Linux CUDA 'Special' App finally available, featuring Low CPU use

Author	Message
Jeff Buck Volunteer tester Send message Joined: 11 Feb 00 Posts: 1441 Credit: 148,764,870 RAC: 0	Message 1889432 - Posted: 12 Sep 2017, 23:08:18 UTC - in response to Message 1889370. I had noted last evening that I had a Cuda50 task on my daily driver that got marked as Invalid for what appeared to be the same problem as cropped up with the Special App. Well, here's another WU that may be worth watching, in order to see if the processing sequence issue extends to the current stock Cuda apps as well as the Special App. The tiebreaker task came to one of my hosts this afternoon because a Cuda42 task and a v8.08 (alt) task appeared to have the same Pulse/Triplet mismatch as seen with the Special App. My machine will be running the tiebreaker overnight using x41p_zi3t2b. My guess would be that the result will match the Cuda42 result. Workunit 2671532815 (12se08aa.22109.14387.14.41.0) Task 6010530319 (S=0, A=0, P=0, T=30, G=0, BG=?) v8.00 (cuda42) windows_intelx86 Task 6010530320 (S=0, A=0, P=24, T=6, G=0, BG=0) v8.08 (alt) windows_x86_64 Yep, the Special App agreed with the Cuda42 (Task 6014562724), reporting 30 Triplets and sticking the v8.08 (alt) host with an Invalid. I also benched this one this afternoon with the stock Windows CPU app (setiathome_8.00_windows_intelx86.exe), and got 24 Pulses and 6 Triplets, which would have matched the v8.08 (alt) result. So, all in all, it appears that maybe the whole Cuda branch is using a different Pulse/Triplet processing sequence than everybody else. If Jason's around, perhaps he can shed some light on this. ID: 1889432 ·

Jeff Buck Volunteer tester Send message Joined: 11 Feb 00 Posts: 1441 Credit: 148,764,870 RAC: 0	Message 1889557 - Posted: 13 Sep 2017, 17:33:11 UTC - in response to Message 1889370. Another one worth watching might be this one, where the Special App superficially agreed with Cuda50 (at least on signal counts) but still was marked Inconclusive. Then SoG disagreed with both, setting up another tiebreaker with v8.08 (alt). I'm guessing that one will probably agree with SoG, leading to........?? Workunit 2669414696 (01mr08ac.6640.20936.13.40.0) Task 6006117152 (S=0, A=0, P=11, T=19, G=0, BG=?) v8.00 (cuda50) windows_intelx86 Task 6006117153 (S=0, A=0, P=11, T=19, G=0, BG=0) x41p_zi3v, Cuda 8.00 special Task 6008209644 (S=0, A=0, P=30, T=0, G=0, BG=0) v8.22 (opencl_nvidia_SoG) windows_intelx86 The 4th host did, indeed, get the same result as SoG, 30 Pulses, which left both the Cuda50 host and my Special App host tagged with Invalids. ID: 1889557 ·

