Message boards :
Number crunching :
The Highest Ranked SETI AMD Host is a MAC: Time for a STOCK MAC APP?
Message board moderation
Previous · 1 . . . 17 · 18 · 19 · 20 · 21 · Next
Author | Message |
---|---|
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
That's odd. The version number has gone up by 2, from 2867 to 2869 yet there aren't any changes listed... So you're hinting that there is something wrong with downloading the latest files from Berkeley's own site? Is there a problem downloading from here, https://setisvn.ssl.berkeley.edu/trac/browser/branches/sah_v7_opt BTW, if the entire folder doesn't download all you have to do is drop down one level and download the folder contents. Tracking down and installing yet more repetitive third party software is not something some people desire. I just spent days eliminating a faulty install of FFTW, I'd like to refrain from installing anything not absolutely needed. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14654 Credit: 200,643,578 RAC: 874 |
That's odd. The version number has gone up by 2, from 2867 to 2869 yet there aren't any changes listed... SVN is just a mechanism for keeping a set of resource files up-to-date on your computer. It doesn't "install" anything - what you do with the resource files after updating them is entirely up to you. Once you have the basic file-set replicated on your computer (which, admittedly, is a big download), only changes (differences) need to be downloaded, which is much, much quicker than downloading the whole darn kit'n'caboodle from scratch every time. And then you can move onto stage 2, which I've barely scratched the surface of myself. For example you can find out who changed what, when. You can compare the previous version with the current version, and - if you're lucky - you can see why the author decided to make that particular change (that does rely on the author remembering to explain the change clearly, though - not all do, or at least not everyone does it every time). For further details (my bedtime looms), try reading http://en.wikipedia.org/wiki/Apache_Subversion. Edit - on second thoughts, that looks a lot heavier than most simple Wikipedia introductions to subjects. Maybe one of the other regular users can come up with a better introduction? 2nd (and positively last) edit on the way to bed. a) I've just downloaded the final three changes to bring my copy of the source code fully up to date. It took 5 seconds. b) The procedure is: use SVN to keep a local copy of everything fully up-to-date - but keep that copy sacrosanct on another partition/drive/computer, well away from any experiments with a compiler. Then make a working scratchpad copy you can fiddle with to your heart's content - that's an 'export' from the local master copy synchronised from Berkeley, and a damn sight faster than a fresh download. That's what I've learned by working on the Lunatics Installer. |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
The entire zipped copy of sah_v7_opt.zip is less than 30mb and takes less than 10 seconds to download. For now, I think I will stick to just downloading the latest version when needed. |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
Any luck on assigning the WG to Apple GPUs yet? If there is a workaround in the repository somewhere has anyone found it? Wait a little, under investigation for now. |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
