Message boards :
Number crunching :
SETI@home v8.12 Windows GPU applications support thread
Message board moderation
Previous · 1 . . . 13 · 14 · 15 · 16 · 17 · Next
Author | Message |
---|---|
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
Nah, I'm getting the massive overflows from tasks from those tape periods on r3525 also. It isn't just you. . . PHEW! . . :) Stephen . |
Kissagogo27 Send message Joined: 6 Nov 99 Posts: 715 Credit: 8,032,827 RAC: 62 |
Hi everyone, i've the same with R3330 ... |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
Hi everyone, i've the same with R3330 ... . . Thanks for replying, I'm glad it isn;t just me. Stephen |
AMDave Send message Joined: 9 Mar 01 Posts: 234 Credit: 11,671,730 RAC: 0 |
Not 90%, but more than noticeable with r3430, and cuda50. |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
. . Hello Raistmer, . . I was wondering if you had RC status for v8.19/r3528 yet? Stephen . |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
I hope to get them released at Tuesday. Up to Eric after that. SETI apps news We're not gonna fight them. We're gonna transcend them. |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
I hope to get them released at Tuesday. Up to Eric after that. . . That's good news, thanks for your hard work. Stephen . |
BilBg Send message Joined: 27 May 07 Posts: 3720 Credit: 9,385,827 RAC: 0 |
The only strange thing of r3528 compared with r3500 (beta vs main) is the report/usage of "LotOfMem path" (Radeon HD 6570 1024MB / Catalyst 11.12 / Windows XP) http://setiweb.ssl.berkeley.edu/beta/show_host_detail.php?hostid=72429 http://setiathome.berkeley.edu/show_host_detail.php?hostid=4832843 Both apps run on default (no cmdline) r3528 reports "LotOfMem path: yes" http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=24879952 Used GPU device parameters are: Number of compute units: 6 Single buffer allocation size: 128MB Total device global memory: 512MB max WG size: 256 local mem type: Real LotOfMem path: yes LowPerformanceGPU path: no HighPerformanceGPU path: no period_iterations_num=50 r3500 "LotOfMem path: no" http://setiathome.berkeley.edu/result.php?resultid=5122876433 Used GPU device parameters are: Number of compute units: 6 Single buffer allocation size: 128MB Total device global memory: 512MB max WG size: 256 local mem type: Real LotOfMem path: no LowPerformanceGPU path: no HighPerformanceGPU path: no period_iterations_num=50 Notes: r3528 is from: http://boinc2.ssl.berkeley.edu/beta/download/setiathome_8.19_windows_intelx86__opencl_ati5_sah.exe http://boinc2.ssl.berkeley.edu/beta/download/MultiBeam_Kernels_r3528.cl bd3682cea1edc9f8973f233839b1fd39 *setiathome_8.19_windows_intelx86__opencl_ati5_sah.exe 277bd652f71c7de826a90064c9280860 *MultiBeam_Kernels_r3528.cl r3500 is from: c9080d2b6e940dd6736f737df6260247 *Lunatics_Win32_v0.45_Beta4-for-SoG_setup.exe dfcf205436ce6b1c85fa35a9a4a66bf8 *MB8_win_x86_SSE2_OpenCL_ATi_HD5_r3500.exe 277bd652f71c7de826a90064c9280860 *MultiBeam_Kernels_r3500.cl (MultiBeam_Kernels_r3528.cl == MultiBeam_Kernels_r3500.cl) Â - ALF - "Find out what you don't do well ..... then don't do it!" :) Â |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
r3528 is SoG while r3500 isn't. "Lot of mem" path has meaning only for SoG. SETI apps news We're not gonna fight them. We're gonna transcend them. |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
r3528 is SoG while r3500 isn't. . . Now I am confused. If r3500 is not SoG what is it?? Stephen . |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