petri33 Volunteer tester Send message Joined: 6 Jun 02 Posts: 1668 Credit: 623,086,772 RAC: 156	Message 1889641 - Posted: 13 Sep 2017, 22:32:08 UTC Here is something interesting (but this has been seen earlier too). If a cuda app (new old speciall any) finds too many signals some of them get ignored or overwritten. My latest exe with debug print out finds a lot of pulses at once. All the pulses that the official exe finds are there + some tens of others too. They are however not logged and transferred to CPU. They get overwritten/discarded. And that is OK. There is nothing scientific in microwave noise ot any other bad noise. It does not matter how many 2-70 second rechecks are needed. The recorded data is bad. This is an example of overly noisy packet. Since there are so many pulses it is natural there are triplets too. All existing apps (intel, SoG, CUDA, arm, ...) would report thousands of signals found. Some random decision has been made what to report then. Choose just N first found. Just choose any of them. They have no meaning. Here is a listing of a test run: The last decimal number on a row is the s/n ratio. The SoG reports 30 of those. I'm sure it finds all of these too. The old Cuda source has not been designed to store this much data before reporting. They all are found at one iteration. I have not counted them. And this is just one iteration. There are a lot more of them in one WU. Current WU: 01mr08ac.6640.20936.13.40.0.wu ---------------------------------------------------------------- Skipping default app setiathome_8.04_i686-pc-linux-gnu, displaying saved result(s) Elapsed Time: ....................... 7 seconds ---------------------------------------------------------------- Running app with command : .......... axo -pfb 32 --device 2 -nobs gCudaDevProps.multiProcessorCount = 20 Work data buffer for fft results size = 320864256 MallocHost G=67108864 T=33554432 P=18874368 (16) MallocHost tmp_PoTP=16777216 MallocHost tmp_PoTP2=16777216 MallocHost tmp_PoTT=16777216 MallocHost tmp_PoTG=4194304 MallocHost best_PoTP=16777216 MallocHost bestPoTG=4194304 Allocating tmp data buf for unroll 20 MallocHost tmp_smallPoT=524288 MallocHost PowerSpectrumSumMax=3145728 CUDA stream priority range: low 0 and high: -1 GPSF 3.365948 3 5.600577 AcIn 16779264 AcOut 33558528 Mallocing blockSums 24576 bytes FoundPulse@2 1.976735 0.000527 0.000071 9.570008 FoundPulse@2 1.623396 0.000540 0.000038 8.638031 FoundPulse@2 1.612676 0.000535 0.000032 8.489481 FoundPulse@2 1.458578 0.000534 0.000048 8.986246 FoundPulse@2 1.606155 0.000531 0.000029 8.399130 FoundPulse@2 1.618457 0.000541 0.000036 8.569588 FoundPulse@2 1.569507 0.000539 0.000009 7.891324 FoundPulse@2 1.844006 0.000533 0.000001 8.269540 FoundPulse@2 2.382838 0.000537 0.000037 9.580583 FoundPulse@2 1.872030 0.000534 0.000016 8.544109 FoundPulse@2 1.598400 0.000530 0.000024 8.291674 FoundPulse@2 1.600082 0.000528 0.000025 8.314984 FoundPulse@2 1.601188 0.000533 0.000026 8.330307 FoundPulse@2 2.321249 0.000536 0.000004 9.153880 FoundPulse@2 1.691558 0.000517 0.000072 9.582502 FoundPulse@2 1.696967 0.000532 0.000077 9.657459 FoundPulse@2 1.582947 0.000517 0.000016 8.077556 FoundPulse@2 1.720131 0.000540 0.000091 9.978423 FoundPulse@2 1.771624 0.000549 0.000121 10.691937 FoundPulse@2 1.663050 0.000532 0.000059 9.187489 FoundPulse@2 1.884131 0.000542 0.000022 8.662674 FoundPulse@2 1.911369 0.000525 0.000036 8.929556 FoundPulse@2 1.679549 0.000533 0.000068 9.416113 FoundPulse@2 1.635044 0.000522 0.000043 8.799426 FoundPulse@2 1.645730 0.000533 0.000050 8.947495 FoundPulse@2 1.650658 0.000539 0.000053 9.015781 FoundPulse@2 1.580639 0.000529 0.000015 8.045572 FoundPulse@2 1.604630 0.000532 0.000028 8.378003 FoundPulse@2 1.461818 0.000540 0.000051 9.049743 FoundPulse@2 1.404458 0.000546 0.000020 7.925724 FoundPulse@2 1.440044 0.000511 0.000037 8.623058 FoundPulse@2 1.435745 0.000520 0.000035 8.538831 FoundPulse@2 1.459193 0.000532 0.000049 8.998309 FoundPulse@2 1.418416 0.000539 0.000027 8.199242 FoundPulse@2 1.649374 0.000538 0.000052 8.997990 FoundPulse@2 1.666896 0.000533 0.000061 9.240775 FoundPulse@2 1.620743 0.000542 0.000037 8.601261 FoundPulse@2 1.941471 0.000537 0.000053 9.224494 FoundPulse@2 1.454564 0.000530 0.000046 8.907591 FoundPulse@2 1.428692 0.000528 0.000032 8.400620 FoundPulse@2 1.425193 0.000533 0.000031 8.332047 FoundPulse@2 1.609333 0.000514 0.000029 8.443158 FoundPulse@2 1.695943 0.000534 0.000077 9.643263 FoundPulse@2 1.520437 0.000532 0.000081 10.198433 FoundPulse@2 1.882622 0.000530 0.000021 8.647898 FoundPulse@2 1.995417 0.000544 0.000083 9.753055 FoundPulse@2 1.856272 0.000539 0.000007 8.389715 FoundPulse@2 1.918873 0.000536 0.000041 9.003082 FoundPulse@2 1.639882 0.000537 0.000047 8.866466 FoundPulse@2 1.699223 0.000525 0.000077 9.688713 FoundPulse@2 1.650857 0.000535 0.000053 9.018543 FoundPulse@2 1.582773 0.000532 0.000016 8.075144 FoundPulse@2 1.670896 0.000542 0.000064 9.296207 FoundPulse@2 1.591117 0.000526 0.000020 8.190753 FoundPulse@2 1.374322 0.000530 0.000003 7.335183 FoundPulse@2 1.595599 0.000527 0.000120 11.671307 FoundPulse@2 1.370728 0.000521 0.000001 7.264764 FoundPulse@2 1.482436 0.000532 0.000061 9.453773 FoundPulse@2 1.821798 0.000549 0.000148 11.387168 FoundPulse@2 1.690228 0.000532 0.000073 