Any luck on assigning the WG to Apple GPUs yet? If there is a workaround in the repository somewhere has anyone found it? You might want to look at this one as well, it also reports "max WG size: 1024", although it appears to work; Not using ap_cmdline.txt-file, using commandline options. shmget in attach_shmem: Invalid argument 05:04:17 (88348): Can't set up shared mem: -1. Will run in standalone mode. Illegal value for gpu_device_num: -1 in BOINC Client 0.0.0 WARNING: boinc_get_opencl_ids failed with code -33 OpenCL platform detected: Apple WARNING: BOINC supplied wrong platform! Number of OpenCL devices found : 3 BOINC assigns slot on device #0. WARNING: BOINC failed to provide OpenCL device, using own enumeration abilities Used GPU device parameters are: Number of compute units: 14 Single buffer allocation size: 256MB Total device global memory: 1024MB max WG size: 1024 local mem type: Real -unroll default value used: 14 -ffa_block default value used: 3584 -ffa_block_fetch default value used: 1792 AstroPulse v7.01 Darwin 10.7+ 64 bit, rel. Rev 1864, OpenCL version by Raistmer, GPU mode V7, by Raistmer ported to OS X by Lunatics.kwsn.net team. Build features: Non-graphics OpenCL OPENCL_WRITE COMBINED_DECHIRP_KERNEL SMALL_CHIRP_TABLE TWIN_FFA FFTW BLANKIT USE_INCREASED_PRECISION SSE4.1 64bit System: Darwin x86_64 Kernel: 12.6.0 CPU : Intel(R) Xeon(R) CPU E5462 @ 2.80GHz GenuineIntel x86, Family 6 Model 23 Stepping 6 Features : FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE MCA CMOV PAT PSE36 CLFSH DS ACPI MMX FXSR SSE SSE2 SS HTT TM PBE SSE3 DTES64 MON DSCPL VMX EST TM2 SSSE3 CX16 TPR PDCM SSE4.1 Number of OpenCL platforms: 1 OpenCL Platform Name: Apple Number of devices: 3 Max compute units: 14 Max work group size: 1024 Max clock frequency: 900Mhz Max memory allocation: 268435456 Cache type: None Cache line size: 0 Cache size: 0 Global memory size: 1073741824 Constant buffer size: 65536 Max number of constant args: 8 Local memory type: Scratchpad Local memory size: 32768 Queue properties: Out-of-Order: No Name: ATI Radeon Barts XT Prototype Vendor: AMD Driver version: 1.0 Version: OpenCL 1.1 Max compute units: 14 Max work group size: 1024 Max clock frequency: 775Mhz Max memory allocation: 268435456 Cache type: None Cache line size: 0 Cache size: 0 Global memory size: 1073741824 Constant buffer size: 65536 Max number of constant args: 8 Local memory type: Scratchpad Local memory size: 32768 Queue properties: Out-of-Order: No Name: ATI Radeon Barts PRO Prototype Vendor: AMD Driver version: 1.0 Version: OpenCL 1.1 Max compute units: 14 Max work group size: 1024 Max clock frequency: 775Mhz Max memory allocation: 268435456 Cache type: None Cache line size: 0 Cache size: 0 Global memory size: 1073741824 Constant buffer size: 65536 Max number of constant args: 8 Local memory type: Scratchpad Local memory size: 32768 Queue properties: Out-of-Order: No Name: ATI Radeon Barts PRO Prototype Vendor: AMD Driver version: 1.0 Version: OpenCL 1.1 INFO: can't open binary kernel file: .//AstroPulse_Kernels_r1864.cl_ATIRadeonBartsXTPrototype.bin_V7_TWIN_FFA_12.6.0_10, continue with recompile... Info : Building Program (binary, clBuildProgram):main kernels: OK code 0 INFO: binary kernel file created WARNING: can't open binary kernel file for oclFFT plan: .//AP_clFFTplan_ATIRadeonBartsXTPrototype_32768_gr64_lr16_wg256_tw0_r1864.bin_12.6.0_10, continue with recompile... single pulses: 0 repetitive pulses: 9 percent blanked: 63.03 Rep. pulse: num_std_devs=7.085 peak_power=2436105 dm=-3072 peak_bin=512 scale=7 ffa_scale=2 period=55.0985 Rep. pulse: num_std_devs=9.825 peak_power=980516.5 dm=2944 peak_bin=11264 scale=7 ffa_scale=2 period=137.7462 Rep. pulse: num_std_devs=8.737 peak_power=491489.5 dm=-2944 peak_bin=14848 scale=7 ffa_scale=2 period=275.4913 Rep. pulse: num_std_devs=8.603 peak_power=979302.1 dm=3072 peak_bin=28672 scale=7 