Posts by Sergei Chernykh

201) Message boards : Number crunching : Newbie questions about the project (Message 816) Posted 30 Apr 2018 by Sergei Chernykh Post: All the amicable pairs under 10²⁰. I meant that it's only needed to check ~2.5% of all numbers below 10²⁰ to do that.
202) Message boards : Number crunching : Newbie questions about the project (Message 814) Posted 30 Apr 2018 by Sergei Chernykh Post: 1) Not through all numbers, only through numbers that can't be proven to not be amicable. ~2.5% of all numbers need to be checked. 2) Distribution of amicable numbers is not uniform and heavily depends on their factorization. 3) Factorization is provided for convenience. It allows to check the numbers quickly. 4) I don't track it, although I save all completed work units and I know how much CPU time each work unit takes on average.
203) Message boards : Random stuff : Project Status Page (Message 809) Posted 29 Apr 2018 by Sergei Chernykh Post: Fixed it, thanks for noticing.
204) Message boards : Number crunching : Unable to get more than 20WU (Message 801) Posted 27 Apr 2018 by Sergei Chernykh Post: No, the number of CPU cores that count for this restriction is what you can see in your Computers belonging to page. P.S. Actually, BOINC client settings can change this number. But "Max # CPUs" setting shouldn't influence it.
205) Message boards : Bug tracker : Not getting GPU workunits (Message 799) Posted 27 Apr 2018 by Sergei Chernykh Post: Check your global BOINC client preferences. You have "Run GPU tasks if user is active" unset.
206) Message boards : Bug tracker : The project send tasks to a video card with only 1G memory (Message 795) Posted 20 Apr 2018 by Sergei Chernykh Post: Decrease kernel size for NVIDIA GPUs in project preferences. Current WUs should work now even on 1 GB cards.
207) Message boards : Number crunching : GPU version requirements for the search up to 10^20 (Message 793) Posted 19 Apr 2018 by Sergei Chernykh Post: Now that WU's allocate less than 900MB, is it possible to relax the GPU memory requirement to 1GB ? I'll try to do it today and then watch if I get any new WU errors on the project.
208) Message boards : Bug tracker : Not getting CPU units (Message 789) Posted 13 Apr 2018 by Sergei Chernykh Post: I didn't change anything on the project side recently, so it's something with your settings. Check here: https://sech.me/boinc/Amicable/prefs.php?subset=project
209) Message boards : Number crunching : Unable to get more than 20WU (Message 787) Posted 3 Apr 2018 by Sergei Chernykh Post: The project has never been down for more than a few minutes so far, it's running on a dedicated server in a data center. Is it a real issue for you? Do you often get disconnected for many hours?
210) Message boards : Number crunching : Unable to get more than 20WU (Message 785) Posted 3 Apr 2018 by Sergei Chernykh Post: The limitation is 20 WU per single GPU and 2 WU per single CPU core.
211) Message boards : News : GPU version fix (Message 782) Posted 27 Mar 2018 by Sergei Chernykh Post: Good news: I was able to narrow down the invalid work units range to only last 9 days after further analysis of this bug.
212) Message boards : News : GPU version fix (Message 781) Posted 27 Mar 2018 by Sergei Chernykh Post: Is this bug the reason for the semi-random work unit runtimes lately? Yes, it was running random percentage of a work unit instead of full 100%.
213) Message boards : News : GPU version fix (Message 779) Posted 27 Mar 2018 by Sergei Chernykh Post: There was a serious bug in GPU version: https://github.com/SChernykh/Amicable/commit/3470ed855d244100f37b0b7ab4c912d2b9365a93 Unfortunately, it skipped some numbers in an unpredictable way for "large primes" work units for the last ~~20 days~~ 9 days. I'll have to reissue all such work units to cover everything missed. The GPU versions for all platforms are already updated. Sorry for the mess, but it's good that the bug is fixed relatively early, because it was hard to notice.
214) Message boards : Number crunching : Auto-ban faulty hosts (Message 778) Posted 27 Mar 2018 by Sergei Chernykh Post: to start banning pc's with just 2 errors out of 242 workunits seems a bit harsh. What's the procedure for the guy to bring his pc back? In the banning process he had more workunits on his pc that he finished crunching but was NOT allowed to return them to even figure out it the problem was fixed or not, he had to abort them all and move to another project, where the graphics card is working just fine. It's not harsh, because extremely few computers actually get banned here. He needs to check his GPU, run stress tests and make sure it doesn't crash or give errors, doesn't overheat, and I'll unban it then. "Working just fine" depends on the type of workload. It may be just on the edge for some projects and run fine, but cross the edge and fail sometimes on other projects.