9.564079 FoundPulse@2 1.482719 0.000542 0.000062 9.459325 FoundPulse@2 2.012115 0.000529 0.000090 9.916666 FoundPulse@2 1.611341 0.000533 0.000032 8.470994 FoundPulse@2 1.826119 0.000537 0.000147 11.447035 FoundPulse@2 1.666753 0.000529 0.000061 9.238797 FoundPulse@2 1.522119 0.000520 0.000080 10.231406 FoundPulse@2 1.615047 0.000535 0.000034 8.522345 FoundPulse@2 1.601047 0.000546 0.000027 8.328350 FoundPulse@2 1.786377 0.000544 0.000127 10.896362 FoundPulse@2 1.706651 0.000539 0.000083 9.791648 FoundPulse@2 1.834663 0.000536 0.000151 11.565433 FoundPulse@2 1.723021 0.000534 0.000091 10.018472 FoundPulse@2 1.745913 0.000534 0.000103 10.335673 FoundPulse@2 1.834882 0.000530 0.000150 11.568469 FoundPulse@2 1.658360 0.000520 0.000055 9.122502 FoundPulse@2 1.212694 0.000530 0.000023 8.335871 FoundPulse@2 1.699895 0.000532 0.000079 9.698027 FoundPulse@2 1.698633 0.000541 0.000079 9.680539 FoundPulse@2 1.702630 0.000539 0.000081 9.735921 FoundPulse@2 1.175943 0.000521 0.000004 6.895549 FoundPulse@2 1.768752 0.000530 0.000115 10.652140 FoundPulse@2 1.700450 0.000528 0.000078 9.705719 FoundPulse@2 1.650478 0.000533 0.000052 9.013291 FoundPulse@2 1.898114 0.000517 0.000029 8.799682 FoundPulse@2 1.684010 0.000537 0.000071 9.477920 FoundPulse@2 1.759106 0.000535 0.000111 10.518478 FoundPulse@2 1.935882 0.000532 0.000050 9.169732 FoundPulse@2 1.936921 0.000526 0.000050 9.179910 FoundPulse@2 1.851263 0.000533 0.000159 11.795451 FoundPulse@2 1.755910 0.000539 0.000110 10.474201 FoundPulse@2 1.837544 0.000529 0.000151 11.605350 FoundPulse@2 1.474259 0.000533 0.000057 9.293543 FoundPulse@2 1.471137 0.000540 0.000056 9.232355 FoundPulse@2 1.425808 0.000546 0.000032 8.344102 FoundPulse@2 1.638894 0.000534 0.000145 12.519721 FoundPulse@2 1.848678 0.000538 0.000160 11.759622 FoundPulse@2 1.774658 0.000542 0.000121 10.733978 FoundPulse@2 1.631125 0.000514 0.000041 8.745118 FoundPulse@2 1.737183 0.000534 0.000099 10.214709 FoundPulse@2 1.599253 0.000528 0.000122 11.742907 FoundPulse@2 1.755173 0.000517 0.000105 10.463976 FoundPulse@2 1.779255 0.000532 0.000121 10.797676 FoundPulse@2 1.763561 0.000542 0.000115 10.580213 FoundPulse@2 1.805176 0.000526 0.000133 11.156842 FoundPulse@2 1.535637 0.000537 0.000090 10.496305 FoundPulse@2 1.592895 0.000535 0.000120 11.618329 FoundPulse@2 1.709927 0.000539 0.000085 9.837032 FoundPulse@2 1.405390 0.000458 0.000017 7.943988 FoundPulse@2 1.680671 0.000533 0.000069 9.431658 FoundPulse@2 1.656950 0.000522 0.000055 9.102960 FoundPulse@2 1.723063 0.000530 0.000091 10.019060 FoundPulse@2 1.666952 0.000532 0.000061 9.241556 FoundPulse@2 2.061483 0.000529 0.000116 10.400365 FoundPulse@2 1.572401 0.000533 0.000109 11.216719 FoundPulse@2 1.380236 0.000521 0.000007 7.451066 FoundPulse@2 1.341293 0.000546 0.000051 9.458194 FoundPulse@2 1.336169 0.000511 0.000045 9.316186 FoundPulse@2 1.356591 0.000533 0.000058 9.882153 FoundPulse@2 1.367106 0.000522 0.000062 10.173537 FoundPulse@2 1.349703 0.000521 0.000053 9.691266 FoundPulse@2 1.398567 0.000533 0.000080 11.045414 FoundPulse@2 1.367047 0.000533 0.000064 10.171892 FoundPulse@2 3.680955 0.000539 0.000311 13.133944 FoundPulse@2 4.710062 0.000529 0.000117 12.852029 FoundPulse@2 3.418431 0.000522 0.000164 11.847843 FoundPulse@2 1.231387 0.000531 0.000034 9.068484 FoundPulse@2 1.276593 0.000514 0.000056 10.840206 FoundPulse@2 1.171105 0.000464 0.000001 6.705898 FoundPulse@2 1.248200 0.000531 0.000043 9.727413 FoundPulse@2 1.276764 0.000514 0.000056 10.846899 FoundPulse@2 1.617722 0.000532 0.000134 12.104819 FoundPulse@2 1.724970 0.000533 0.000093 10.045484 FoundPulse@2 1.749004 0.000533 0.000106 10.378510 FoundPulse@2 1.402142 0.000442 0.000016 7.880332 FoundPulse@2 1.546361 0.000532 0.000096 10.706445 FoundPulse@2 1.529766 0.000533 0.000087 10.381251 FoundPulse@2 1.555368 0.000534 0.000101 10.882949 FoundPulse@2 1.748975 0.000534 0.000204 14.676847 FoundPulse@2 1.708569 0.000530 0.000181 13.885062 FoundPulse@2 1.478010 0.000514 0.000057 9.367037 FoundPulse@2 1.558586 0.000534 0.000103 10.946007 FoundPulse@2 1.530365 0.000539 0.000088 10.392982 FoundPulse@2 1.488839 0.000531 0.000065 9.579253 FoundPulse@2 1.503385 0.000520 0.000071 9.864291 FoundPulse@2 1.612716 0.000532 0.000131 12.006738 FoundPulse@2 1.583842 0.000542 0.000118 11.440924 FoundPulse@2 1.690363 0.000532 0.000075 9.565948 FoundPulse@2 1.519599 0.000535 0.000082 10.182016 FoundPulse@2 1.581248 0.000530 0.000114 11.390091 FoundPulse@2 1.576281 0.000525 0.000110 11.292755 FoundPulse@2 1.931715 0.000532 0.000203 12.910217 FoundPulse@2 1.624529 0.000533 0.000138 12.238214 FoundPulse@2 1.633149 0.000542 0.000144 12.407137 FoundPulse@2 1.572188 0.000533 0.000110 11.212556 FoundPulse@2 1.617853 0.000539 0.000135 12.107402 FoundPulse@2 1.570236 0.000540 0.000110 11.174293 FoundPulse@2 1.608043 0.000532 0.000129 11.915168 FoundPulse@2 1.388271 0.000459 0.000010 7.608532 FoundPulse@2 1.822513 0.000534 0.000146 11.397076 FoundPulse@2 1.647225 0.000533 0.000150 12.682960 FoundPulse@2 1.672371 0.000521 0.000064 9.316650 FoundPulse@2 1.392193 0.000441 0.000011 7.685372 FoundPulse@2 1.530179 0.000541 0.000089 10.389351 FoundPulse@2 1.583618 0.000517 0.000112 11.436528 FoundPulse@2 1.874083 0.000532 0.000270 17.128452 FoundPulse@2 1.641492 0.000533 0.000146 12.570631 