Posts by Sebastian*

41) Message boards : News : Running Modfit on MilkyWay@home (Message 64783) Posted 30 Jun 2016 by Sebastian* Post: Thanks for the reply Jake :) Good luck with the Linux binary. I hope you find a useful documentary about the server config :) Take your time, i can still remember the time when Milkyway@home was one of the very unstable and unreliable projects. We have come a long way, keep up the good work :)
42) Message boards : News : Running Modfit on MilkyWay@home (Message 64777) Posted 29 Jun 2016 by Sebastian* Post: Jake Weiss I still only get 25 WUs per computer, and i run out of them quickly. Even now and then the time to the next request goes up, and i have to wait several minutes until the next request is done. Could you set the server so hold 3000 to 5000 WUs ready for sending? I have set my computing preferences: Maintain enough tasks to keep busy for at least "1 day" ... and up to an additional "1 day" I think holding "only" 1000 workunits is not enough for such short WUs Milkyway@Home provides.
43) Message boards : News : Running Modfit on MilkyWay@home (Message 64701) Posted 21 Jun 2016 by Sebastian* Post: Just to clarify, what is the name of the new task, in the Boinc Projects directory? I am getting computation errors since yesterday, but i think it has to do with the PrimeGrid apps i am running since then.
44) Message boards : Number crunching : What am I missing in order to get more GPU tasks? (Message 64659) Posted 15 Jun 2016 by Sebastian* Post: Hi, just got this message: 15/06/2016 12:13:18 \| Milkyway@Home \| Not sending work - last request too recent: 3296 sec i would like to go back to the old 1 minute checkup time. Can the project increase WU-generation? So there will be work for everyone?
45) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 63832) Posted 26 Jul 2015 by Sebastian* Post: Thanks for the heads up Cliff. Aborting is the best way, glad you could do that. Good luck with the rest of the components, i hope most survived. [TA]Assimilator1 i am using the 13.1 driver on the 6950s. I haven't tested other projects yet with it. But what is intresting, if you run a 5970 (2 5870s in one card) then you have to go back to 12.4. Because the other drivers have trouble with the integrated Crossfire when running milkyway. But for games works the 15.7 fine.
46) Message boards : Number crunching : All Milkyway@Home 1.02 tasks ending in computation error on HD6950. (Message 63819) Posted 21 Jul 2015 by Sebastian* Post: Is there an Update on what driver to use for the 6950 GPUs? If you uninstall the AMD drivers, use the AMD cleanup utility, which completely removes the driver and openCL stuff. If you have run 14.x for example, and just uninstall it, the openCL stuff will stay on your system and the lower driver version (13.1 for example) still uses it, so causing your computation errors.
47) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 63817) Posted 21 Jul 2015 by Sebastian* Post: This Processs Lasso tool seems interesting, but i think it is way overkill for me. Since most my computers only have windows and boinc running. The tool seems to avoid hyper treading, which i also do in only simulating 4 CPU cores in boinc, instead of the 8 cores Windows shows me. I haven't read something about the AMD CPU architecture on the tool website. The AMD CPUs are a little complicated to optimize on because of their design. I don't think the tool would help me there, since only a program which is programmed for the Module design would fully utilize it. By the way, does anyone know, what AMD driver to use for the 6950 / 6970 GPUs, so that all Milkyway WUs run fine. With the current driver the long tasks just create Computation Errors after a few seconds. And i didn't have time yet to search the forum.
48) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 63798) Posted 14 Jul 2015 by Sebastian* Post: Theoretical yes. But windows throws around the Threads at random. I don't know any option or program that can reserve one Module. The 323 second times apply when i only run Milkyway@home, no CPU tasks from Boinc. But i should keep the times, when i could reserve one Module exclusively.
49) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 63791) Posted 8 Jul 2015 by Sebastian* Post: An AMD FX 8150 Desktop CPU has 4 Modules. A Module are 2 Integer processing units and one Floating point processing unit. But windows sees a module as 2 cores. As long as integer stuff is running it works fine this way. But when you run Floating Point things, you have to share the one unit for 2 cores. When i run Asteroids@home on 7 cores then one Floating Point unit is shared with the one Milkyway@home needs from time to time. So GPU-Times go up. The windows sceduler is also a mess. It works ok for real cores. But when you have Hyper Threading on Intel CPUs enabled, it does not make a difference between real and virtual cores. Windows also trows tasks around on the cores at will, which causes delays (empty cpu pipeline on a core and reload the stuff on another one). This is quite frustrating in my opinion, since a lot of time for work on the cpu is wasted.
50) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 63789) Posted 7 Jul 2015 by Sebastian* Post: There are currently 291 valid WUs from the 660TI GPU. But i let 7 CPU cores run Asteroids at home. So there is some upwards variation from the 383 seconds. AMD FX 8150 CPU "8 Cores / 4 Modules" I have no means to let the GPU run on one Module exclusively. This is why the times vary.
51) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 63784) Posted 2 Jul 2015 by Sebastian* Post: I let the 660TI run some Milkyway, almost all times are at 383.2x seconds. You can check with the link from your last post. 97 valid WUs just now.
52) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 63781) Posted 29 Jun 2015 by Sebastian* Post: the 114.2 and 383.3 had only those times, as good as no variations. The lastest times from the 660ti are with Vlc running, and trying to run 2 Work Units at once. Still impressive to see so little difference between 1 WU 383s and 2 WUs 410s.
53) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 63778) Posted 29 Jun 2015 by Sebastian* Post: I looked at more then 5 results, and guessed the times. But looked more at the longer times. So the times might be a few seconds longer than average. Here are the times again with clocks: GTX 580: 114,2s CPU 21.3s (815 GPU / 1025 Mem) (Stock would be 772/1002) and 90%+ GPU Load GTX 460: 292s CPU 52s (763 GPU / 950 Mem) (Stock would be 675/800) and 95%+ GPU Load GTX 680: 343s CPU 1,8s (At Stock speeds 1006 GPU / 1500 Mem, but GPU Boost around 1070 to 1080) and 55% GPU Load GTX 660ti 383,3s CPU 2,4s (At Stock speeds 915 GPU / 1500Mem, but GPU Boost runs at 1058) and 50%+ GPU Load
54) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 63773) Posted 26 Jun 2015 by Sebastian* Post: Here we have some times: GTX 580: 114,2s CPU 21.3s GTX 460: 292s CPU 52s GTX 680: 343s CPU 1,8s GTX 660ti 383,3s CPU 2,4s All times are from running one WU on one GPU The GTX 460 outcrunches the 680. I find that a little surprising. But when i think about other projects, the 4xx and 5xx series were stronger crunchers than the 6xx series, at least until the apps were optimizes. In case of milkyway i thing that the 4xx serie also has more double precision power. The CPU-Times are also intresting, and i link them to the architecture of the GPU. The 680 runs at stock clocks. the other cards have slight factory overclocks. I have to check them tomorrow.
55) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 63767) Posted 24 Jun 2015 by Sebastian* Post: I haven't seen a GTX 680 or 660ti on your list. So we would have times for those GPUs. I can imagine, that even a GTX 580 would outcrunch a 680 here in milkyway. We can compare the GTX 680 to the Titan (both Kepler) so we might get an idea why there are so high CPU-Times on the Titan I will post the times later here. I can also confirm the HD 4850 and HD 5850 times. The times from both cards are still valid, when i look at the times from my cards.
56) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 63764) Posted 24 Jun 2015 by Sebastian* Post: To khryl and [TA]Assimilator1: I could run some WUs on a GTX 680 or 660Ti, or even on a 580. So we have some GPU and CPU times to compare. One WU on the Titan would also be possible. One problem about the downtime in Milkyway is, that the GPUs got more powerful over time, but the WUs are still need the same amount of crunching. I wonder how a Fire Pro W9100 would do. It has 2,5 Terraflop double precision crunching power. 2.5 times more than a 7970. By the way, can you run Milkyway on Hawaii GPUs?
57) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 63763) Posted 24 Jun 2015 by Sebastian* Post: Hi khryl, the Nvidia GTX Titan Black is a special case in my opinion. I think the Nvidia WUs are not optimized for a Nvidia card with that much double precision power. (Titan Black has 1,5 Terraflop, but normal cards are more aroun 0.1 Tflop to max 0.2 Tflop) When i just run one WU on the Titan Black the card runs at 12% load (double precision enabled in Nvidia control center) or at 25% load (when double precision is not enabled) I have a i7 4790k CPU in this computer, and when double precision is disabled the CPU-load is near to 100% when i run 8 WUs at once, and the times are insanely high then. khryl you can run 4 WUs at once on the 7970 / 280X so you have no downtime. And i think not that this is cheating, since the Milkyway@home project gets more WUs returned. But you should also set the CPU to .25, so you have one core reserved for all 4 Milkyway WUs. But keep in mind that your GPU will run at 100% all the time. When i take a look at wikipedia and compare the 7970 (at 1GHz) to the Titan Black at stock, then the AMD card is way better than the Nvidia card. 7970 has 1 Tflop and the Titan Black has 1.5 Tflop. But they both have comparable times. [TA]Assimilator1 you have to edit the list, i have a Titan Black, not a normal Titan. The Titan Black has more shader units and a slightly higher GPU Clock. I am not sure what to think about the high CPU times. I have no idea how they are calculated. And the enabled Hyper Treading might mess up the times as well. But i only run Boinc on 4 cores (3 for CPU stuff and 1 for Milkyway exclusive)
58) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 63743) Posted 19 Jun 2015 by Sebastian* Post: DR4GOON, i can feel with you, a few month ago my 7870XT stopped working. I have updated times for the GTX Titan Black GPU at stock speed of 889MHz. The MilkyWay@Home v1.02 (opencl_nvidia) tasks need an average time of 208.82 seconds when i run 8 WUs at once. And the Milkyway@Home Separation (Modified Fit) v1.36 (opencl_nvidia_101) need 60.624 seconds when 8 WUs are ran at once. I need to run 8 WUs in parallel, so i can get 99% GPU usage. Average time per WU, if it would one be one: MilkyWay@Home v1.02 (opencl_nvidia) 25.23s (201.82s/8) Milkyway@Home Separation (Modified Fit) v1.36 (opencl_nvidia_101) 7.58s (60.624s/8) But you have to enable the Double Precison feature in the Nvidia driver. Otherwise the time are much much longer, and the CPU is heavily used. The times are from this Computer: (Not overclocked) http://milkyway.cs.rpi.edu/milkyway/show_host_detail.php?hostid=593293 PS: [TA]Assimilator1 look at the links in your GPU-Times list. For some reasons the are two "http://" in the beginning of the links. List of GPU-Times Example: http://"http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=3551&postid=63236"
59) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 63700) Posted 11 Jun 2015 by Sebastian* Post: In the linked list, 7th and 8th row for above: Microsoft Windows running on an AMD x86_64 or Intel EM64T CPU 1.02 (opencl_amd_ati) 10 Feb 2012, 7:11:09 UTC 56,748 GigaFLOPS Microsoft Windows running on an AMD x86_64 or Intel EM64T CPU 1.02 (opencl_nvidia) 10 Feb 2012, 7:11:06 UTC 17,646 GigaFLOPS In the last section of the list "Milkyway@Home Separation (Modified Fit)" it gets even worse for Nvidia cards: Microsoft Windows running on an AMD x86_64 or Intel EM64T CPU 1.36 (opencl_ati_101) 6 Oct 2014, 20:18:32 UTC 15,667 GigaFLOPS Microsoft Windows running on an AMD x86_64 or Intel EM64T CPU 1.36 (opencl_nvidia_101) 6 Oct 2014, 20:18:35 UTC 4,951 GigaFLOPS I think the shown gigaflops mean the average computing power needed to finish a task. Ok, looks like i got fooled by thinking about this list too much. I really seems that it is the current crunching power donated to Milkyway@home. But for some reason the Mac crunching at the end of the list seems way off. Update for my GTX Titan Black: I just found out how useful the app_config.xml is :) Runnin 10 tasks on the Titan, and i have enabled Double Precision in the Nvidia Driver. I hope i can compare the times with the 7970s in a day or so, when i have enough finished tasks to average the times. Theoretically it should be 30% faster than i 7970. But i've read that it is difficult to access the full power with opencl. I hope for the same crunching power then.
60) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 63697) Posted 10 Jun 2015 by Sebastian* Post: Thanks for this nice list Assimilator1. But did you take in account, that the AMD GPU apps need 3 times the crunching power than the Nvidia gpu apps? Check this list_ http://milkyway.cs.rpi.edu/milkyway/apps.php I am struggeling to get my Titan Black fully utilized. And stumbled over this list while finding a way to use the 1,3 Terraflop crunching power.

Previous 20 · Next 20