Message boards :
Number crunching :
all 14au18ac GPU tasks on one system running 3 seconds?
Message board moderation
Previous · 1 · 2 · 3
Author | Message |
---|---|
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
Nice to hear. Now let's keep an eye on inconclusives and nvalids. even you have hybrid cards, your VRMs and mem chips are still air cooled. be mindful of that, especially with your 4x 1080ti system where you are significantly impeding airflow on 3 cards for the fan responsible for getting airflow to the mem and VRMs Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
Yes but the 1080Ti have a modified copperplate that extends over those. Not so with the 1070s or the 9X0s series. I learned that when I was modifying those over the past few years. When I modified the 1070 FTWs to hybirds I ended up modifying the kit to included those. When the 1080Ti hybrids come out, it was almost as if they had been paying attention to what we had done. So I don't worry much about the 1080Ti hybrids as I used to. |
petri33 Send message Joined: 6 Jun 02 Posts: 1668 Credit: 623,086,772 RAC: 156 |
Nice to hear. Now let's keep an eye on inconclusives and nvalids. Lucky you with hybrids. My temps are at 82C. To overcome Heisenbergs: "You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13161 Credit: 1,160,866,277 RAC: 1,873 |
We got that and understood eventually Petri. Mostly just doing a little light editing of my app_info files to change over to the Pascal or the multiverse app and remove the old cuda library references. Just using the app_names you gave them in the zip files. All looking good so far and VERY impressed by the task speedups. I have always run my fans at 100%. The Inconclusives look good and invalids look good so far. The one system with the very large sudden Inconclusive count was from the headless system that I didn't notice overflowing a ton of tasks on the zi3v application early yesterday. I chalk that up to too much memory overclock for air cooled 1070Ti cards that are butt-to-butt in the system. I backed off the overclock to what was stable before on the zi3v app before I got greedy. Looking good so far on the new 0.97 app. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
Nice to hear. Now let's keep an eye on inconclusives and nvalids. . . Those are temps to be envied ... Stephen :) |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
All the 14au tasks I got seemed to take "forever" not run extra fast :( Tom A proud member of the OFA (Old Farts Association). |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
All the 14au tasks I got seemed to take "forever" not run extra fast :( . . If you take a look at the results for those tasks in the SETI database and look at the stderr info you will probably find they have lower than the "normal" AR of 0.44. The lower the AR, the longer they take. . . I have had a look and there are no tasks for 14au for the machine with the single 750ti, but there are a lot of errors. I am guessing that was from when you had the SoG tuning parameters in your app_info. There are task results from 14au for the machine with the two 750ti cards but the run times seem pretty normal to me. Stephen . . |
arkayn Send message Joined: 14 May 99 Posts: 4438 Credit: 55,006,323 RAC: 0 |
All the 14au tasks I got seemed to take "forever" not run extra fast :( Also if you are running the SOG app, it will take longer on Arecibo tasks than than the Guppi's, the CUDA app is reversed on those accounts. |
MajorTom Send message Joined: 25 Aug 03 Posts: 33 Credit: 78,247,091 RAC: 46 |
working app_info.xml just the cuda92 part 4 the multi gpu version maxwell chipset setiV0.97.multi_sm.linux_X86_64_cuda92 needs embed in <app_info></app_info> https://pastebin.com/P7qXkmiy |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
I am getting lots of very short tasks. And they are validating. On both the gpu and cpu sides. Anyone else? Tom A proud member of the OFA (Old Farts Association). |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
I am getting lots of very short tasks. And they are validating. On both the gpu and cpu sides. . . They are very short because it is a very noisy signal segment (for whatever reason) and it is abandoned after 30 signal hitss which occcur very quickly because of all the noise. But because the result is correct they will validate (mostly) and earn you ... ooohhhh ... 0.5 points :) Stephen :) |
©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.