Posts by mpyusko

InfoMessage
1) Message boards : GPUs : Projects for HD5450 & HD 5670
Message 44979
Posted 20 Jul 2012 by mpyusko 		OK, so I signed up for Primegrid. I ran a WU on the HD 5450 and it took 15.5 hours. For kicks, I ran 1 WU through my HD 6950.... 30 minutes. WOW! I guess there's a sort of benchmark. Geez.

Thanks.
2) Message boards : GPUs : Projects for HD5450 & HD 5670
Message 44952
Posted 19 Jul 2012 by mpyusko 		I was wondering what projects are available for my HD5450 and HD5670 GPU's.

I've been crunching Seti@home for many years on my numeous CPU's, and I just recently started crunch Milkyway@home on my HD6950 GPU. However I'm having difficulty sifting through what projects are available for these other two. The HD5450 runs on Win7x64 and the HD 5670 runs on XP (32-bit).

Thanks.
3) Message boards : GPUs : GPU still not working. HD6950
Message 44805
Posted 9 Jul 2012 by mpyusko 		The usually's are because there are exceptions and it would be incorrect to speak in absolutes. If you are a Linux enthusiast, then you would notice how MS has copied parts of X widows based GUI's and incorporated them into Windows. There are also GUIs available for Linux, like "Lindows" and "XPde" that attempt to mimic the windows interface as much as possible to minimize the learning curve and entice people to switch. Wine exists for those users who refuse to spend money on a Windows license but want to use software where a Linux port is not available, for example GAMES! In fact, I keep my computer dual-boot between Win7x64pro and Slackware 13.37, just so I can reboot and play games. The advantage is if I need to be productive, I use Linux, where my games aren't available. And when I game, I use Windows where it's a clean OS environment without the bloat and overhead of productivity software backends, auto-updates and such.

In Linux, the computer can be fully functional without the overhead of a GUI. In windows, you must have a GUI, even if you want it to be a Web,MySQL,PHP server. The benefits running Linux should be rather obvious.... no GUI, more resources available for other things.

There are however benefits to using a GUI with Linux. For example, a guru could create a substantial firewall via the command line. However, for someone who doesn't indulge in network security every day, creating a firewall can be confusing and messy, unless you use a GUI frontend. Also, for things like SAMBA (windows network sharing), SWAT, the GUI configuration tool, is great to setup basic shares. However, there are a lot of options and directions you can go which SWAT cannot provide. That is where the console and command line come in. In the case of updates, Debian apt is a great tool. Two basic console commands can perform updates for all your installed software. There is also aptitude that will let you search for new software via the console, but I fund it clunky (IMHO) compared to the GUI option Synaptic.

Linux is a whole different culture and frame of mind vs. Windows. I read it once somewhere, In windows the goal is to make the best, most feature-rich program, all-in-one kind of solution and sell it for lots of money. In Linux, the goal is to focus on a specific task and make the best program you can, make it free, and make it so other programs can easily use it. I find the latter model to be more efficient because a team/project will focus their time and resources on making one small task. Then they will make it very extensible, compatible and resource efficient. Let someone else make a GUI frontend for it. Then they will focus on making the best, most intuitive GUI for it. Let someone else make the Desktop environment. Then they can focus on either being feature-rich, or minimalist. (Compare KDE and Gnome to Windowmaker and Fluxbox). If you really want to taste Linux, I heartily suggest you try a Live CD distro like Knoppix. and then DSL. Both of them can run off a CD or DVD, cache completely into RAM (if you have enough) and do not even require a hard drive in your computer. (This makes them 100% virusproof.) Knoppix is a full-feature Linux environment with a KDE interface. Focused on fitting on a standard CD or DVD. DSL is a minimalist distro which can run on older hardware, with many of the same features as larger distros, but focused on fitting on one 50MB "Business Card" CD.

I could go on and on about Linux vs. Windows. Both have their place. I just prefer Linux.