215) Message boards : Number crunching : Auto-ban faulty hosts (Message 776) Posted 27 Mar 2018 by Sergei Chernykh Post: 1 error detected in the compilation of "C:\Users\steve\AppData\Local\Temp\OCL2252T1.cl". Frontend phase failed compilation. OpenCL.cpp, line 397: Trying to disable 'goto' and build again 01:16:06 (2252): called boinc_finish(0) please tell me how this is a USER CAUSED ERROR? Because I too could be banned if this is a problem caused by the users!! That PC has this message in all tasks' logs, including a few hundred valid ones. It's normal: in this case it failed to compile because of a known issue with amd drivers and a workaround was applied, so it compiled and ran fine after that. But incorrect result was returned for a few tasks out of a few hundred valid tasks. This is a direct indication of unstable hardware. Many similar configurations run with the same message in logs, but without problems (validation failures) on this project. Only validation failures are counted - it's when a client didn't report any errors on exit, but returned incorrect result. This can jeopardize the whole project's target and can't be tolerated. Only "Completed, marked as invalid" status is counted for banning. Because I too could be banned if this is a problem caused by the users!! No, your computers have 0 "Completed, marked as invalid" tasks. P.S. There are over 5000 hosts that run Amicable Numbers every day. Only 3 hosts are currently banned, so please don't exaggerate this problem.
216) Message boards : Number crunching : Auto-ban faulty hosts (Message 774) Posted 25 Mar 2018 by Sergei Chernykh Post: I did 1300 good work units and 4 invalid and you banned my computer!!!!!!! BYE!!! Do you realise this ban was automatic? Unless you fix this PC, it will stay banned.
217) Message boards : Random stuff : Social Numbers (Message 771) Posted 20 Mar 2018 by Sergei Chernykh Post: I would assume that the Amicable Numbers project has a database with the sum of the proper divisors for many numbers. No, it's not even theoretically possible to save this amount of data. The project only stores found amicable pairs. Social numbers require different search algorithm.
218) Message boards : Number crunching : GPU version: kernel size tuning and less UI lags (Message 769) Posted 18 Mar 2018 by Sergei Chernykh Post: There was not enough GPU memory. Try to close all other programs. Work units that are being issued now require less memory than before, so there is a good chance you won't have this error now. P.S. You can also try to reduce kernel size.
219) Message boards : Number crunching : >96% of v2.08 aarch64-unknown-linux-gnu (mt) WUs failing (Message 766) Posted 13 Mar 2018 by Sergei Chernykh Post: You'll need a 2 GB swap file for current work units.
220) Message boards : Getting started : how to run on C2070 & K80 (Message 763) Posted 12 Mar 2018 by Sergei Chernykh Post: BOINC didn't recognize these cards on your computers. You can try to reinstall graphics drivers.
221) Message boards : Bug tracker : Computation error at 6 seconds (Message 759) Posted 9 Mar 2018 by Sergei Chernykh Post: MacOS GPU version broke again when the project switched to the new type of work units, sorry for the mess. I'll fix it later today when I have time.
222) Message boards : Random stuff : Command Line Change (Message 757) Posted 7 Mar 2018 by Sergei Chernykh Post: It's a new type of work units (I call them "large prime WUs"). The search now goes through all primes between first and second number in the command line and looks for amicable numbers of the form "np" where p is a prime from this range and np < 10²⁰. https://github.com/SChernykh/Amicable/blob/boinc-version-128-bit/Amicable/Amicable.cpp#L131
223) Message boards : Random stuff : Up to date list of project badges (Message 755) Posted 6 Mar 2018 by Sergei Chernykh Post: Added new badge for finding 100,000 amicable pairs:
224) Message boards : Random stuff : Which OS this project support? Also, Can you help me Mapping BOINC community? (Message 754) Posted 3 Mar 2018 by Sergei Chernykh Post: https://boinc.berkeley.edu/projects.php Here is written too Nvidia/AMD gpu is incorrect, right?. Sorry for asking dumb questions. It's correct. The project supports AMD and NVIDIA GPUs.
225) Message boards : Bug tracker : Computation error at 6 seconds (Message 751) Posted 3 Mar 2018 by Sergei Chernykh Post: Lamplighter55 Can you try again with version 2.09?
226) Message boards : Bug tracker : Computation error at 6 seconds (Message 750) Posted 3 Mar 2018 by Sergei Chernykh Post: MacOS didn't like new OpenCL code: <program source>:1516:29: error: comparison of integers of different signs: 'ulong' (aka 'unsigned long') and 'int' if (amicableCandidateIndex >= candidatesDataHighBitOffsets.x) value_ulong \|= 0x100000000UL; I'll fix it ASAP.