r3528 is SoG while r3500 isn't. It's non-SoG AMD build: Build features: SETI8 Non-graphics OpenCL USE_OPENCL_HD5xxx OCL_ZERO_COPY OCL_CHIRP3 FFTW AMD specific USE_SSE2 x86 CPUID: AMD Athlon(tm) II X3 455 Processor Cache: L1=64K L2=512K CPU features: FPU TSC PAE CMPXCHG8B APIC SYSENTER MTRR CMOV/CCMP MMX FXSAVE/FXRSTOR SSE SSE2 HT SSE3 SSE4A OpenCL-kernels filename : MultiBeam_Kernels_r3500.cl ar=0.429339 NumCfft=195489 NumGauss=1101398310 NumPulse=226362612175 NumTriplet=452728333399 Currently allocated 209 MB for GPU buffers In v_BaseLineSmooth: NumDataPoints=1048576, BoxCarLength=8192, NumPointsInChunk=32768 Windows optimized setiathome_v8 application Based on Intel, Core 2-optimized v8-nographics V5.13 by Alex Kan SSE2xj Win32 Build 3500 , Ported by : Raistmer, JDWhale SETI8 update by Raistmer OpenCL version by Raistmer, r3500 AMD HD5 version by Raistmer SETI apps news We're not gonna fight them. We're gonna transcend them. |
BilBg Send message Joined: 27 May 07 Posts: 3720 Credit: 9,385,827 RAC: 0 |
r3528 is SoG while r3500 isn't. You mean that setiathome_8.19_windows_intelx86__opencl_ati5_sah.exe is ATI AMD SoG build? I didn't expected "Total device global memory: 512MB" to be considered "LotOfMem" EDIT: OK, I now see "SIGNALS_ON_GPU" SETI8 Non-graphics OpenCL USE_OPENCL_HD5xxx OCL_ZERO_COPY SIGNALS_ON_GPU OCL_CHIRP3 FFTW AMD specific USE_SSE2 x86 Â - ALF - "Find out what you don't do well ..... then don't do it!" :) Â |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
Yes. And well, it means app considered enough memory to chose that path. BTW, try to experiment if selection works properly - increase -sbs value - will you see switch to non-LoM path? SETI apps news We're not gonna fight them. We're gonna transcend them. |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
r3528 is SoG while r3500 isn't. . . OK I will not pretend I understand what all that means apart from basically it is a CPU based OpenCL app (am I right?). But it is running pretty well on my Nvidia cards. It is running on all my rigs with GT730, GTX950 and 2 x GTX970. So confused or not I am impressed :) Stephen . |
BilBg Send message Joined: 27 May 07 Posts: 3720 Credit: 9,385,827 RAC: 0 |
   OK I will not pretend I understand what all that means apart from basically it is a CPU based OpenCL app (am I right?). No - I posted 2 links to tasks run on my ATI AMD GPU: Radeon HD 6570 on Computer 4832843 and the same physical Computer 72429 on SETI@home Beta. Read second line of my post - "second line" of text, not the empty one ;) : http://setiathome.berkeley.edu/forum_thread.php?id=79760&postid=1820358#1820358 "USE_OPENCL_HD5xxx" and "AMD HD5" means ATI AMD Radeon HD 5xxx series (and above) But it is running pretty well on my Nvidia cards The MultiBeam_Kernels_r35xx.cl file is the same for ATI AMD and NVIDIA but the .exe is different.  - ALF - "Find out what you don't do well ..... then don't do it!" :)  |
BilBg Send message Joined: 27 May 07 Posts: 3720 Credit: 9,385,827 RAC: 0 |
BTW, try to experiment if selection works properly - increase -sbs value - will you see switch to non-LoM path? Пробовал, не берьет ;) The switch -sbs is "accepted" but have no real effect if > of the default 128 All of: Maximum single buffer size set to:256MB Maximum single buffer size set to:144MB Maximum single buffer size set to:160MB Maximum single buffer size set to:136MB Maximum single buffer size set to:130MB Maximum single buffer size set to:320MB Result in: Currently allocated 201 MB for GPU buffers ... Restarted at x.xx percent. Used GPU device parameters are: Number of compute units: 6 Single buffer allocation size: 128MB Total device global memory: 512MB max WG size: 256 local mem type: Real LotOfMem path: yes LowPerformanceGPU path: no HighPerformanceGPU path: no period_iterations_num=50 and the video RAM usage is the same: 236 MB (from ~20 at just Desktop) -sbs really works for me only if < 128 Maximum single buffer size set to:120MB ... ar=0.012633 NumCfft=146033 NumGauss=0 NumPulse=50210701440 NumTriplet=68004583584 Currently allocated 193 MB for GPU buffers In v_BaseLineSmooth: NumDataPoints=1048576, BoxCarLength=8192, NumPointsInChunk=32768 Restarted at 14.61 percent. Used GPU device parameters are: Number of compute units: 6 Single buffer allocation