But I digress... I guess automatically starting BOINC at boot with CPU only and then a run a restart script at X login to enable GPU computing isn't entirely out-of the question, I just wish there was a way to use the GPU at boot without X running. For instance, with X running BOINC is only using 90% of my GPU. Leaving some resources available for the Desktop Environment. (Honestly, using the GPU all the time, doesn't interfere with me at all. No lag, so skips or jerks.... just smooth operation)
4) Message boards : GPUs : GPU still not working. HD6950
Message 44801
Posted 9 Jul 2012 by mpyusko 		Linux and Windows are like apples and oranges. A remote login with Linux is usually done through an SSH connection via a text-based console. When the system boots it starts a SSH server so client computer can connect. X (the GUI) doesn't usually start until a local user has logged in via kdm, gdm, etc... (X can be forwarded via SSH for RDP or other thin-client uses, but it's usually not the norm). System administration is usually done through a SSH console connection because a GUI interface eats up bandwidth and system resources. Also the utilities that are used for system management are usually written for the console first, then sometimes adapted to a GUI usually with reduced features.

It would be nice if there was a way to utilize the GPU without loading X first.
5) Message boards : GPUs : GPU still not working. HD6950
Message 44793
Posted 8 Jul 2012 by mpyusko 		OK, now what did I do? I had to remotely reboot the computer and now it's not detecting the GPU again. It auto starts kdm, though I have to login locally. Does the X server need to be running to detect/use the GPU?
6) Message boards : GPUs : GPU still not working. HD6950
Message 44791
Posted 8 Jul 2012 by mpyusko 		I figured out the milkyway@home part. On their page, under account profile you can select Milkyway@home settings only and unselect use CPU, and select use GPUs. Now I'm crunching like crazy http://stats.free-dc.org/stats.php?page=hostbycpid&cpid=96aa9449453ca87dcbb6218801d75340

I only crunch in Linux, because if I'm running windows it's to play a game.

Thanks again.

clinfo
Number of platforms:                             1
  Platform Profile:                              FULL_PROFILE
  Platform Version:                              OpenCL 1.2 AMD-APP (938.1)
  Platform Name:                                 AMD Accelerated Parallel Processing
  Platform Vendor:                               Advanced Micro Devices, Inc.
  Platform Extensions:                           cl_khr_icd cl_amd_event_callback cl_amd_offline_devices


  Platform Name:                                 AMD Accelerated Parallel Processing
Number of devices:                               2
  Device Type:                                   CL_DEVICE_TYPE_GPU
  Device ID:                                     4098
  Board name:                                    AMD Radeon HD 6900 Series 
  Device Topology:                               PCI[ B#1, D#0, F#0 ]
  Max compute units:                             22
  Max work items dimensions:                     3
    Max work items[0]:                           256
    Max work items[1]:                           256
    Max work items[2]:                           256
  Max work group size:                           256
  Preferred vector width char:                   16
  Preferred vector width short:                  8
  Preferred vector width int:                    4
  Preferred vector width long:                   2
  Preferred vector width float:                  4
  Preferred vector width double:                 2
  Native vector width char:                      16
  Native vector width short:                     8
  Native vector width int:                       4
  Native vector width long:                      2
  Native vector width float:                     4
  Native vector width double:                    2
  Max clock frequency:                           0Mhz
  Address bits:                                  32
  Max memory allocation:                         268435456
  Image support:                                 Yes
  Max number of images read arguments:           128
  Max number of images write arguments:          8
  Max image 2D width:                            8192
  Max image 2D height:                           8192
  Max image 3D width:                            2048
  Max image 3D height:                           2048
  Max image 3D depth:                            2048
  Max samplers within kernel:                    16
  Max size of kernel argument:                   1024
  Alignment (bits) of base address:              2048
  Minimum alignment (bytes) for any datatype:    128
  Single precision floating point capability
    Denorms:                                     No
    Quiet NaNs:                                  Yes
    Round to nearest even:                       Yes
    Round to zero:                               Yes
    Round to +ve and infinity:                   Yes
    IEEE754-2008 fused multiply-add:             Yes
  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
  Kernel Preferred work group size multiple:     64
  Error correction support:                      0
  Unified memory for Host and Device:            0
  Profiling timer resolution:                    1
  Device endianess:                              Little
  Available:                                     Yes
  Compiler available:                            Yes
  Execution capabilities:                                
    Execute OpenCL kernels:                      Yes
    Execute native function:                     No
  Queue properties:                              
    Out-of-Order:                                No
    Profiling :                                  Yes
  Platform ID:                                   0x7f1f1809f0e0
  Name:                                          Cayman
  Vendor:                                        Advanced Micro Devices, Inc.
  Device OpenCL C version:                       OpenCL C 1.2 
  Driver version:                                CAL 1.4.1741
  Profile:                                       FULL_PROFILE
  Version:                                       OpenCL 1.2 AMD-APP (938.1)
  Extensions:                                    cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_gl_sharing cl_ext_atomic_counters_32 cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_popcnt 