ffa_scale=3 period=275.4903 Rep. pulse: num_std_devs=7.544 peak_power=490654.6 dm=-2944 peak_bin=0 scale=7 ffa_scale=1 period=137.7467 Rep. pulse: num_std_devs=9.826 peak_power=1955434 dm=-2944 peak_bin=0 scale=7 ffa_scale=3 period=137.7451 Rep. pulse: num_std_devs=7.253 peak_power=246268.9 dm=3072 peak_bin=8192 scale=7 ffa_scale=1 period=275.4924 Rep. pulse: num_std_devs=6.666 peak_power=2435438 dm=-3072 peak_bin=512 scale=7 ffa_scale=1 period=27.54915 Rep. pulse: num_std_devs=7.222 peak_power=4866263 dm=3072 peak_bin=512 scale=7 ffa_scale=2 period=27.54894 remove_radar total=9.942E+08 , N=1 , <>=9.942E+08 , min=9.942E+08 , max=9.942E+08 main_loop_L1 total=1.167E+11 , N=2 , <>=5.837E+10 , min=5.835E+10 , max=5.840E+10 FFT_forward total=1.949E+08 , N=2352 , <>=8.286E+04 , min=2.062E+04 , max=4.025E+05 remove_radar_randomize total=0.000E+00 , N=0 , <>=0.000E+00 , min=1.845E+19 , max=0.000E+00 build_chirp_table total=0.000E+00 , N=0 , <>=0.000E+00 , min=1.845E+19 , max=0.000E+00 DataWrite total=0.000E+00 , N=0 , <>=0.000E+00 , min=1.845E+19 , max=0.000E+00 DataWrite_ns total=0.0000E+00, N=0 , <>=0 , min=0 , max=0 oclReadBuf total=0.000E+00 , N=0 , <>=0.000E+00 , min=1.845E+19 , max=0.000E+00 ChirpWrite total=0.000E+00 , N=0 , <>=0.000E+00 , min=1.845E+19 , max=0.000E+00 ChirpWrite_ns total=0.0000E+00, N=0 , <>=0 , min=0 , max=0 dechirp total=1.148E+08 , N=2352 , <>=4.881E+04 , min=2.699E+04 , max=8.592E+04 Dechirp_ns total=0.0000E+00, N=0 , <>=0 , min=0 , max=0 Half_ns total=0.0000E+00, N=0 , <>=0 , min=0 , max=0 PC_single_pulse_kernel_FFA_update total=7.850E+10 , N=2352 , <>=3.338E+07 , min=3.088E+07 , max=6.946E+07 PC_ns total=0.0000E+00, N=0 , <>=0 , min=0 , max=0 oclReadBuf total=0.000E+00 , N=0 , <>=0.000E+00 , min=1.845E+19 , max=0.000E+00 oclWriteBuf total=0.000E+00 , N=0 , <>=0.000E+00 , min=1.845E+19 , max=0.000E+00 FFT_inverse total=0.000E+00 , N=0 , <>=0.000E+00 , min=1.845E+19 , max=0.000E+00 ffa total=3.393E+10 , N=18 , <>=1.885E+09 , min=7.572E+08 , max=1.087E+10 FFA blocks counters: FFA_fetch total=0.000E+00 , N=0 , <>=0.000E+00 , min=1.845E+19 , max=0.000E+00 FFA_tt_build total=0.000E+00 , N=0 , <>=0.000E+00 , min=1.845E+19 , max=0.000E+00 FFA_compare total=1.926E+08 , N=40 , <>=4.815E+06 , min=8.445E+05 , max=2.608E+07 FFA_coadd total=3.191E+07 , N=2526 , <>=1.263E+04 , min=6.993E+03 , max=6.266E+04 FFA_stride_add total=3.589E+06 , N=36 , <>=9.968E+04 , min=6.511E+04 , max=1.184E+05 GPU_buffer_read_backs total=0.0000E+00, N=0 , <>=0 , min=0 , max=0 TWIN_FFA USE_OPENCL OPENCL_WRITE USE_INCREASED_PRECISION SMALL_CHIRP_TABLE COMBINED_DECHIRP_KERNEL BLANKIT rev 1864 GPU device sync requested... ...GPU device synched 05:05:03 (88348): called boinc_finish(0) Why ap_7.01, is the Blankit folder the correct folder? At least the benchmark appears to work for AP; KWSN-OSX-APbench v2.0.24 Running on TomsMacPro.local at Wed Mar 11 09:03:29 2015 ------------------------------------------------------------------------- Starting benchmark run... ------------------------------------------------------------------------- Listing wu-file(s) in /testWUs : sigind_v5.wu Listing executable(s) in /testAPPs : ap_7.01r1864_sse41_clGPU_x86_64-apple-darwin Listing executable in /refAPPs : astropulse_7.04_x86_64-apple-darwin__opencl_ati_mac Listing additional reference results in /testData/refResults : No additional reference results in /testData/refResults found. ------------------------------------------------------------------------- Current WU: sigind_v5.wu ------------------------------------------------------------------------- Running default app with command : ... astropulse_7.04_x86_64-apple-darwin__opencl_ati_mac Elapsed Time: ........................ 