size: 120MB Total device global memory: 512MB max WG size: 256 local mem type: Real LotOfMem path: yes LowPerformanceGPU path: no HighPerformanceGPU path: no period_iterations_num=50 Maximum single buffer size set to:32MB ... OpenCL-kernels filename : MultiBeam_Kernels_r3528.cl ar=0.012633 NumCfft=146033 NumGauss=0 NumPulse=50210701440 NumTriplet=68004583584 Currently allocated 105 MB for GPU buffers In v_BaseLineSmooth: NumDataPoints=1048576, BoxCarLength=8192, NumPointsInChunk=32768 Restarted at 15.63 percent. Used GPU device parameters are: Number of compute units: 6 Single buffer allocation size: 32MB Total device global memory: 512MB max WG size: 256 local mem type: Real LotOfMem path: yes LowPerformanceGPU path: no HighPerformanceGPU path: no period_iterations_num=50  - ALF - "Find out what you don't do well ..... then don't do it!" :)  |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
So,it just can't allow more. And older revision, non-SoG one? SETI apps news We're not gonna fight them. We're gonna transcend them. |
BilBg Send message Joined: 27 May 07 Posts: 3720 Credit: 9,385,827 RAC: 0 |
And older revision, non-SoG one? The same. The previous Beta task: 27jl16aa.5587.171137.8.42.195.vlar_1 This here: blc5_2bit_guppi_57449_46372_HIP81348_OFF_0020.18939.416.17.26.223.vlar_1 Video RAM used 220 MB (by "ATI MemoryViewer" (31.05.2011)): Maximum single buffer size set to:132MB ... Info: CPU affinity mask used: 1; system mask is 7 ... OpenCL-kernels filename : MultiBeam_Kernels_r3500.cl ar=0.010761 NumCfft=119383 NumGauss=0 NumPulse=50212798336 NumTriplet=63189698720 Currently allocated 185 MB for GPU buffers In v_BaseLineSmooth: NumDataPoints=1048576, BoxCarLength=8192, NumPointsInChunk=32768 Restarted at 5.21 percent. Used GPU device parameters are: Number of compute units: 6 Single buffer allocation size: 128MB Total device global memory: 512MB max WG size: 256 local mem type: Real LotOfMem path: no LowPerformanceGPU path: no HighPerformanceGPU path: no period_iterations_num=50 Maximum single buffer size set to:124MB ... OpenCL-kernels filename : MultiBeam_Kernels_r3500.cl ar=0.010761 NumCfft=119383 NumGauss=0 NumPulse=50212798336 NumTriplet=63189698720 Currently allocated 181 MB for GPU buffers In v_BaseLineSmooth: NumDataPoints=1048576, BoxCarLength=8192, NumPointsInChunk=32768 Restarted at 7.95 percent. Used GPU device parameters are: Number of compute units: 6 Single buffer allocation size: 124MB Total device global memory: 512MB max WG size: 256 local mem type: Real LotOfMem path: no LowPerformanceGPU path: no HighPerformanceGPU path: no period_iterations_num=50 Seems this is device || driver || OS limitation, see: "Max memory allocation: 134217728" OpenCL Platform Name: AMD Accelerated Parallel Processing Number of devices: 1 Max compute units: 6 Max work group size: 256 Max clock frequency: 800Mhz Max memory allocation: 134217728 Cache type: None Cache line size: 0 Cache size: 0 Global memory size: 536870912 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: Turks Vendor: Advanced Micro Devices, Inc. Driver version: CAL 1.4.1646 Version: OpenCL 1.1 AMD-APP (831.4) Â - ALF - "Find out what you don't do well ..... then don't do it!" :) Â |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
   OK I will not pretend I understand what all that means apart from basically it is a CPU based OpenCL app (am I right?). . . OK, so the AMD version of r3500 is not SoG but the Nvidia version is? . . Always so much to learn. Stephen . |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
Ah, indeed. App checks that value and will not try to aquire more than runtime could give. So, this GPU can't use more than 128M as single allocation. But it has 512Mb total so can allocate enough number of separate buffers for LoM path. SETI apps news We're not gonna fight them. We're gonna transcend them. |
©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.