Message |
---|---|
TimeLord04 Send message Joined: 9 Mar 06 Posts: 21140 Credit: 33,933,039 RAC: 23 |
At last check, (been a few months since I've looked), the CUDA Driver from NVIDIA for MAC OS X was 7.5.26; this is what I'm on. I still encounter "Validation Inconclusives" and "Invalids" on CUDA65; but NOT as many as when I was stock on OpenCL. The one "Error" I still have showing is from a "Ghost" that was created when I switched from OpenCL to CUDA65 and my entire cache got wiped. TL TimeLord04 Have TARDIS, will travel... Come along K-9! Join Calm Chaos |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
Take your current steady state inconclusives and divide by pendings. Inconclusives / Pendings < 5% = good. >= 5% = questionable. >10% = there is a problem somewhere. "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
TimeLord04 Send message Joined: 9 Mar 06 Posts: 21140 Credit: 33,933,039 RAC: 23 |
Take your current steady state inconclusives and divide by pendings. Inconclusives / Pendings < 5% = good. >= 5% = questionable. >10% = there is a problem somewhere. Well, looking yesterday afternoon, Out of 200 Units in queue there were 28 Pendings, (now 33), 8 Inconclusives, (now 7), and the one Error mentioned... Yesterday, then, was 3.5% and today is 4.71%. This seems to be the new "Normal" for my system since switching to CUDA from OpenCL. [EDIT:] Forgot to include: 2 Invalids showing on recent numbers... [End EDIT.] TL TimeLord04 Have TARDIS, will travel... Come along K-9! Join Calm Chaos |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
Take your current steady state inconclusives and divide by pendings. Inconclusives / Pendings < 5% = good. >= 5% = questionable. >10% = there is a problem somewhere. Nice. You're golden dude :-D ------ > Thread closed ? "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
TimeLord04 Send message Joined: 9 Mar 06 Posts: 21140 Credit: 33,933,039 RAC: 23 |
Take your current steady state inconclusives and divide by pendings. Inconclusives / Pendings < 5% = good. >= 5% = questionable. >10% = there is a problem somewhere. I have no complaints; was just clarifying things for Stubbles69/Rob. KUDOS to TBar for his hard work!!! :-) Thanks again, everyone. <<--- Still feel STUPID about that rubber protector thing.... :-O Thread can be closed, unless someone else would like to comment. TL TimeLord04 Have TARDIS, will travel... Come along K-9! Join Calm Chaos |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
My only comment is that this demonstrates a lot of things we do need to care about. Accuracy & precision are important (though different). We learned that to have an idea of whether something is working right or not, you need a reference. For the general case inconclusive/pending works, just because pending will average itself depending on project health. So, for example, you are less than 5% mutated than project average, if your inconclusives/pendings is less than 5% "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
LOL....that IS a funny missing the cap on the connector. And I am laughing with you, not at you. I would be fibbing if I did not admit that over the years I have had one or two similar mishaps by not having my eyes open or my thinking cap on. Trying to mate the wrong PSU connector with a mobo socket, etc.. And then giving myself the big ol' facepalm when I figured out what I was trying to do. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
At last check, (been a few months since I've looked), the CUDA Driver from NVIDIA for MAC OS X was 7.5.26; this is what I'm on. I still encounter "Validation Inconclusives" and "Invalids" on CUDA65; but NOT as many as when I was stock on OpenCL. The one "Error" I still have showing is from a "Ghost" that was created when I switched from OpenCL to CUDA65 and my entire cache got wiped. Sometimes the Driver is updated a couple times a month, sometimes it goes a couple months without being updated. It's Real easy to check. Just go to the Dock and open System Preferences. Then open CUDA. Why didn't you do that simple check before posting? This is the current driver, http://www.nvidia.com/object/macosx-cuda-7.5.30-driver.html I just had a problem with the latest Special App and the 750Ti, it was solved by updating the driver. You should try it. |
TimeLord04 Send message Joined: 9 Mar 06 Posts: 21140 Credit: 33,933,039 RAC: 23 |
