Message boards :
Number crunching :
Seti@home v8
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 . . . 10 · Next
Author | Message |
---|---|
Jord Send message Joined: 9 Jun 99 Posts: 15184 Credit: 4,362,181 RAC: 3 |
I posted mine for my HD7870 in the other thread. It's a combination of Lunatics OpenCL GPU only applications for Multibeam and Astropulse, have added the new stuff to the app_info.xml and it's working for me. |
tbret Send message Joined: 28 May 99 Posts: 3380 Credit: 296,162,071 RAC: 40 |
My GPU's did not find 30 autocorrelations with these: Take this with a grain of salt: If I remember correctly, 30 (of anything) was decided as an upper-limit so that if a program found 30, chances are the processing was wrong and it would abort. Therefore; you're computations are probably more correct than the ones with "30" showing as an autocorrelation amount (which I am at a loss to define). Please note that I said,"probably." But I feel fairly safe saying the the other computer's process was terminated due to an abnormally high count indicating a crunching error. Yours may yet be correct when checked against another machine's result. |
tullio Send message Joined: 9 Apr 04 Posts: 8797 Credit: 2,930,782 RAC: 1 |
I have v8 tasks running on a 32-bit Linux box and a 64-bit Windows 10 PC. The Linux box has no screensaver while the Windows 10 PC shows a nice screensaver for SETI@home. Tullio |
Iona Send message Joined: 12 Jul 07 Posts: 790 Credit: 22,438,118 RAC: 0 |
I thought I had, but obviously not, as no tasks were showing when I restarted BOINC. Reverted to stock and predictably, no tasks showing....all gone. Project reset....yet more V8 CPU tasks. However, there were now 191 V8 tasks listed as in progress for one of my PCs, in spite of the reset and me only seeing about 60! Those 191 V8 tasks, will now just have to time out - as my patience did. I also noticed that my S@H preferences no longer seemed to work under the latest BOINC - if they had, ironically, not one V8 task would've even downloaded!!! I've given up and removed the project completely. Don't take life too seriously, as you'll never come out of it alive! |
Marco Franceschini Send message Joined: 4 Jul 01 Posts: 54 Credit: 69,877,354 RAC: 135 |
I have v8 tasks running on a 32-bit Linux box and a 64-bit Windows 10 PC. The Linux box has no screensaver while the Windows 10 PC shows a nice screensaver for SETI@home. Hi Tullio...beware that screensaver it'a cache memory "destructor" !!!! Marco. |
tullio Send message Joined: 9 Apr 04 Posts: 8797 Credit: 2,930,782 RAC: 1 |
Well, I have a 20 GB RAM on that PC which runs SETI both CPU 8 and GPU 7, vLHC@home, Atlas@home at 4 GB/task, CMS-dev and the CERN Challenge. I monitor memory usage by the Task Manager. I forgot, also CPDN with its yearly deadline. Tullio |
Marco Franceschini Send message Joined: 4 Jul 01 Posts: 54 Credit: 69,877,354 RAC: 135 |
Well, I have a 20 GB RAM on that PC which runs SETI both CPU 8 and GPU 7, vLHC@home, Atlas@home at 4 GB/task, CMS-dev and the CERN Challenge. I monitor memory usage by the Task Manager. I forgot, also CPDN with its yearly deadline. Not central memory but cpu and now gpu cache memory depletion due to limits in size for both L1 and L2 level. Marco. |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
Except there ISN'T any Version 8 CUDA Source code out yet, making it Impossible to have a Version 8 CUDA App. Tried compiling in Yosemite but I'm getting the same Error in a number of sections; /usr/include/dispatch/object.h(143): error: expected an identifier /usr/include/dispatch/object.h(362): error: identifier "dispatch_block_t" is undefined Because of those errors it ends with; clang: error: no such file or directory: 'cudaAcceleration.o' clang: error: no such file or directory: 'cudaAcc_CalcChirpData.o' clang: error: no such file or directory: 'cudaAcc_fft.o' clang: error: no such file or directory: 'cudaAcc_gaussfit.o' clang: error: no such file or directory: 'cudaAcc_PowerSpectrum.o' clang: error: no such file or directory: 'cudaAcc_pulsefind.o' clang: error: no such file or directory: 'cudaAcc_summax.o' clang: error: no such file or directory: 'cudaAcc_transpose.o' clang: error: no such file or directory: 'cudaAcc_utilities.o' clang: error: no such file or directory: 'cudaAcc_autocorr.o' make[2]: [seti_cuda] Error 1 (ignored) Later today I'll try it in Mountain Lion. |
petri33 Send message Joined: 6 Jun 02 Posts: 1668 Credit: 623,086,772 RAC: 156 |
Except there ISN'T any Version 8 CUDA Source code out yet, making it Impossible to have a Version 8 CUDA App. See if this helps ... http://stackoverflow.com/questions/27976312/how-to-cope-with-non-gcc-compatible-code-in-os-x-yosemite-core-headers/28014302#28014302 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 |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
See if this helps ... No help, same Errors. Hmmmm, what's this; Seti Cuda MB, INITIAL v8 compatibility updates... Would it help if I used those files? I suppose at this point it couldn't hurt. |
tullio Send message Joined: 9 Apr 04 Posts: 8797 Credit: 2,930,782 RAC: 1 |