  Device Type:                                   CL_DEVICE_TYPE_CPU
  Device ID:                                     4098
  Board name:                                    
  Max compute units:                             2
  Max work items dimensions:                     3
    Max work items[0]:                           1024
    Max work items[1]:                           1024
    Max work items[2]:                           1024
  Max work group size:                           1024
  Preferred vector width char:                   16
  Preferred vector width short:                  8
  Preferred vector width int:                    4
  Preferred vector width long:                   2
  Preferred vector width float:                  4
  Preferred vector width double:                 0
  Native vector width char:                      16
  Native vector width short:                     8
  Native vector width int:                       4
  Native vector width long:                      2
  Native vector width float:                     4
  Native vector width double:                    0
  Max clock frequency:                           3203Mhz
  Address bits:                                  64
  Max memory allocation:                         2147483648
  Image support:                                 Yes
  Max number of images read arguments:           128
  Max number of images write arguments:          8
  Max image 2D width:                            8192
  Max image 2D height:                           8192
  Max image 3D width:                            2048
  Max image 3D height:                           2048
  Max image 3D depth:                            2048
  Max samplers within kernel:                    16
  Max size of kernel argument:                   4096
  Alignment (bits) of base address:              1024
  Minimum alignment (bytes) for any datatype:    128
  Single precision floating point capability
    Denorms:                                     Yes
    Quiet NaNs:                                  Yes
    Round to nearest even:                       Yes
    Round to zero:                               Yes
    Round to +ve and infinity:                   Yes
    IEEE754-2008 fused multiply-add:             Yes
  Cache type:                                    Read/Write
  Cache line size:                               64
  Cache size:                                    32768
  Global memory size:                            8371740672
  Constant buffer size:                          65536
  Max number of constant args:                   8
  Local memory type:                             Global
  Local memory size:                             32768
  Kernel Preferred work group size multiple:     1
  Error correction support:                      0
  Unified memory for Host and Device:            1
  Profiling timer resolution:                    1
  Device endianess:                              Little
  Available:                                     Yes
  Compiler available:                            Yes
  Execution capabilities:                                
    Execute OpenCL kernels:                      Yes
    Execute native function:                     Yes
  Queue properties:                              
    Out-of-Order:                                No
    Profiling :                                  Yes
  Platform ID:                                   0x7f1f1809f0e0
  Name:                                          Pentium(R) Dual-Core  CPU      E6700  @ 3.20GHz
  Vendor:                                        GenuineIntel
  Device OpenCL C version:                       OpenCL C 1.2 
  Driver version:                                2.0 (sse2)
  Profile:                                       FULL_PROFILE
  Version:                                       OpenCL 1.2 AMD-APP (938.1)
  Extensions:                                    cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_byte_addressable_store cl_khr_gl_sharing cl_ext_device_fission cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_popcnt
7) Message boards : GPUs : GPU still not working. HD6950
Message 44788
Posted 7 Jul 2012 by mpyusko 		The thing is, I installed 7.0.25 in /usr/local/bin with boinc as a non-privileged user with /home/boinc as it's data directory. so when I run boinc from the command line, it will either try to use the current folder as the data directory, or I have to specify the data directory 
boinc --dir /home/boinc


I think that's the problem!