FoundPulse@2 1.663845 0.000537 0.000160 13.008656 FoundPulse@2 1.687775 0.000529 0.000170 13.477576 FoundPulse@2 1.760616 0.000534 0.000211 14.904960 FoundPulse@2 1.606095 0.000521 0.000125 11.876986 FoundPulse@2 1.560110 0.000532 0.000103 10.975865 FoundPulse@2 1.462208 0.000511 0.000049 9.057384 FoundPulse@2 1.339552 0.000537 0.000092 13.307646 FoundPulse@2 1.177856 0.000443 0.000005 6.970521 FoundPulse@2 1.178306 0.000447 0.000005 6.988143 FoundPulse@2 1.203385 0.000464 0.000017 7.971035 FoundPulse@2 1.362162 0.000532 0.000104 14.193811 FoundPulse@2 1.359334 0.000539 0.000103 14.082946 FoundPulse@2 1.172183 0.000464 0.000002 6.748181 FoundPulse@2 1.191525 0.000443 0.000011 7.506200 FoundPulse@2 1.487700 0.000511 0.000063 9.556918 FoundPulse@2 1.526271 0.000520 0.000084 10.312771 FoundPulse@2 1.194648 0.000453 0.000012 7.628613 FoundPulse@2 1.387985 0.000533 0.000118 15.205826 FoundPulse@2 1.382726 0.000530 0.000114 14.999716 FoundPulse@2 1.168040 0.000436 0.000000 6.585807 FoundPulse@2 1.346147 0.000514 0.000092 13.566149 FoundPulse@2 1.340563 0.000520 0.000090 13.347272 FoundPulse@2 1.203283 0.000439 0.000016 7.967043 FoundPulse@2 1.184896 0.000457 0.000008 7.246432 FoundPulse@2 1.171934 0.000437 0.000002 6.738399 FoundPulse@2 1.169665 0.000456 0.000001 6.649496 FoundPulse@2 1.416634 0.000511 0.000087 11.546108 FoundPulse@2 1.383381 0.000531 0.000073 10.624569 FoundPulse@2 1.435164 0.000520 0.000098 12.059611 FoundPulse@2 1.186431 0.000453 0.000009 7.306552 FoundPulse@2 1.354350 0.000537 0.000100 13.887609 FoundPulse@2 1.190938 0.000456 0.000011 7.483204 FoundPulse@2 1.393475 0.000533 0.000121 15.420992 FoundPulse@2 1.396394 0.000530 0.000121 15.535420 FoundPulse@2 1.168350 0.000433 0.000000 6.597959 FoundPulse@2 1.168542 0.000461 0.000000 6.605458 FoundPulse@2 1.183443 0.000462 0.000007 7.189454 FoundPulse@2 1.401417 0.000511 0.000120 15.732285 FoundPulse@2 1.369589 0.000531 0.000107 14.484860 FoundPulse@2 1.431171 0.000520 0.000137 16.898373 FoundPulse@2 1.189886 0.000435 0.000010 7.441981 FoundPulse@2 1.184322 0.000462 0.000008 7.223927 FoundPulse@2 2.641188 0.000532 0.000293 13.129504 FoundPulse@2 2.608662 0.000517 0.000268 12.869292 FoundPulse@2 2.284326 0.000541 0.000105 10.274607 FoundPulse@2 2.489584 0.000542 0.000217 11.916672 FoundPulse@2 2.454545 0.000533 0.000194 11.636361 FoundPulse@2 2.637069 0.000537 0.000294 13.096554 FoundPulse@2 2.395074 0.000525 0.000160 11.160595 FoundPulse@2 3.296475 0.000540 0.000308 12.990826 FoundPulse@2 2.963412 0.000535 0.000127 11.106736 FoundPulse@2 3.080935 0.000546 0.000194 11.771549 FoundPulse@2 3.231863 0.000527 0.000266 12.625320 FoundPulse@2 3.135302 0.000532 0.000218 12.079088 FoundPulse@2 3.077044 0.000511 0.000179 11.749535 FoundPulse@2 3.483701 0.000531 0.000402 14.049937 FoundPulse@2 3.015670 0.000520 0.000150 11.402350 FoundPulse@2 3.205062 0.000530 0.000253 12.473715 FoundPulse@2 3.030780 0.000533 0.000162 11.487824 FoundPulse@2 4.348483 0.000520 0.000274 13.393932 FoundPulse@2 2.972420 0.000540 0.000133 11.157690 FoundPulse@2 3.183635 0.000549 0.000251 12.352501 FoundPulse@2 3.348328 0.000532 0.000331 13.284149 FoundPulse@2 4.091917 0.000544 0.000147 12.367666 FoundPulse@2 4.409654 0.000536 0.000315 13.638615 FoundPulse@2 3.165585 0.000533 0.000234 12.250395 FoundPulse@2 3.183913 0.000530 0.000242 12.354074 FoundPulse@2 4.107477 0.000539 0.000154 12.429907 FoundPulse@2 3.449244 0.000534 0.000386 13.855015 FoundPulse@2 3.732662 0.000534 0.000537 15.458272 FoundPulse@2 3.520773 0.000530 0.000421 14.259648 FoundPulse@2 2.763338 0.000541 0.000365 14.106700 FoundPulse@2 3.159050 0.000538 0.000233 12.213431 FoundPulse@2 3.585775 0.000533 0.000458 14.627350 FoundPulse@2 3.011280 0.000540 0.000497 16.090240 FoundPulse@2 2.857469 0.000535 0.000411 14.859750 FoundPulse@2 2.708736 0.000546 0.000338 13.669887 FoundPulse@2 2.212057 0.000456 0.000056 9.696456 FoundPulse@2 2.126701 0.000447 0.000016 9.013610 FoundPulse@2 2.944295 0.000527 0.000450 15.554360 FoundPulse@2 2.875532 0.000521 0.000409 15.004254 FoundPulse@2 2.717790 0.000532 0.000334 13.742314 FoundPulse@2 2.203531 0.000453 0.000051 9.628248 FoundPulse@2 3.281728 0.000533 0.000296 12.907405 FoundPulse@2 3.385394 0.000537 0.000354 13.493824 FoundPulse@2 3.525179 0.000529 0.000422 14.284567 FoundPulse@2 2.707284 0.000511 0.000316 13.658269 FoundPulse@2 2.895343 0.000531 0.000428 15.162742 FoundPulse@2 2.674402 0.000520 0.000304 13.395212 FoundPulse@2 3.179533 0.000532 0.000241 12.329303 FoundPulse@2 3.212977 0.000542 0.000264 12.518492 FoundPulse@2 3.005086 0.000526 0.000146 11.342480 FoundPulse@2 3.109256 0.000517 0.000198 11.931750 FoundPulse@2 3.469300 0.000532 0.000395 13.968467 FoundPulse@2 3.534985 0.000517 0.000418 14.340042 FoundPulse@2 3.282642 0.000514 0.000286 12.912571 FoundPulse@2 3.279329 0.000534 0.000295 12.893832 FoundPulse@2 3.324021 0.000520 0.000311 13.146646 FoundPulse@2 2.931068 0.000530 0.000445 15.448545 FoundPulse@2 2.728113 0.000528 0.000337 13.824901 FoundPulse@2 2.819107 0.000533 0.000389 14.552854 FoundPulse@2 2.237934 0.000456 0.000068 9.903471 FoundPulse@2 2.130280 0.000456 0.000018 9.042243 FoundPulse@2 2.969148 0.000540 0.000475 15.753187 FoundPulse@2 3.179366 0.000549 0.000598 17.434927 FoundPulse@2 3.118658 0.000532 0.000547 16.949265 