48 seconds ------------------------------------------------------------------------- Running app with command : ........... ap_7.01r1864_sse41_clGPU_x86_64-apple-darwin Elapsed Time: ........................ 48 seconds Speed compared to default : .......... 100 % ----------------- Comparing results Actual results : ./testData/ref-pulse.astropulse_7.04_x86_64-apple-darwin__opencl_ati_mac.sigind_v5.wu.out: <ap_signal>19,<pulses>9,<best_pulses>10 ./pulse.out: <ap_signal>19,<pulses>9,<best_pulses>10 All Signals: Checked 19, 19 , Strongly Similar Pulses: Checked 9, 9 , Strongly Similar Best Pulses: Checked 10, 10 , Strongly Similar -(./testData/ref-pulse.astropulse_7.04_x86_64-apple-darwin__opencl_ati_mac.sigind_v5.wu.out)- Reportable Single Pulses: 0 [OK], 0 above threshold*THRESHOLD_FUDGE Reportable Repeating Pulses: 9 [OK] Single Pulses (Best): 0 [OK], 0 above threshold*THRESHOLD_FUDGE -(./pulse.out)- Reportable Single Pulses: 0 [OK], 0 above threshold*THRESHOLD_FUDGE Reportable Repeating Pulses: 9 [OK] Single Pulses (Best): 0 [OK], 0 above threshold*THRESHOLD_FUDGE I really didn't expect much change with the GPU AP App, the CPU App will be interesting... |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
Seems to be hanging on the AP CPU App; ld: warning: directory not found for option '-L/opt/local/lib' Undefined symbols for architecture x86_64: "Astropulse::number_of_logged_signals", referenced from: Astropulse::add_signal(std::vector<ap_signal, std::allocator<ap_signal> >&, ap_signal) in ap_client-ap_fileio.o ld: symbol(s) not found for architecture x86_64 clang: error: linker command failed with exit code 1 (use -v to see invocation) make: [ap_client] Error 1 (ignored) /bin/cp ap_client ap_7.01r1864_sse41_x86_64-apple-darwin cp: ap_client: No such file or directory make: [ap_7.01r1864_sse41_x86_64-apple-darwin] Error 1 (ignored) strip ap_7.01r1864_sse41_x86_64-apple-darwin error: strip: can't open file: ap_7.01r1864_sse41_x86_64-apple-darwin (No such file or directory) make: [ap_7.01r1864_sse41_x86_64-apple-darwin] Error 1 (ignored) /bin/rm -f astropulse-7.01_x86_64-apple-darwin.debug /bin/ln ap_client astropulse-7.01_x86_64-apple-darwin.debug ln: ap_client: No such file or directory make: [astropulse-7.01_x86_64-apple-darwin.debug] Error 1 (ignored) make all-am There doesn't seem to be an ap_signal file. Strange it didn't have any problems building the GPU App. If you use the other folder, "AP", it will build you a nice AP CPU 6.09 App...if you ever need an APv6 CPU App. But it hangs on the APv7 CPU App... |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
Ok, try revision: 2870 for OpenCL MultiBeam. That WG size limiting code wasn't on my workcopy too so I added it. Also other improvements. Regarding AP: If you download whole folder you need to delete AP folder after and rename AP_BLANKIT folder to AP one. All this should be not needed if SVN would be used properly. AP and AP_BLANKIT belongs to different branches actually. |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
And after that check revision 2871 - will it work on OS X ? It has more generalized approach to kernel launch geometry, but implements this approach only on FindSpike32 kernel. All others will use max device limit for now. |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