mpyusko@wopr:/opt/BOINC$ ./boinc
07-Jul-2012 12:26:34 [---] Starting BOINC client version 7.0.30 for x86_64-pc-linux-gnu
07-Jul-2012 12:26:34 [---] log flags: file_xfer, sched_ops, task, coproc_debug, cpu_sched_debug, work_fetch_debug
07-Jul-2012 12:26:34 [---] Libraries: libcurl/7.21.4 OpenSSL/0.9.8n zlib/1.2.5 libidn/1.19
07-Jul-2012 12:26:34 [---] Data directory: /opt/BOINC
07-Jul-2012 12:26:34 [---] Processor: 2 GenuineIntel Pentium(R) Dual-Core  CPU      E6700  @ 3.20GHz [Family 6 Model 23 Stepping 10]
07-Jul-2012 12:26:34 [---] Processor: 2.00 MB cache
07-Jul-2012 12:26:34 [---] Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm xsave lahf_lm dts tpr_shadow vnmi flexpriority
07-Jul-2012 12:26:34 [---] OS: Linux: 2.6.37.6
07-Jul-2012 12:26:34 [---] Memory: 7.80 GB physical, 0 bytes virtual
07-Jul-2012 12:26:34 [---] Disk: 37.25 GB total, 32.90 GB free
07-Jul-2012 12:26:34 [---] Local time is UTC -4 hours
07-Jul-2012 12:26:34 [---] ATI GPU 0: AMD Radeon HD 6900 series (Cayman) (CAL version 1.4.1741, 2048MB, 1835MB available, 50 GFLOPS peak)
07-Jul-2012 12:26:34 [---] OpenCL: ATI GPU 0: AMD Radeon HD 6900 series (Cayman) (driver version CAL 1.4.1741, device version OpenCL 1.2 AMD-APP (938.1), 2048MB, 1835MB available)
07-Jul-2012 12:26:34 [---] No NVIDIA library found
07-Jul-2012 12:26:34 [---] Config: use all coprocessors
07-Jul-2012 12:26:34 [---] Config: ignoring NVIDIA GPU 1
07-Jul-2012 12:26:34 [---] Config: GUI RPC allowed from:
07-Jul-2012 12:26:34 [---] Config:   192.168.1.3
07-Jul-2012 12:26:34 [---] Config:   192.168.1.30
07-Jul-2012 12:26:34 [---] Config:   192.168.1.31
07-Jul-2012 12:26:34 [---] Config:   192.168.1.22
07-Jul-2012 12:26:34 [---] Version change (7.0.25 -> 7.0.30)
07-Jul-2012 12:26:34 [Milkyway@Home] URL http://milkyway.cs.rpi.edu/milkyway/; Computer ID 453236; resource share 100
07-Jul-2012 12:26:34 [SETI@home] URL http://setiathome.berkeley.edu/; Computer ID 5974796; resource share 100
07-Jul-2012 12:26:34 [Milkyway@Home] General prefs: from Milkyway@Home (last modified 27-May-2011 11:50:22)
07-Jul-2012 12:26:34 [Milkyway@Home] Host location: none
07-Jul-2012 12:26:34 [Milkyway@Home] General prefs: using your defaults
07-Jul-2012 12:26:34 [---] Reading preferences override file
07-Jul-2012 12:26:34 [---] Preferences:
07-Jul-2012 12:26:34 [---]    max memory usage when active: 3991.96MB
07-Jul-2012 12:26:34 [---]    max memory usage when idle: 7185.52MB
07-Jul-2012 12:26:34 [---]    max disk usage: 10.00GB
07-Jul-2012 12:26:34 [---]    suspend work if non-BOINC CPU load exceeds 75 %
07-Jul-2012 12:26:34 [---]    (to change preferences, visit the web site of an attached project, or select Preferences in the Manager)
07-Jul-2012 12:26:34 [---] [cpu_sched_debug] Request CPU reschedule: Prefs update
07-Jul-2012 12:26:34 [---] [work_fetch] Request work fetch: Prefs update
07-Jul-2012 12:26:34 [---] [cpu_sched_debug] Request CPU reschedule: Startup
07-Jul-2012 12:26:34 [---] [work_fetch] Request work fetch: Startup
07-Jul-2012 12:26:34 [---] Not using a proxy
Initialization completed
07-Jul-2012 12:26:35 [---] Running CPU benchmarks
07-Jul-2012 12:26:35 [---] [cpu_sched_debug] Request CPU reschedule: Idle state change
07-Jul-2012 12:26:35 [---] Suspending computation - CPU benchmarks in progress
07-Jul-2012 12:27:06 [---] Benchmark results:
07-Jul-2012 12:27:06 [---]    Number of CPUs: 2
07-Jul-2012 12:27:06 [---]    3491 floating point MIPS (Whetstone) per CPU
07-Jul-2012 12:27:06 [---]    17934 integer MIPS (Dhrystone) per CPU


I logged out from root and switched to my standard user account. It detected the GPU! THANK YOU!!! Now I have two more questions.