FoundPulse@2 2.139271 0.000459 0.000023 9.114169 FoundPulse@2 2.131280 0.000443 0.000018 9.050243 FoundPulse@2 3.360310 0.000537 0.000341 13.351932 FoundPulse@2 3.093683 0.000525 0.000193 11.843657 FoundPulse@2 3.001695 0.000535 0.000147 11.323295 FoundPulse@2 3.402549 0.000544 0.000368 13.590869 FoundPulse@2 3.387249 0.000539 0.000356 13.504322 FoundPulse@2 3.525861 0.000536 0.000428 14.288427 FoundPulse@2 3.183043 0.000533 0.000243 12.349156 FoundPulse@2 3.561772 0.000539 0.000450 14.491570 FoundPulse@2 3.127786 0.000529 0.000212 12.036577 FoundPulse@2 3.058472 0.000534 0.000517 16.467772 FoundPulse@2 3.394793 0.000534 0.000697 19.158344 FoundPulse@2 3.171607 0.000530 0.000573 17.372852 FoundPulse@2 2.866142 0.000533 0.000414 14.929137 FoundPulse@2 2.759463 0.000522 0.000350 14.075701 FoundPulse@2 2.866398 0.000530 0.000411 14.931184 FoundPulse@2 2.106157 0.000433 0.000007 8.849252 FoundPulse@2 2.242244 0.000463 0.000071 9.937949 FoundPulse@2 2.922145 0.000538 0.000448 15.377163 FoundPulse@2 3.310622 0.000533 0.000651 18.484968 FoundPulse@2 2.888442 0.000542 0.000433 15.107538 FoundPulse@2 2.163557 0.000457 0.000034 9.308453 FoundPulse@2 1.787007 0.000465 0.000038 8.903969 FoundPulse@2 3.195474 0.000533 0.000589 17.563795 FoundPulse@2 2.932799 0.000537 0.000453 15.462390 FoundPulse@2 3.130942 0.000529 0.000550 17.047537 FoundPulse@2 3.055107 0.000532 0.000175 11.625443 FoundPulse@2 3.083886 0.000539 0.000193 11.788239 FoundPulse@2 2.940140 0.000532 0.000452 15.521116 FoundPulse@2 3.057419 0.000542 0.000524 16.459349 FoundPulse@2 2.899826 0.000526 0.000426 15.198611 FoundPulse@2 2.136022 0.000441 0.000020 9.088177 FoundPulse@2 2.932218 0.000517 0.000436 15.457743 FoundPulse@2 3.173748 0.000532 0.000576 17.389982 FoundPulse@2 3.429989 0.000517 0.000693 19.439911 FoundPulse@2 2.182720 0.000443 0.000041 9.461760 FoundPulse@2 2.816949 0.000514 0.000374 14.535591 FoundPulse@2 2.932202 0.000534 0.000450 15.457618 FoundPulse@2 3.006984 0.000520 0.000477 16.055868 FoundPulse@2 2.201365 0.000439 0.000049 9.610922 FoundPulse@2 1.723444 0.000463 0.000008 8.184834 FoundPulse@2 1.839752 0.000441 0.000059 9.500705 FoundPulse@2 1.783345 0.000443 0.000034 8.862534 FoundPulse@2 2.953725 0.000533 0.000461 15.629800 FoundPulse@2 3.158204 0.000539 0.000575 17.265629 FoundPulse@2 2.940977 0.000529 0.000450 15.527819 FoundPulse@2 2.154262 0.000442 0.000029 9.234093 FoundPulse@2 2.229566 0.000437 0.000061 9.836531 FoundPulse@2 3.265495 0.000544 0.000640 18.123959 FoundPulse@2 2.936312 0.000539 0.000456 15.490493 FoundPulse@2 3.354506 0.000536 0.000678 18.836046 FoundPulse@2 3.201926 0.000537 0.000598 17.615412 FoundPulse@2 2.880461 0.000525 0.000415 15.043683 FoundPulse@2 2.836920 0.000535 0.000399 14.695362 FoundPulse@2 2.134585 0.000439 0.000020 9.076682 FoundPulse@2 1.790466 0.000441 0.000038 8.943100 FoundPulse@2 2.161813 0.000464 0.000033 9.294498 FoundPulse@2 2.879083 0.000532 0.000420 15.032666 FoundPulse@2 3.131023 0.000541 0.000564 17.048182 FoundPulse@2 2.868010 0.000539 0.000420 14.944081 FoundPulse@2 2.293484 0.000437 0.000089 10.347869 FoundPulse@2 1.885244 0.000437 0.000079 10.015388 FoundPulse@2 2.615463 0.000539 0.000490 18.276882 FoundPulse@2 1.831125 0.000453 0.000057 9.403111 FoundPulse@2 1.778937 0.000439 0.000032 8.812670 FoundPulse@2 1.219372 0.000461 0.000004 7.019907 FoundPulse@2 1.379671 0.000462 0.000032 8.590966 FoundPulse@2 1.148649 0.000462 0.000002 6.727070 FoundPulse@2 1.348988 0.000457 0.000017 7.896694 FoundPulse@2 1.349797 0.000441 0.000017 7.915010 FoundPulse@2 1.347705 0.000436 0.000016 7.867673 FoundPulse@2 1.315705 0.000458 0.000002 7.143591 FoundPulse@2 1.216560 0.000458 0.000003 6.929918 FoundPulse@2 1.212920 0.000458 0.000001 6.813433 FoundPulse@2 1.345507 0.000456 0.000016 7.817935 FoundPulse@2 1.155155 0.000458 0.000005 7.021531 FoundPulse@2 1.348878 0.000441 0.000017 7.894209 FoundPulse@2 1.253445 0.000451 0.000019 8.110238 FoundPulse@2 1.321100 0.000437 0.000004 7.265654 FoundPulse@2 1.207769 0.000464 0.000010 7.433360 FoundPulse@2 1.213769 0.000440 0.000012 7.648015 FoundPulse@2 1.204557 0.000458 0.000008 7.318470 FoundPulse@2 1.207852 0.000458 0.000010 7.436326 FoundPulse@2 1.201369 0.000454 0.000007 7.204398 processPulseResults Elapsed Time : ...................... 2 seconds Speed compared to default : ......... 350 % ----------------- Comparing results ------------- R1:R2 ------------ ------------- R2:R1 ------------ Exact Super Tight Good Bad Exact Super Tight Good Bad Spike 0 0 0 0 0 0 0 0 0 0 Gaussian 0 0 0 0 0 0 0 0 0 0 Pulse 0 11 11 11 19 0 11 11 11 0 Triplet 0 0 0 0 0 0 0 0 0 19 Best Spike 0 0 0 0 0 0 0 0 0 0 Best Gaussian 0 0 0 0 0 0 0 0 0 0 Best Pulse 0 0 0 0 0 0 0 0 0 0 Best Triplet 0 0 0 0 0 0 0 0 0 0 ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- 0 11 11 11 19 0 11 11 11 19 Unmatched signal(s) in R1 at line(s) 642 667 703 729 755 780 805 831 867 903 933 963 988 1023 1059 1094 1129 1165 1201 Unmatched signal(s) in R2 at line(s) 342 359 376 393 410 427 444 461 478 495 512 529 546 563 580 597 614 631 648 For R1:R2 matched signals only, Q= 99.83% Result : Different. ---------------------------------------------------------------- Done with 01mr08ac.6640.20936.13.40.0.wu 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 ID: 1889641 ·