Ok, try revision: 2870 for OpenCL MultiBeam. Nice work. 2870 is running in standalone in Mountain Lion. I'll have to test it awhile before seeing if it will work in BOINC; 20:07:52 (60901): Can't open init data file - running in standalone mode 20:07:52 (60901): Can't open init data file - running in standalone mode Not using mb_cmdline.txt-file, using commandline options. 20:07:52 (60901): Can't open init data file - running in standalone mode WARNING: init_data.xml missing OpenCL platform detected: Apple WARNING: BOINC supplied wrong platform! Number of OpenCL devices found : 3 BOINC assigns slot on device #0. WARNING: BOINC failed to provide OpenCL device, using own enumeration abilities Build features: SETI7 Non-graphics OpenCL USE_OPENCL_HD5xxx OCL_CHIRP3 FFTW SSE4.1 64bit System: Darwin x86_64 Kernel: 12.6.0 CPU : Intel(R) Xeon(R) CPU E5462 @ 2.80GHz GenuineIntel x86, Family 6 Model 23 Stepping 6 OpenCL-kernels filename : MultiBeam_Kernels_r2870.cl INFO: can't open binary kernel file: .//MultiBeam_Kernels_r2870.clHD5_ATIRadeonBartsXTPrototype.bin_V7_12.6.0_10, continue with recompile... Info : Building Program (binary, clBuildProgram):main kernels: OK code 0 INFO: binary kernel file created WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_524288_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_8_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_16_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_32_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_64_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_128_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_256_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_512_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_1024_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_2048_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_4096_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_8192_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_16384_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_32768_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_65536_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: .//MB_clFFTplan_ATIRadeonBartsXTPrototype_131072_gr64_lr16_wg256_tw0_r2870.bin_12.6.0_10, continue with recompile... ar=1.185405 NumCfft=101385 NumGauss=0 NumPulse=56257780816 NumTriplet=56257780816 Currently allocated 121 MB for GPU buffers In v_BaseLineSmooth: NumDataPoints=1048576, BoxCarLength=8192, NumPointsInChunk=32768 OS X optimized S@H v7 application (based on S@H Enhanced by Alex Kan) Version info: SSE4.1ux (Intel, Core 2-optimized v8-nographics) V5.13 by Alex Kan SSE4.1ux OS X 64bit Build 2870 , Ported by : Raistmer, JDWhale, Urs Echternacht OpenCL version by Raistmer, r2870 AMD HD5 version by Raistmer Number of OpenCL platforms: 1 OpenCL Platform Name: Apple Number of devices: 3 Max compute units: 14 Max work group size: 1024 Max clock frequency: 900Mhz Max memory allocation: 268435456 Cache type: None Cache line size: 0 Cache size: 0 Global memory size: 1073741824 Constant buffer size: 65536 Max number of constant args: 8 Local memory type: Scratchpad Local memory size: 32768 Queue properties: Out-of-Order: No Name: ATI Radeon Barts XT Prototype Vendor: AMD Driver version: 1.0 Version: OpenCL 1.1 Max compute units: 14 Max work group size: 1024 Max clock frequency: 775Mhz Max memory allocation: 268435456 Cache type: None Cache line size: 0 Cache size: 0 Global memory size: 1073741824 Constant buffer size: 65536 Max number of constant args: 8 Local memory type: Scratchpad Local memory size: 32768 Queue properties: Out-of-Order: No Name: ATI Radeon Barts PRO Prototype Vendor: AMD Driver version: 1.0 Version: OpenCL 1.1 Max compute units: 14 Max work group size: 1024 Max clock frequency: 775Mhz Max memory allocation: 268435456 Cache type: None Cache line size: 0 Cache size: 0 Global memory size: 1073741824 Constant buffer size: 65536 Max number of constant args: 8 Local memory type: Scratchpad