1, How do I get it load right from rc.local (or at boot). I believe rc.local executes as root.

2, Is there a way to configure it so Seti@home uses the CPU's, and Milkyway@home only uses the GPU?
8) Message boards : GPUs : GPU still not working. HD6950
Message 44782
Posted 7 Jul 2012 by mpyusko 		Yes, well I knew what he meant. Regardless, it still says:

07-Jul-2012 00:02:39 [---] No NVIDIA library found
07-Jul-2012 00:02:39 [---] No usable CAL devices found
07-Jul-2012 00:02:39 [---] clGetDeviceIDs for platform #0 returned error -1
07-Jul-2012 00:02:39 [---] No usable GPUs found
07-Jul-2012 00:02:39 [---] Config: use all coprocessors
07-Jul-2012 00:02:39 [---] Config: ignoring NVIDIA GPU 1


Is there some setup script I'm missing. I mean one day I had an HD 5670, then the next I swapped it out and installed a HD 6950. I down loaded BOINC 7.0.30 source code, compiled it, installed it, ran it. It's never detected my GPU (even the old one). Then I upgraded from 12.4 to 12.6 Beta drivers. No Luck. I'm now on 12.6 Final. I done both AMD-APP-SDK-v2.7-RC-lnx64.tgz & AMD-APP-SDK-v2.7-lnx64.tgz with no luck.

I did run across this in the sourcecode (http://boinc.berkeley.edu/svn/trunk/boinc/client/gpu_amd.cpp)...
if (!numDevices) {
        warnings.push_back("No usable CAL devices found");
        return;

Which means it's not detecting my device. I have /dev/ati/card0.


cat /proc/ati/0/biosversion

BIOS_CREATION_DATE="03/22/12 01:30"
BIOS_MSG="CAYMAN PRO Hill GDDR5 BIOS                                                 "
BIOS_KIT_VERSION="BK-ATI VER013.012.000.008.000000"
BIOS_PN="113-1E186J20-O4B


But clinfo still only detects my cpu, not my GPU. I've run installation scripts, but not any configuration scripts, other than 
 aticonfig --uninstall=force
then I installed the 12.6 driver and ran
 aticonfig --initial


Not sure what else to do.
9) Message boards : GPUs : GPU still not working. HD6950
Message 44775
Posted 6 Jul 2012 by mpyusko
Device Type: CL_DEVICE_TYPE_CPU


It is only listing the Specs for the CPU.

Max compute units: 2

It's a Dual-Core CPU (2 threads)
Name: Pentium(R) Dual-Core CPU E6700 @ 3.20GHz

See?

Anyway, there is not a single spec on that list that is GPU related. My GPU has like 1408 cores @ 820MHz with 2GB of DDR5.

I just went and did a forced uninstall of 12.6 Beta and a clean install of 12.6. Still No Usable GPUs Found.

Fri 06 Jul 2012 05:44:20 PM EDT |  | No usable CAL devices found


Thanks though.
10) Message boards : GPUs : GPU still not working. HD6950
Message 44773
Posted 6 Jul 2012 by mpyusko 		Shouldn't Number of Devices be "2"?

root@wopr:/# clinfo
No protocol specified
Number of platforms:                             1
  Platform Profile:                              FULL_PROFILE
  Platform Version:                              OpenCL 1.2 AMD-APP (938.1)
  Platform Name:                                 AMD Accelerated Parallel Processing
  Platform Vendor:                               Advanced Micro Devices, Inc.
  Platform Extensions:                           cl_khr_icd cl_amd_event_callback cl_amd_offline_devices