227) Message boards : Random stuff : Which OS this project support? Also, Can you help me Mapping BOINC community? (Message 748) Posted 3 Mar 2018 by Sergei Chernykh Post: The project supports Widows, Linux and MacOS: https://sech.me/boinc/Amicable/apps.php Requirements: 64-bit OS, 2 GB of free memory.
228) Message boards : Random stuff : Report on search up to 2^64 (Message 746) Posted 27 Feb 2018 by Sergei Chernykh Post: https://sech.me/boinc/Amicable/forum_thread.php?id=91 Yes, I also have an analysis of amicable pairs distribution based on the results of this search, which gave me an approximate formula for the amount of amicable pairs below any given number. I'll publish an article describing it soon, when I have more free time. Probably some time in March.
229) Message boards : Number crunching : WUs names - are there different types of WUs? (Message 744) Posted 22 Feb 2018 by Sergei Chernykh Post: In short, WU names are made random because they have to be unique. amicable_10_20 - app name 20198 - PID of the process that generated this WU 1519206302.310434 - timestamp 177 - number in the batch
230) Message boards : Number crunching : Computers list is gone? (Message 742) Posted 18 Feb 2018 by Sergei Chernykh Post: I try to keep the database from bloating. Even now, hosts.gz is 2.5 MB: https://sech.me/boinc/Amicable/stats/
231) Message boards : Number crunching : Computers list is gone? (Message 740) Posted 18 Feb 2018 by Sergei Chernykh Post: Hosts that weren't active for more than 3 months are automatically deleted.
232) Message boards : Number crunching : Computers list is gone? (Message 738) Posted 18 Feb 2018 by Sergei Chernykh Post: Are you talking about this one: https://sech.me/boinc/Amicable/top_hosts.php ? Or this one: https://sech.me/boinc/Amicable/hosts_user.php?userid=2516 ?
233) Message boards : News : More than 1000 new amicable pairs found! (Message 734) Posted 10 Feb 2018 by Sergei Chernykh Post: It's a type of amicable pair. Detailed explanation: https://en.wikipedia.org/wiki/Amicable_numbers#Regular_pairs. Short explanation: first digit is the amount of unique prime factors in the first number, second digit is the amount of unique prime factors in the second number. In this pair: - 13, 29, 193, 293, 1979, 5623, 20599 (7 factors) are unique to the first number - 73, 1249, 798503, 74397311 (4 factors) are unique to the second number
234) Message boards : Bug tracker : Amicable doesn't respect usage limits (Message 731) Posted 6 Feb 2018 by Sergei Chernykh Post: It's better to set the number of CPU cores to use and run only one task at a time. Try setting 2, 3, 4 cores to find which setting doesn't overheat your CPU. Setting the % of CPU time to use sometimes gives an invalid result for WU, and I have no idea why, so I had to turn off this setting. P.S. Actually, it shouldn't overheat when running Amicable. Amicable uses only integer arithmetic, no heavy SSE/AVX instructions at all. Check you CPU cooler.
235) Message boards : Number crunching : Drop in WU size? (Message 729) Posted 6 Feb 2018 by Sergei Chernykh Post: The speed increase for GPU over CPU depends on how well a WU can be parallelized, but it's hard to predict without actually running each WU. It's even different between various GPU models. CPUs are more stable because they don't run that many parallel threads. That's why WUs were normalized for CPUs.
236) Message boards : Random stuff : Up to date list of project badges (Message 726) Posted 1 Feb 2018 by Sergei Chernykh Post: That's definitely not an e-mail sent by the project. Check "From: " field again. It must have been someone else - free-dc.org, signature.statseb.fr or some other BOINC stats tracking site. You should ask them to fix this.
237) Message boards : Random stuff : Up to date list of project badges (Message 724) Posted 1 Feb 2018 by Sergei Chernykh Post: What Turbo badge?
238) Message boards : Number crunching : OpenCL version (Message 721) Posted 26 Jan 2018 by Sergei Chernykh Post: No, the program runs the same code for all OpenCL versions. It does matter however how old GPU drivers are. Newer drivers generally have better OpenCL compiler, and the program should run faster with them.
239) Message boards : News : More than 1000 new amicable pairs found! (Message 719) Posted 23 Jan 2018 by Sergei Chernykh Post: I take it when only one person is named as discoverer that they ran the work unit on two of their own computers ? Just checking as you didn't counter the 'by two different users' statement. Exactly. It makes no sense to repeat the name in this case.