Local memory size: 32768 Queue properties: Out-of-Order: No Name: ATI Radeon Barts PRO Prototype Vendor: AMD Driver version: 1.0 Version: OpenCL 1.1 Work Unit Info: ............... Credit multiplier is : 2.85 WU true angle range is : 1.185405 Used GPU device parameters are: Number of compute units: 14 Single buffer allocation size: 64MB Total device global memory: 1024MB max WG size: 256 local mem type: Real period_iterations_num=23 Spike: peak=24.47872, time=83.89, d_freq=1421126615.33, chirp=0.2514, fft_len=64k Spike: peak=24.07581, time=87.24, d_freq=1421126616.18, chirp=0.25232, fft_len=128k Spike: peak=25.0097, time=87.24, d_freq=1421126616.19, chirp=0.25325, fft_len=128k Spike: peak=25.16989, time=87.24, d_freq=1421126616.2, chirp=0.25417, fft_len=128k Spike: peak=24.65972, time=83.89, d_freq=1421126615.35, chirp=0.2551, fft_len=64k Spike: peak=24.59426, time=87.24, d_freq=1421126616.2, chirp=0.2551, fft_len=128k Spike: peak=24.38239, time=6.711, d_freq=1421125858.53, chirp=-8.7435, fft_len=128k Spike: peak=24.15315, time=6.711, d_freq=1421125858.53, chirp=-8.7444, fft_len=128k Spike: peak=24.46851, time=6.711, d_freq=1421125858.55, chirp=-8.7527, fft_len=128k Spike: peak=24.65365, time=6.711, d_freq=1421125858.54, chirp=-8.7537, fft_len=128k Spike: peak=24.40533, time=6.711, d_freq=1421119059.91, chirp=-14.735, fft_len=128k Spike: peak=24.47265, time=6.711, d_freq=1421119059.9, chirp=-14.735, fft_len=128k I think the task it's on now is the 19/2 shorty. Yes, it found the correct numbers; Spike count: 19 Autocorr count: 2 Pulse count: 0 Triplet count: 0 Gaussian count: 0 Time cpu in use since last restart: 102.8 seconds GPU device sync requested... ...GPU device synched 20:17:00 (60901): called boinc_finish(0) Success! |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
Check next rev please too. |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
Seems to be a number of differences with r2871. The first difference is a new numbered .cl file is being created. You no longer have to use the already present MultiBeam_Kernels.cl and add numbers to it. In standalone testing I noticed the max WG size is still listed as 1024 instead of the old 256 in r2870. When running in BOINC it seems to take much less time to build the binaries, about half as long. So far it appears to be working well, possibly a little slower though. We'll have to see how it goes, seems the new builds start out slow but speed up after a while. You can see the results here; http://setiathome.berkeley.edu/results.php?hostid=6796479&offset=160&show_names=0&state=0&appid=0 Thanks. |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
Well, I've tried about 4 different builds and it doesn't appear to get any better than the current build. When comparing the results to r2839 this build produces almost the exact same numbers and I've run thousands of r2839 tasks. The only problem I've noticed is you can't use over 128 on the settings or it will produce errors on launch. I'm not sure if that's even a problem and it seems to work best with low numbers anyway, just as the other Mac MBv7 Apps I've tested. What I haven't tested is running more than one task at a time and I think I'll leave that for others to test. Just as I'll leave it to others to comment on any screen lag ;-) |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