  Platform Name:                                 AMD Accelerated Parallel Processing
Number of devices:                               1
  Device Type:                                   CL_DEVICE_TYPE_CPU
  Device ID:                                     4098
  Board name:                                    
  Max compute units:                             2
  Max work items dimensions:                     3
    Max work items[0]:                           1024
    Max work items[1]:                           1024
    Max work items[2]:                           1024
  Max work group size:                           1024
  Preferred vector width char:                   16
  Preferred vector width short:                  8
  Preferred vector width int:                    4
  Preferred vector width long:                   2
  Preferred vector width float:                  4
  Preferred vector width double:                 0
  Native vector width char:                      16
  Native vector width short:                     8
  Native vector width int:                       4
  Native vector width long:                      2
  Native vector width float:                     4
  Native vector width double:                    0
  Max clock frequency:                           3203Mhz
  Address bits:                                  64
  Max memory allocation:                         2147483648
  Image support:                                 Yes
  Max number of images read arguments:           128
  Max number of images write arguments:          8
  Max image 2D width:                            8192
  Max image 2D height:                           8192
  Max image 3D width:                            2048
  Max image 3D height:                           2048
  Max image 3D depth:                            2048
  Max samplers within kernel:                    16
  Max size of kernel argument:                   4096
  Alignment (bits) of base address:              1024
  Minimum alignment (bytes) for any datatype:    128
  Single precision floating point capability
    Denorms:                                     Yes
    Quiet NaNs:                                  Yes
    Round to nearest even:                       Yes
    Round to zero:                               Yes
    Round to +ve and infinity:                   Yes
    IEEE754-2008 fused multiply-add:             Yes
  Cache type:                                    Read/Write
  Cache line size:                               64
  Cache size:                                    32768
  Global memory size:                            8371740672
  Constant buffer size:                          65536
  Max number of constant args:                   8
  Local memory type:                             Global
  Local memory size:                             32768
  Kernel Preferred work group size multiple:     1
  Error correction support:                      0
  Unified memory for Host and Device:            1
  Profiling timer resolution:                    1
  Device endianess:                              Little
  Available:                                     Yes
  Compiler available:                            Yes
  Execution capabilities:                                
    Execute OpenCL kernels:                      Yes
    Execute native function:                     Yes
  Queue properties:                              
    Out-of-Order:                                No
    Profiling :                                  Yes
  Platform ID:                                   0x7fe8b83680e0
  Name:                                          Pentium(R) Dual-Core  CPU      E6700  @ 3.20GHz
  Vendor:                                        GenuineIntel
  Device OpenCL C version:                       OpenCL C 1.2 
  Driver version:                                2.0 (sse2)
  Profile:                                       FULL_PROFILE
  Version:                                       OpenCL 1.2 AMD-APP (938.1)
  Extensions:                                    cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_byte_addressable_store cl_khr_gl_sharing cl_ext_device_fission cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_popcnt
11) Message boards : GPUs : GPU still not working. HD6950
Message 44771
Posted 6 Jul 2012 by mpyusko 		I just found this....

http://boinc.berkeley.edu/dev/forum_thread.php?id=7619&nowrap=true#44402

But I'm actually using 12.6 (because 12.4 wasn't working either). Perhaps there is the same issue with 12.6 (????)

However, if I read the thread correctly, the problem was resolved by updating to 7.0.28. I compiled 7.0.30 from source and installed it. Hmmmmmmm. An HD 6950 has not only been labeled as compatible, but turns out really high numbers even when compared to professional GPU's. I've got to be missing a detail somewhere.