240) Message boards : Number crunching : GPU version: kernel size tuning and less UI lags (Message 716) Posted 22 Jan 2018 by Sergei Chernykh Post: I think you got it a bit wrong. You can can increase kernel size as long as you have enough GPU RAM (monitor the usage with GPU-Z or similar tools). Bigger kernel size will mean faster processing of WUs on any GPU if it has enough memory (but more UI lags if it's attached to monitor). Kernel size = 23 should be enough for any GPU, even 1080 Ti. The way to set this option is to try different sizes and pick the largest number that still give you responsive UI and GPU load at 100% (can check it in GPU-Z). Once you have 100% GPU load, increasing kernel size further will not speed it up, because it's already at 100%.
241) Message boards : Number crunching : GPU version: kernel size tuning and less UI lags (Message 713) Posted 22 Jan 2018 by Sergei Chernykh Post: You get the same WU. The bigger kernel size just means there is less overhead on GPU because it switches to the next kernel less often. But it also makes UI laggy, if you use the same GPU to connect to your monitor.
242) Message boards : Number crunching : Drop in WU size? (Message 711) Posted 20 Jan 2018 by Sergei Chernykh Post: We've had a bunch of very GPU-friendly (highly parallelizable) workunits recently, but as far as I can tell, we're back to regular workunits again. All workunits were calibrated to have approximately the same amount of computations for CPU version. Run times on GPU are usually stable too, but sometimes they can vary a lot.
243) Message boards : News : The search up to 10^20 (Message 707) Posted 19 Jan 2018 by Sergei Chernykh Post: What would the RAM requirement be for the search to 10^21? I would guess that it wouldn't quite be a 10x jump to 15GB, but it could still be quite substantial. The memory required is O(sqrt(N)), so it will not be a 10x jump: CPU version would require ~1400 MB of memory. GPU version would require ~3700 MB of GPU memory.
244) Message boards : Random stuff : Amicable Pair Reporting Format (Message 702) Posted 29 Dec 2017 by Sergei Chernykh Post: Thanks - I presume X means exotic? ( in relation to type) Yes.
245) Message boards : News : Credit badges (Message 700) Posted 29 Dec 2017 by Sergei Chernykh Post: "Pool" user is opaque, i.e. the project doesn't know who exactly runs tasks behind pool's user account. It can only be done on corresponding pool's side.
246) Message boards : Random stuff : Amicable Pair Reporting Format (Message 697) Posted 29 Dec 2017 by Sergei Chernykh Post: The first line has format "Type Discoverer Year". You can read how the type of amicable pair is defined here: https://en.wikipedia.org/wiki/Amicable_numbers#Regular_pairs There are two discoverers for each amicable pair in this BOINC project because each work unit is processed by two different users and then the results from both users are compared. These two users are then listed in the order they completed the corresponding work unit.
247) Message boards : Bug tracker : Computation error at 6 seconds (Message 695) Posted 26 Dec 2017 by Sergei Chernykh Post: It's a 560 Ti so no it has 1024 of memory. Also I added another project and allocated this card towards it to see if same error. Did not fail, so successfully crunching with this card on another project. If that 2GB memory minimum is a project specific thing then I guess all is working as intended, must have missed the line that mentioned it. You need to have 2 GB of GPU memory to run GPU version of Amicable Numbers, so it's normal that 560 Ti errors out.
248) Message boards : Bug tracker : Computation error at 6 seconds (Message 693) Posted 24 Dec 2017 by Sergei Chernykh Post: OpenCL.cpp, line 440: clEnqueueWriteBuffer returned error -4 Error -4 is CL_MEM_OBJECT_ALLOCATION_FAILURE. What are your second GPU's specs? Does it have at least 2 GB of memory?
249) Message boards : Number crunching : Auto-ban faulty hosts (Message 691) Posted 23 Dec 2017 by Sergei Chernykh Post: It can still run CPU tasks. But if you consider only GPUs for Amicable Numbers, this host will be auto-removed after 3 months of inactivity.
250) Message boards : Number crunching : Very long WUs? (Message 686) Posted 8 Dec 2017 by Sergei Chernykh Post: more than 1 hr to finish a single wu ? with 100% cpu and gpu usage ? used to be only 15 mins, all of my cards are showing this. something wrong ??????? Actually, there is something wrong with your settings. You've set kernel size for NVIDIA GPUs to 18 which is too small for GTX 1070/1080. Set it to at least 21 here: https://sech.me/boinc/Amicable/prefs.php?subset=project P.S. I've set it to 21 for you, but you can always change it to whatever value you want. Let's hope it improves WU run times.

Previous 50 · Next 50