...Regarding AP: If you download whole folder you need to delete AP folder after and rename AP_BLANKIT folder to AP one. All this should be not needed if SVN would be used properly. AP and AP_BLANKIT belongs to different branches actually. It's still hanging at; checking if clang supports -c -o file.o... (cached) yes Strange it doesn't have a problem making the AP GPU App. I made the AP GPU App, ran make clean, ran the AP CPU CL, then make -i -k, and it says; "Astropulse::number_of_logged_signals", referenced from: Astropulse::add_signal(std::vector<ap_signal, std::allocator<ap_signal> >&, ap_signal) in ap_client-ap_fileio.o ld: symbol(s) not found for architecture x86_64 clang: error: linker command failed with exit code 1 (use -v to see invocation) make: [ap_client] Error 1 (ignored) /bin/cp ap_client ap_7.01r2871_sse41_x86_64-apple-darwin cp: ap_client: No such file or directory make: [ap_7.01r2871_sse41_x86_64-apple-darwin] Error 1 (ignored) strip ap_7.01r2871_sse41_x86_64-apple-darwin error: strip: can't open file: ap_7.01r2871_sse41_x86_64-apple-darwin (No such file or directory) make: [ap_7.01r2871_sse41_x86_64-apple-darwin] Error 1 (ignored) /bin/rm -f astropulse-7.01_x86_64-apple-darwin.debug /bin/ln ap_client astropulse-7.01_x86_64-apple-darwin.debug ln: ap_client: No such file or directory make: [astropulse-7.01_x86_64-apple-darwin.debug] Error 1 (ignored) make all-am g++ -I/opt/local/include -mmacosx-version-min=10.7 -O3 -mtune=generic -I./ -I./Users/Tom/sah_v7_opt/src -I./Users/Tom/sah_v7_opt/AKv8 -march=corei7 -msse4.1 -mmacosx-version-min=10.7 -O3 -mtune=generic -I./ -I./Users/Tom/sah_v7_opt/src -I./Users/Tom/sah_v7_opt/AKv8 -DHAVE_CONFIG_H -DTEXT_UI -DNDEBUG -DCLIENT -I../server/db -I/Users/Tom/sah_v7_opt/AKv8/../src -I/Users/Tom/sah_v7_opt/AKv8 -I/Users /Tom/sah_v7_opt/AKv8/db -I/Users/Tom/sah_v7_opt/AKv8/client -I/Users/Tom/boinc -I/Users/Tom/boinc/api -I/Users/Tom/boinc/lib -I/Users /Tom/boinc/sched -I/Users/Tom/boinc/db -D_THREAD_SAFE -D_THREAD_SAFE -mmacosx-version-min=10.7 -ldl -lz -lpthread -framework Carbon -lm -L/opt/local/lib -mmacosx-version-min=10.7 -ldl -lz -lpthread -framework Carbon -o ap_client ap_client-sqlblob.o ap_client-sqlrow.o ap_client-xml_util.o ap_client-lcgamm.o ap_client-ap_schema.o ap_client-ap_client_main.o ap_client-ap_science.o ap_client-ap_fileio.o ap_client-ap_fold.o ap_client-ap_timer.o ap_client-ap_debug.o ap_client-mtrand.o ap_client-ap_version.o ap_client-malloc_a.o ap_client GPU_lock.o ap_client-sbtf.o ap_client-ap_shmem.o ap_client-ap_remove_radar.o -L/Users/Tom/boinc/api -L/Users/Tom/boinc/api/.libs -lboinc_api -L/Users/Tom/boinc/lib -L/Users/Tom/boinc/lib/.libs -lboinc -lfftw3f ../../lib/OSX64/libfftw3f.a -lz -lstdc++ -lm -lpthread -ldl /Users/Tom/sah_v7_opt/lib/OSX64/libfftw3f.a -lz -lstdc++ -lm -lpthread -ldl /Users/Tom/sah_v7_opt/lib/OSX64/libfftw3f.a ld: warning: directory not found for option '-L/opt/local/lib' Undefined symbols for architecture x86_64: "Astropulse::number_of_logged_signals", referenced from: Astropulse::add_signal(std::vector<ap_signal, std::allocator<ap_signal> >&, ap_signal) in ap_client-ap_fileio.o ld: symbol(s) not found for architecture x86_64 clang: error: linker command failed with exit code 1 (use -v to see invocation) make[1]: [ap_client] Error 1 (ignored) TomsMacPro:client Tom$ Which is the same as before... |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
I also had the same problem with the CPU AP build in Mavericks. You can build the AP GPU App in ML & Mavericks, but the CPU build hangs with the same error. The MBv7 ATI5 GPU App seems to work better in Mavericks. Considering the AP creation rate, you might need the MBv7 ATI5 App for a while. I posted it here; MBv7_r2871_ati5.zip Works for me... |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
