Posts by Eric Driver

1) Message boards : News : Future of the Project (Message 3471) Posted 15 days ago by Eric Driver Post: At Gerasim, SerVal has integrated the apps for nvidia gpu (on windows), amd gpu (on windows), and linux cpu. These apps are now processing WUs for subfield 3. On Numberfields, I am sprinkling in some WUs for sf6, in order to get some updated timing information. This will help in creating the new batches for sf6 to come later in the year.
2) Message boards : Number crunching : Creating a config file for NumberFields@home / setting aside enough resources for the GPU? (Message 3470) Posted 23 Feb 2023 by Eric Driver Post: You can control the number of cpus through the BOINC manager. To run more than 1 GPU task, see this link for my app_config, but I think the one you quoted should also work if you change the cpu_usage and gpu_usage fields: https://numberfields.asu.edu/NumberFields/forum_thread.php?id=558&postid=3380#3380
3) Message boards : Number crunching : Credits aren't updated on BOINCstats, Free-DC (Message 3465) Posted 20 Feb 2023 by Eric Driver Post: Why no 500k Badge? There is one and it looks like you got it. The badge assignment script only runs once per day, so maybe that was the problem.
4) Message boards : News : Future of the Project (Message 3462) Posted 19 Feb 2023 by Eric Driver Post: It sounds like you need to talk to the User Petri33 of the Gpu Users Group Team and see if he can help you, he's optimized several Projects Nvidia apps so far, Linux only though. The problem is integrating the existing gpu app into the Gerasim server. Optimizing the CUDA or OpenCL code is a different problem altogether, not to mention a very difficult problem - it would take a guru of the highest order to improve the current code.
5) Message boards : News : 2022 Year End Summary (Message 3458) Posted 19 Feb 2023 by Eric Driver Post: Is there any way to guesstimate percentage of total work done ... or have all workunits not been generated? Would be cool to see updates every month or two, so folks have an idea of how close we are to completion! Very exciting, all the best!! I can only make estimates on those batches that have been generated. So this means the current batch sf7_DS16x11 and the last of sf3. You should be able to get estimates from the batch status page for those cases. I have generated the first 3 million from set sf7_DS16x12 and from my tests, I estimate that set will take approximately 1 year at the current rate. I have not generated any of the remaining sf6 cases, so those are unknown, but I can say that sf6 will be the hardest subfield (and hence why we do him last).
6) Message boards : News : Future of the Project (Message 3453) Posted 18 Feb 2023 by Eric Driver Post: Hello if you want to continue using the Gerasim server, it would be better to keep applications for Linux and Mac. thank you With the help of SerVal I dropped a batch of tasks on the Gerasim server. Currently just for windows cpu. The tasks are for the remainder of the subfield 3 search. Those users that are more interested in sf3 can get tasks from Gerasim - the only caveat is that you get Gerasim credit instead. I think of this as a trial period, to see how it goes. Numberfields is not going away anytime soon, and Gerasim can run the sf3 tasks in parallel to the sf7 tasks currently running on NF. I will keep this up for as long as Gerasim and/or SerVal allows.
7) Message boards : Number crunching : Project Admin(s) (Message 3451) Posted 15 Feb 2023 by Eric Driver Post: Project Admin(s) BOINCGames.com has a Sprint coming up in approx. 14 hours: Skiing Alpine World Ski Championships 02/15/2023 14:00 (UTC) - 02/19/2023 14:00 (UTC) Your project is elgible for being randomly selected (by a computer). This Sprint may bring a slightly higher than usual load on your servers for the duration, but BOINCGames.com is still under BETA testing and the particpiation level is still relatively low. This is just a heads-up that you might want to ensure there is enough work for the participants if your project is selected. Thanks for letting me know! We have plenty of work available.
8) Message boards : News : Future of the Project (Message 3449) Posted 14 Feb 2023 by Eric Driver Post: I think it's best not to wait when the current server dies. That's another option if/when the current server dies. Yes. Why not use the public server gerasim.boinc.ru ? (you can upload tasks to the server and receive the results calculated by the participants). @Eric: If you are interested, I will tell you how to do it. Best regards, SerVal (gerasim.boinc.ru admin). @SerVal: I thought gerasim was a stand alone project. I am interested to learn more. I will respond to your pm so we can discuss offline.
9) Message boards : News : Future of the Project (Message 3444) Posted 6 Jan 2023 by Eric Driver Post: Have you considered to rent a dedicated server. For Example Hetzner Auction offeres good price/performace offerings If taking cheapest server with XEON processor avalable, 32 GB RAM and 10 TB HDD (RAID), the price is 62 dollars + VAT per month. https://www.hetzner.com/sb?drives_size_from=10000&drives_size_to=16000 For this price you get unlimited traffic, with 1 Gbps uplink. Several years ago I was renting 1-2 servers from them to run BOINC projects, mainly primegrid and it was great service quality. One time PSU failed and it got replaced in 1 day with no additional cost. Sure this idea still needs a monthly donation, but if monthly price would be 1 dollar or more (if anyone wants), I think there will be plenty of people who can donate that sum each month (me counted in :) ). That's another option if/when the current server dies.
10) Message boards : News : 2022 Year End Summary (Message 3442) Posted 1 Jan 2023 by Eric Driver Post: 2022 was the year of ℚ(√-10) (subfield 7), as most of our time was spent on that search. Many new fields were found and we are now entering the final row of the search. In the next 3 to 6 months we should be at the point where we can say with almost certainty that it is complete. The goals for 2023 are simple: 1. Get subfield 7 to the point where it is essentially complete. 2. Go back and officially complete subfield 3, so we can say with 100% certainty that it is done. 3. Do more on subfield 7 or move on to subfield 6 (the final search). Thanks everyone for your contributions and have a wonderful New Year!