At last check, (been a few months since I've looked), the CUDA Driver from NVIDIA for MAC OS X was 7.5.26; this is what I'm on. I still encounter "Validation Inconclusives" and "Invalids" on CUDA65; but NOT as many as when I was stock on OpenCL. The one "Error" I still have showing is from a "Ghost" that was created when I switched from OpenCL to CUDA65 and my entire cache got wiped. Thanks for that, TBar! :-) Just Updated to CUDA 7.5.30... We'll see tonight how it goes. :-) TL [EDIT:] Also noticed that the system seems to lag, (especially in Firefox), since adding the second 750TI SC... (Maybe a system memory issue???) System only has 8GB, DDR2 800MHz RAM... Just ordered new Patriot 4GB RAM sticks totaling 16GB, (4 sticks), also at 800MHz. New RAM should be here just before the WOW Event starts up. Hopefully this improves system stability issues. (Memory Leak(s) maybe???) Footnote: Nothing has crashed, yet... TL [EDIT 2:] The new Patriot RAM was purchased through EBAY and was half the price of my local Fry's Electronics store in Concord, CA. Couldn't find anything through Amazon that did NOT have ECC on it... This MOBO does NOT use ECC... So, I ordered through the EBAY source. $164.24 for four 4GB DDR2 800MHz sticks of Patriot RAM. TL TimeLord04 Have TARDIS, will travel... Come along K-9! Join Calm Chaos |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
I'm guessing you're still running a CUDA 5.5 task and CUDA 6.5 task at the same time? That probably won't work very well. You should only run One type of task at a time. My testing shows the best results are with running One task per GPU with the -poll option on the Baseline App. The Special App uses a full CPU already so the -poll option offers no advantage. On the Baseline Apps the -poll option speeds up the SINGLE task quite a bit. I'm running three cards with 6 GBs of RAM, No Lag here. If you were using the CUDA 7.5 App your app_info.xml set to use the -poll option would look as this; <app_info> <app> <name>setiathome_v8</name> </app> <file_info> <name>setiathome_x41zi_x86_64-apple-darwin_cuda75</name> <executable/> </file_info> <file_info> <name>libcudart.6.5.dylib</name> <executable/> </file_info> <file_info> <name>libcufft.6.5.dylib</name> <executable/> </file_info> <app_version> <app_name>setiathome_v8</app_name> <platform>x86_64-apple-darwin</platform> <version_num>802</version_num> <plan_class>cuda65</plan_class> <avg_ncpus>0.1</avg_ncpus> <max_ncpus>0.1</max_ncpus> <cmdline>-poll</cmdline> <coproc> <type>CUDA</type> <count>1</count> </coproc> <file_ref> <file_name>setiathome_x41zi_x86_64-apple-darwin_cuda75</file_name> <main_program/> </file_ref> <file_ref> <file_name>libcudart.6.5.dylib</file_name> </file_ref> <file_ref> <file_name>libcufft.6.5.dylib</file_name> </file_ref> </app_version> </app_info> Since you are only using the GPUs using a full CPU per task shouldn't be a concern, and it will complete the GPU tasks faster. It would be even faster using the Special App.... |
spitfire_mk_2 Send message Joined: 14 Apr 00 Posts: 563 Credit: 27,306,885 RAC: 0 |
At last check, (been a few months since I've looked), the CUDA Driver from NVIDIA for MAC OS X was 7.5.26; this is what I'm on. I still encounter "Validation Inconclusives" and "Invalids" on CUDA65; but NOT as many as when I was stock on OpenCL. The one "Error" I still have showing is from a "Ghost" that was created when I switched from OpenCL to CUDA65 and my entire cache got wiped. My limited experience is that mobo with DDR2 ram only support 8 gb. Did you check documentation? |
TimeLord04 Send message Joined: 9 Mar 06 Posts: 21140 Credit: 33,933,039 RAC: 23 |
...(Edited for space and readability.) Yes, the docs state that 16 GB is the limit for this board. :-) TimeLord04 Have TARDIS, will travel... Come along K-9! Join Calm Chaos |
TimeLord04 Send message Joined: 9 Mar 06 Posts: 21140 Credit: 33,933,039 RAC: 23 |
16 GB Patriot DDR2 PC2 6400 RAM installed and working perfectly. :-) Lagging has been greatly reduced; but, NOT fully eliminated. It still takes some time to maneuver within my Hushmail e-Mail Account, moving things, deleting things, making folders, etc... On 8 GB DDR2 RAM, these things would end up taking 5 Min sometimes; resulting in a gray spinning wheel in the corner of the open Tab in Firefox. Now, under 16 GB RAM, it took 30 seconds to get into the e-Mail Account and one item therein took 1.5 Min... MUCH IMPROVED. :-) Under the 8 GB of RAM; almost 4 GB was being taken up by system resources and open/active programs. Now, this has dropped 1 GB to 3 GB of resources used and the rest of the 16 GB is free. :-) I will find out more tonight when BOINC resumes crunching. I still have a couple of Einstein Units left in queue. So, when things pick up tonight; I'll have to wait until those finish before BOINC dedicates both 750's to SETI. TL TimeLord04 Have TARDIS, will travel... Come along K-9! Join Calm Chaos |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.