baron_iv Volunteer tester Send message Joined: 4 Nov 02 Posts: 109 Credit: 104,905,241 RAC: 0	Message 1889646 - Posted: 13 Sep 2017, 22:46:26 UTC Hey Petri, if I sent you an AMD GPU (AMD 290 or R9 Fury), could you come up with a "special" app that helps AMD GPUs perform better (on Windows or Linux, I don't care which)? I'd love to be able to fully utilize my R9 Fury GPUs to the extent that my Nvidia GTX 1070s are used. My 1070s are performing almost 2x as well as my 2 R9 Fury cards, but the Fury cards have over 7100 cores (vs "only" 3800 for the 1070s). -baron_iv Proud member of: GPU Users Group ID: 1889646 ·

petri33 Volunteer tester Send message Joined: 6 Jun 02 Posts: 1668 Credit: 623,086,772 RAC: 156	Message 1889688 - Posted: 14 Sep 2017, 3:58:13 UTC - in response to Message 1889646. @baron_iv Thanks for the offer, but I do have a windows laptop that I could try with. It has a amd gpu. For now I do not think I have enough time to set up a development environment, the latest source and get to know how it works. Maybe next summer... 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 ID: 1889688 ·

baron_iv Volunteer tester Send message Joined: 4 Nov 02 Posts: 109 Credit: 104,905,241 RAC: 0	Message 1889714 - Posted: 14 Sep 2017, 11:04:18 UTC - in response to Message 1889688. @baron_iv Thanks for the offer, but I do have a windows laptop that I could try with. It has a amd gpu. For now I do not think I have enough time to set up a development environment, the latest source and get to know how it works. Maybe next summer... Well, it's an open offer. If you change your mind, let me know and thank you for all the wonderful work that you've done on the NVidia app. :) -baron_iv Proud member of: GPU Users Group ID: 1889714 ·