My CPU is an AMD A10-6700 at 3.7 GHz but can reach 4.14 GHz. Its L1 is 192 KB, L2 4 MB. Present graphic board is Radeon HD 8470, plus an onboard graphic 8670D Devastator. I am not sure about their caches, but I plan to upgrade to a Geforge GTX 750. I have seen in stderr.txt that SETI@home GPU tasks use both, while Einstein@home GPU tasks see only the onboard graphic. But I am not a fan of graphic boards, since the CERN projects use only VirtualBox. CPDN too uses only CPU. Tullio |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14654 Credit: 200,643,578 RAC: 874 |
Hmmmm, what's this; Do heed the WARNINGS (and other notes) |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
See if this helps ... Running just the new files from r3312 I get one early Error; analyzeFuncs.cpp:1320:75: error: use of undeclared identifier 'FILE_VERSION' retval=outfile.printf("<setiathome_version>%s</setiathome_version>\n",FILE_VERSION); ^ 3 warnings and 1 error generated. make[2]: [seti_cuda-analyzeFuncs.o] Error 1 (ignored) I can't find where FILE_VERSION is declared.....anyone? |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14654 Credit: 200,643,578 RAC: 874 |
I can't find where FILE_VERSION is declared.....anyone? In the changeset we've both linked, it's defined in win_build/seti_boinc_private.h (line 14, just above the copyright date) Other builds, and other branches, will probably have a similar header file - I'll edit here after switching to another machine which has a fully updated repository. The primary source is line 14 of https://setisvn.ssl.berkeley.edu/trac/browser/seti_boinc/client/win_build/seti_boinc_private.h Jason just copied that to the matching location in his build tree, as you will have to. |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
I can't find where FILE_VERSION is declared.....anyone? I just added it to the analyzeFuncs.cpp includes and that seemed to work. Then it ran all the way to the end and gave the same response Urs and myself have been receiving with many v8 builds; ld: symbol(s) not found for architecture x86_64 clang: error: linker command failed with exit code 1 (use -v to see invocation) make[2]: [seti_cuda] Error 1 So, I guess I'll just have to switch to Mountain Lion and try it there. |
gs Send message Joined: 18 May 99 Posts: 45 Credit: 5,412,660 RAC: 8 |
What is the justification for the jpg files? I doubt that they are required for number crunching. Can I omit them? |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14654 Credit: 200,643,578 RAC: 874 |
What is the justification for the jpg files? I doubt that they are required for number crunching. Can I omit them? The question arises from my sample app_info files, in the sticky "for information only" thread at the top of the board. I asked for discussion to take place outside that thread, so here we are. The sample app_info files were posted in two versions, one without graphics (shorter and simpler), and the other with graphics available for display via the command button in BOINC Manager, or as a screensaver. If you choose the 'with graphics' version, then the jpg files have to be included - they form an integral part of the display. BOINC will probably refuse to run the application at all, with an error message about the missing files. But if you just use the simple version without eye candy, the jpg files aren't needed either. |
gs Send message Joined: 18 May 99 Posts: 45 Credit: 5,412,660 RAC: 8 |
Aaaah, now I got you. My bad, I misinterpreted the "graphics"-thing with the GPU option. Of course, it's still too early for that. |
Urs Echternacht Send message Joined: 15 May 99 Posts: 692 Credit: 135,197,781 RAC: 211 |
I just added it to the analyzeFuncs.cpp includes and that seemed to work. Then it ran all the way to the end and gave the same response Urs and myself have been receiving with many v8 builds; So it compiles with clang. That's one down and one more to go. Which version of clang are you using on OS X ML ? I have clang 3.2 (425.xxx), which does not link our apps. Is there a newer version available in the OS X open source downloads section ? _\|/_ U r s |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
I just added it to the analyzeFuncs.cpp includes and that seemed to work. Then it ran all the way to the end and gave the same response Urs and myself have been receiving with many v8 builds; I'm using the one that comes with Xcode 5.1.1, it says clang 5.1 (clang-503.0.40). I haven't had any luck with boinc-master 7.7 though. Once I went back to ver 7.5 the compiles would finish. I just removed the NV cards, copied the sah_v7_opt from Yosemite to ML, ran make, and it compiled on the first try. I now have another setiathome_x41zc_x86_64-apple-darwin_cuda65 this time compiled from the fresh Xbranch 3312. Of course I'll have to reinstall the cards and boot to Yosemite to try it, it should work though. I suppose it would be best to start again from a fresh folder and run the compile again....and try Petri's files too. If you have an Apple ID you can download all the different versions of the Tools from the download site. First you'll have to login but eventually you'll want to get to here; https://developer.apple.com/downloads/ From there you just look for Xcode 5.1.1. I'm using the version of command line tools that came with the Xcode download. The link probably won't work, but I'm using this one in ML; http://adcdownload.apple.com/Developer_Tools/xcode_5.1.1/xcode_5.1.1.dmg |
©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.