11) Message boards : Cafe : Happy New Year and we are closing in on the current batch DS15x12 (Message 3441) Posted 1 Jan 2023 by Eric Driver Post: Happy New Year everyone. I occasionally check on our batch progress and found that we may finish this set on new year's day a while back. Day after day the hope of it dwindled and dwindled, but the est. kept dropping rapidly. Yesterday I already knew it would not finish today as the number started stabilizing. I know the est. is inaccurate during the final crunch but it would have been a banger to conclude 2022 with it. Did other people notice it? Was my estimate very wrong, I do think I was very optimistic tbh. Happy New Year! Yes, I noticed it too. From past experience, I know it is hard to estimate the final time with all the ebbs and flows of the volunteers. Looks like a little more than a day left. But it will still take about a week for the remaining WUs to trickle in before I can officially collate the results.
12) Message boards : Number crunching : How many Gfx memory needed by Intel GPU OpenCL on Ubuntu 22.04 (Message 3437) Posted 20 Dec 2022 by Eric Driver Post: Updated: I downgraded (rollbacked) to an earlier release of Intel's NEO compute runtime, and then the OpenCL seems to work normally Check it out at https://github.com/intel/compute-runtime/releases Good to hear you got it working!
13) Message boards : News : Future of the Project (Message 3431) Posted 28 Nov 2022 by Eric Driver Post: It is interesting what is the bandwidth should internet have in order this project works properly? Is it enough 100 MB/s (I have such internet at my home). Or it needs 1 GB/s or greater? It's hard to say, but the more the better. When there is a period of heavy usage, such as during a competition, is when you will have problems. Also, remember it's important for projects to have a fast upload speed. My ISP gives a pretty fast download speed, but the upload is only ~10MB/s. To get more, they want me to purchase a "business" account and pay through the nose.
14) Message boards : News : Future of the Project (Message 3429) Posted 21 Nov 2022 by Eric Driver Post: It interesing, which search are you going to calculate after DS15x12 The plan is to do DS16x10 and DS16x11, and then back to finish off sf3.
15) Message boards : News : Support for Intel GPUs (Message 3423) Posted 14 Nov 2022 by Eric Driver Post: I seem to be having problems in getting GPU work. I have an Intel HD 630 GPU as far as I am aware the driver is up to date, but I keep getting the message no GPU work available for some reason. I can run Einstein WU's no problem so BOINC knows the GPU is there. Thanks It could be several things. You need to have beta apps selected and the card must support openCL 1.1
16) Message boards : News : Support for Intel GPUs (Message 3420) Posted 14 Nov 2022 by Eric Driver Post: I was comparing to a default Windows Get Decic Fields v4.00 (default)windows_x86_64 You may have a point about Driver or implementation Bill F The GPU apps should be faster than the default CPU app. However, some of the older cards can be slower, if they work at all.
17) Message boards : News : Support for Intel GPUs (Message 3418) Posted 13 Nov 2022 by Eric Driver Post: bunch of WU processed :) https://numberfields.asu.edu/NumberFields/result.php?resultid=161210884 https://numberfields.asu.edu/NumberFields/result.php?resultid=161211661 and plenty others Thx! Good to see it's working, and that run times are reasonable on Intel's newest GPUs. Run times are longer for the Intel GPU tasks. Are they doing larger amounts of science and should the credit amount be higher ? Bill F Run times are longer compared to what? Compared to the previous dataset, run times are longer and credits have gone up accordingly. If comparing to a different gpu then average times will probably be different. Each WU is a fixed amount of work and therefore has a fixed credit. Just because they run longer doesn't mean they deserve more credit. If it's running longer than expected then it could be an inefficient driver or bad OpenCL implementation.
18) Message boards : News : Support for Intel GPUs (Message 3416) Posted 11 Nov 2022 by Eric Driver Post: bunch of WU processed :) https://numberfields.asu.edu/NumberFields/result.php?resultid=161210884 https://numberfields.asu.edu/NumberFields/result.php?resultid=161211661 and plenty others Thx! Good to see it's working, and that run times are reasonable on Intel's newest GPUs.
19) Message boards : News : Support for Intel GPUs (Message 3410) Posted 6 Nov 2022 by Eric Driver Post: Adding an Intel ARC A750 to the pool, watch these spaces => https://numberfields.asu.edu/NumberFields/show_host_detail.php?hostid=2835341 and https://numberfields.asu.edu/NumberFields/results.php?hostid=2835341 It's been several hours and nothing returned yet, I hope that's not a bad sign. I would expect the Arc A750 to do well, assuming Intel puts out a decent openCL driver. The machine is busy with Einstein@home tasks :D Arc A750 doesn't have FP64 though I'll pause Einstein and let some tasks run I saw some cpu tasks go through. Any problems with tasks on the Arc A750?
20) Message boards : News : Support for Intel GPUs (Message 3408) Posted 5 Nov 2022 by Eric Driver Post: Adding an Intel ARC A750 to the pool, watch these spaces => https://numberfields.asu.edu/NumberFields/show_host_detail.php?hostid=2835341 and https://numberfields.asu.edu/NumberFields/results.php?hostid=2835341 It's been several hours and nothing returned yet, I hope that's not a bad sign. I would expect the Arc A750 to do well, assuming Intel puts out a decent openCL driver.

Next 20