Stephen "Heretic" Volunteer tester Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628	Message 1889803 - Posted: 14 Sep 2017, 20:54:06 UTC . . Hi people, . . Here is an interesting one ..... http://setiathome.berkeley.edu/result.php?resultid=6015653373 . . The two CUDA hosts came up as overflows yet greatly disagree with each other, while the two 8.08(alt) hosts came up as valid tasks with far less hits. . . Cuda50 30 spikes . . Cuda 80 S-17, P-6, T-7 . . 8.08 Alt S-1, P-9, T-2 Stephen :( ? ID: 1889803 ·

Mike Volunteer tester Send message Joined: 17 Feb 01 Posts: 34258 Credit: 79,922,639 RAC: 80	Message 1889811 - Posted: 14 Sep 2017, 21:13:04 UTC - in response to Message 1889803. . . Hi people, . . Here is an interesting one ..... http://setiathome.berkeley.edu/result.php?resultid=6015653373 . . The two CUDA hosts came up as overflows yet greatly disagree with each other, while the two 8.08(alt) hosts came up as valid tasks with far less hits. . . Cuda50 30 spikes . . Cuda 80 S-17, P-6, T-7 . . 8.08 Alt S-1, P-9, T-2 Stephen :( ? Well thats well known. The cuda 5 is run by a old 9500 so i would say its the GPU. The CPU app is more precise. With each crime and every kindness we birth our future. ID: 1889811 ·

Wiggo Send message Joined: 24 Jan 00 Posts: 34744 Credit: 261,360,520 RAC: 489	Message 1889815 - Posted: 14 Sep 2017, 21:15:24 UTC - in response to Message 1889803. . . Hi people, . . Here is an interesting one ..... http://setiathome.berkeley.edu/result.php?resultid=6015653373 . . The two CUDA hosts came up as overflows yet greatly disagree with each other, while the two 8.08(alt) hosts came up as valid tasks with far less hits. . . Cuda50 30 spikes . . Cuda 80 S-17, P-6, T-7 . . 8.08 Alt S-1, P-9, T-2 Stephen :( ? Well the rig that did the Cuda50 task is a Win10 machine running a pre-Fermi GPU so it just trashes all GPU work with false overflows, but why your task overflowed I have no idea. Cheers. ID: 1889815 ·

Stephen "Heretic" Volunteer tester Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628	Message 1889818 - Posted: 14 Sep 2017, 21:30:15 UTC - in response to Message 1889815. Well the rig that did the Cuda50 task is a Win10 machine running a pre-Fermi GPU so it just trashes all GPU work with false overflows, but why your task overflowed I have no idea. Cheers. . . Clearly I did not look closely enough at that host, surely he should be running only Cuda32 ... :(, {or 23} Stephen :( ID: 1889818 ·

Keith Myers Volunteer tester Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873	Message 1889819 - Posted: 14 Sep 2017, 21:31:39 UTC I have been getting the consensus opinion that all CUDA derived apps behave similarly but differ from the CPU and SoG apps in pulse search order which leads to invalids. And the solution is to have all apps perform the pulse search in the same order. Now the question is ...... which developer needs to change their app to match the preferred norm? Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) ID: 1889819 ·

Wiggo Send message Joined: 24 Jan 00 Posts: 34744 Credit: 261,360,520 RAC: 489	Message 1889822 - Posted: 14 Sep 2017, 21:52:34 UTC . . Clearly I did not look closely enough at that host, surely he should be running only Cuda32 ... :(, {or 23} Stephen That person really needs to update to a post-Fermi GPU or go back to Win7 where there are drivers that properly support that old card. ;-) ....Now the question is ...... which developer needs to change their app to match the preferred norm? Obviously the Cuda apps seeing as they're not matching the CPU app and the SoG app is. This also means that tasks that have been validated by Cuda only apps are suspect. Cheers. ID: 1889822 ·

Jeff Buck Volunteer tester Send message Joined: 11 Feb 00 Posts: 1441 Credit: 148,764,870 RAC: 0	Message 1889824 - Posted: 14 Sep 2017, 21:57:00 UTC - in response to Message 1889803. Last modified: 14 Sep 2017, 22:05:41 UTC . . Hi people, . . Here is an interesting one ..... http://setiathome.berkeley.edu/result.php?resultid=6015653373 . . The two CUDA hosts came up as overflows yet greatly disagree with each other, while the two 8.08(alt) hosts came up as valid tasks with far less hits. . . Cuda50 30 spikes . . Cuda 80 S-17, P-6, T-7 . . 8.08 Alt S-1, P-9, T-2 Stephen :( ? It appears you got bitten by the Restart bug that seems to exist in x41p_zi3v. If you look at the signals listed in your Stderr up until the task restarted, you'll see they pretty much match one of the validated results. However, after the restart the app started reporting lots of Spikes, and then Triplets, which aren't found in the validated tasks. I've been getting something similar on my host 8289033 ever since I started running the Special App back in late June. Because my crunch-only machines shut down every weekday afternoon and then restart later in the evening, that machine averages about 3 of those Invalids each week (out of a possible 20 restarted tasks), although mine have always been all Spikes or all Triplets (with that "peak=-nan" value), not a mixed bag like yours. I don't see that bug on my other two Linux boxes which are running x41p_zi3t2b, so it seems to be specific to the latest version. EDIT: I have to qualify that last statement about the bug not existing in zi3t2b. I just looked at one of my earlier posts on the topic (Message 1875523) and see that it has happened once on those two machines with that version. Still much better than 3 per week on the single machine running the later version. ID: 1889824 ·