After 2 weeks MBv7_r2871_ati5 is still working fine. It's now up to Consecutive valid tasks: 5025. Considering some "Lost" tasks are going to start expiring in the next day or so, I guess that is about as high as the Consecutive valid tasks will go. It would be nice if someone was to enable resend lost tasks for a few hours a week. As far as I can tell, the ATI OSX MB App works about as well as the ATI MB App in Windows, which is much better than the OSX ATI MB App on Beta at present... ;-) |
Dena Wiltsie Send message Joined: 19 Apr 01 Posts: 1628 Credit: 24,230,968 RAC: 26 |
I think you have made the big time. I went over to Beta for a few AP work units and when they ran out I was about to switch back to production when I noticed I was getting MB units on my graphic processor. The graphic processor seems to be about five times faster that the CPU on my system and the few records I have looked at indicate the application compares well with the other already existing MB applications. I am not sure if this is right but this should be the link to my work units on Beta if you are interested. Thanks for your effort to make this possible. |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
Unfortunately the ATI App r2728 on Beta at present was one of Joe Fox's early attempts. His later version r2760 was much better and was tested months ago, 2760 running 2 tasks at once. The main problem with version 2760 was Low GPU load and High CPU usage. The Lower Angle Range tasks ran Slow and running 2 at a time produced massive screen Lag. The newer Apps are better and produce good GPU load while running 1 at a time. According to Raistmer's post we should see newer Apps soon, Let's hope they will emerge here as stock ones quite soon. The latest version of my Apps at Crunchers Anonymous are the best yet on my MacPro. We'll see how the others work. It was a frustrating start, but it appears to have worked out rather nicely ;-) |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
It's good to see that the Beta MB apps seem to be working. At least the last few that have been processed today. Both Intel and Nvidia GPUs are crunching away after that full day of all errored out MBs yesterday. Was worried it was the apps until I looked at the history of all those MBs and saw every different configure (CPU, ATI, NV, Intel) all errored so it was a problem with the work units themselves. I would have posted in Beta but there isn't a MB Mac OS discussion thread. Just in case anyone with an iMac was wondering, lol Zalster |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
That didn't stop me, http://setiweb.ssl.berkeley.edu/beta/forum_thread.php?id=2182&postid=53953#53953. As you mentioned, it's not really a Mac problem. That place is a minefield, I'm aborting many more tasks than I'm running. Here's a comparison between the r2728 that's on Beta and my latest r2871 both at 0.416 Angle Range. MBv7 2871 http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=6950487 Run time: 22 min 24 sec CPU time: 7 min 6 sec MBv7 2728 http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=6949536 Run time: 59 min 3 sec CPU time: 33 min 25 sec That's a fair difference in my book. Dena is a Wingperson on the 2871 task. |
Dena Wiltsie Send message Joined: 19 Apr 01 Posts: 1628 Credit: 24,230,968 RAC: 26 |
That didn't stop me, http://setiweb.ssl.berkeley.edu/beta/forum_thread.php?id=2182&postid=53953#53953. As you mentioned, it's not really a Mac problem. That place is a minefield, I'm aborting many more tasks than I'm running. I will crunch that work unit shortly. It seems like my RAC is only about 10,000 crunching MB units where I could hit 20,000 crunching AP units. Of that number, about 5,000 credits come from crunching MB units on the CPU. |
©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.