Keith Myers Volunteer tester Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873	Message 1889828 - Posted: 14 Sep 2017, 22:20:37 UTC Jeff, would increasing your task save interval to a value longer than the typical task completion time reduce your invalids upon restart. If the task has to start from scratch each time you wouldn't see the error ... right? Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) ID: 1889828 ·

Stephen "Heretic" Volunteer tester Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628	Message 1889829 - Posted: 14 Sep 2017, 22:22:21 UTC - in response to Message 1889824. It appears you got bitten by the Restart bug that seems to exist in x41p_zi3v. If you look at the signals listed in your Stderr up until the task restarted, you'll see they pretty much match one of the validated results. However, after the restart the app started reporting lots of Spikes, and then Triplets, which aren't found in the validated tasks. I've been getting something similar on my host 8289033 ever since I started running the Special App back in late June. Because my crunch-only machines shut down every weekday afternoon and then restart later in the evening, that machine averages about 3 of those Invalids each week (out of a possible 20 restarted tasks), although mine have always been all Spikes or all Triplets (with that "peak=-nan" value), not a mixed bag like yours. I don't see that bug on my other two Linux boxes which are running x41p_zi3t2b, so it seems to be specific to the latest version. EDIT: I have to qualify that last statement about the bug not existing in zi3t2b. I just looked at one of my earlier posts on the topic (Message 1875523) and see that it has happened once on those two machines with that version. Still much better than 3 per week on the single machine running the later version. . . Ahh, a confluence of traps. One host running running an antiquated GPU with the wrong drivers and the wrong version of the app, and my machine doing a restart. That was probably yesterday when I shut it down to remove the 3rd GPU. . . I wonder if someone should mention to the owner of the other host that they need to do something with their rig? Otherwise they will have a flood of dud results. Stephen ?? ID: 1889829 ·

Jeff Buck Volunteer tester Send message Joined: 11 Feb 00 Posts: 1441 Credit: 148,764,870 RAC: 0	Message 1889832 - Posted: 14 Sep 2017, 22:38:54 UTC - in response to Message 1889828. Jeff, would increasing your task save interval to a value longer than the typical task completion time reduce your invalids upon restart. If the task has to start from scratch each time you wouldn't see the error ... right? Well, that's been Tbar's suggestion for dealing with it. However, there's a tradeoff in lost processing time going that route. Currently I have my checkpoint interval set to 120 seconds. That means, on average, that every time BOINC restarts, each restarted task has to back up about half that interval, or 60 seconds. With 4 GPU tasks and 3 CPU tasks running on that machine, it means I already lose about 7 minutes of total processing time each weekday. Doubling or tripling the checkpoint interval might reduce the Invalids, but at a corresponding cost of additional lost processing time, which likely would exceed the processing time lost by the 2-3 tasks that get marked Invalid each week. A better solution would probably be to back up to the earlier release of the Special App, since the problem seems to be less prominent with that version. I just haven't gotten around to making that change. ID: 1889832 ·

Keith Myers Volunteer tester Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873	Message 1889841 - Posted: 14 Sep 2017, 23:52:14 UTC - in response to Message 1889832. Which leads to the other question I had. You picked up the zi3t2b version directly from Petri, correct? Or was that posted somewhere for download that I missed. I am using the x41p_zi3v version from TBar that I picked up from Crunchers Anonymous Linux thread. I too am running the 120 second task save interval as a compromise. I will have to look again at my inconclusives and analyze whether my increase is from restart reordering of the search. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) ID: 1889841 ·

Stephen "Heretic" Volunteer tester Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628	Message 1889847 - Posted: 15 Sep 2017, 0:16:54 UTC - in response to Message 1889841. Which leads to the other question I had. You picked up the zi3t2b version directly from Petri, correct? Or was that posted somewhere for download that I missed. I am using the x41p_zi3v version from TBar that I picked up from Crunchers Anonymous Linux thread. I too am running the 120 second task save interval as a compromise. I will have to look again at my inconclusives and analyze whether my increase is from restart reordering of the search. . . If you look back in this thread TBar posted a link for that version, but I am not sure if that is still valid. Stephen ?? ID: 1889847 ·

Keith Myers Volunteer tester Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873	Message 1889853 - Posted: 15 Sep 2017, 0:59:13 UTC - in response to Message 1889847. Last modified: 15 Sep 2017, 1:08:09 UTC Thanks Stephen, guess I missed it. Will do the search. [Edit] Guess I missed that version which must be earlier. I thought it was the latest based on the numbering. From that link, I ended up with the zi3v version. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) ID: 1889853 ·

Jeff Buck Volunteer tester Send message Joined: 11 Feb 00 Posts: 1441 Credit: 148,764,870 RAC: 0	Message 1889855 - Posted: 15 Sep 2017, 1:13:53 UTC - in response to Message 1889853. You can find a link to Petri's version in Message 1861418. ID: 1889855 ·

©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.