Edit: (It was also a windows machine)
12) Message boards : GPUs : GPU still not working. HD6950
Message 44770
Posted 6 Jul 2012 by mpyusko 		I started out by posting on Milkyway@home, but didn't get any help there. You've at least gotten me closer. ;)
13) Message boards : GPUs : GPU still not working. HD6950
Message 44768
Posted 6 Jul 2012 by mpyusko 		both should be lib64 in my case. However, it is still not detecting any usable GPU's.
root@wopr:/opt/BOINC# ./boinc
06-Jul-2012 13:30:23 [---] Starting BOINC client version 7.0.30 for x86_64-pc-linux-gnu
06-Jul-2012 13:30:23 [---] log flags: file_xfer, sched_ops, task, coproc_debug, cpu_sched_debug, work_fetch_debug
06-Jul-2012 13:30:23 [---] Libraries: libcurl/7.21.4 OpenSSL/0.9.8n zlib/1.2.5 libidn/1.19
06-Jul-2012 13:30:23 [---] Data directory: /opt/BOINC
06-Jul-2012 13:30:23 [---] Processor: 2 GenuineIntel Pentium(R) Dual-Core  CPU      E6700  @ 3.20GHz [Family 6 Model 23 Stepping 10]
06-Jul-2012 13:30:23 [---] Processor: 2.00 MB cache
06-Jul-2012 13:30:23 [---] Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm xsave lahf_lm dts tpr_shadow vnmi flexpriority
06-Jul-2012 13:30:23 [---] OS: Linux: 2.6.37.6
06-Jul-2012 13:30:23 [---] Memory: 7.80 GB physical, 0 bytes virtual
06-Jul-2012 13:30:23 [---] Disk: 37.25 GB total, 32.90 GB free
06-Jul-2012 13:30:23 [---] Local time is UTC -4 hours
No protocol specified
06-Jul-2012 13:30:23 [---] No NVIDIA library found
06-Jul-2012 13:30:23 [---] No usable CAL devices found
06-Jul-2012 13:30:23 [---] clGetDeviceIDs for platform #0 returned error -1
06-Jul-2012 13:30:23 [---] No usable GPUs found
06-Jul-2012 13:30:23 [---] Config: use all coprocessors
06-Jul-2012 13:30:23 [---] Config: GUI RPC allowed from:
06-Jul-2012 13:30:23 [---] Config:   192.168.1.3
06-Jul-2012 13:30:23 [---] Config:   192.168.1.30
06-Jul-2012 13:30:23 [---] Config:   192.168.1.31
06-Jul-2012 13:30:23 [---] Config:   192.168.1.22
06-Jul-2012 13:30:23 [Milkyway@Home] URL http://milkyway.cs.rpi.edu/milkyway/; Computer ID 453236; resource share 100
06-Jul-2012 13:30:23 [SETI@home] URL http://setiathome.berkeley.edu/; Computer ID 5974796; resource share 100
06-Jul-2012 13:30:23 [Milkyway@Home] General prefs: from Milkyway@Home (last modified 27-May-2011 11:50:22)
06-Jul-2012 13:30:23 [Milkyway@Home] Host location: none
06-Jul-2012 13:30:23 [Milkyway@Home] General prefs: using your defaults
06-Jul-2012 13:30:23 [---] Reading preferences override file
06-Jul-2012 13:30:23 [---] Preferences:
06-Jul-2012 13:30:23 [---]    max memory usage when active: 3991.96MB
06-Jul-2012 13:30:23 [---]    max memory usage when idle: 7185.52MB
06-Jul-2012 13:30:23 [---]    max disk usage: 10.00GB
06-Jul-2012 13:30:23 [---]    suspend work if non-BOINC CPU load exceeds 75 %
06-Jul-2012 13:30:23 [---]    (to change preferences, visit the web site of an attached project, or select Preferences in the Manager)
06-Jul-2012 13:30:23 [---] [cpu_sched_debug] Request CPU reschedule: Prefs update
06-Jul-2012 13:30:23 [---] [work_fetch] Request work fetch: Prefs update
06-Jul-2012 13:30:23 [---] [cpu_sched_debug] Request CPU reschedule: Startup
06-Jul-2012 13:30:23 [---] [work_fetch] Request work fetch: Startup
06-Jul-2012 13:30:23 [---] Not using a proxy
Initialization completed

My current cc_config.xml file:
<cc_config>
<options>
<use_all_gpus>1</use_all_gpus>
<no_gpus>0</no_gpus>
</options>
<log_flags>
<cpu_sched_debug>1</cpu_sched_debug>
<work_fetch_debug>1</work_fetch_debug>
<coproc_debug>1</coproc_debug> 
</log_flags> 
</cc_config>


???
14) Message boards : GPUs : GPU still not working. HD6950
Message 44766
Posted 6 Jul 2012 by mpyusko 		I have both those files. :/

root@wopr:/opt/BOINC# whereis libaticalrt.so
libaticalrt: /usr/lib/libaticalrt.so /usr/lib64/libaticalrt.so                                                       
root@wopr:/opt/BOINC# whereis libOpenCL.so                                                                           
libOpenCL: /usr/lib/libOpenCL.so /usr/lib64/libOpenCL.so
15) Message boards : GPUs : GPU still not working. HD6950
Message 44676
Posted 30 Jun 2012 by mpyusko 		I compiled it from source then placed it in /opt/BOINC It runs from there. It starts automatically using /etc/rc.d/rc.local

# Begin BOINC Client Startup

cd /opt/BOINC
./run_client --daemon
cd /
echo "BOINC running"
# End BOINC Client Startup


What is it looking for so I can point to it properly?
16) Message boards : GPUs : GPU still not working. HD6950
Message 44671
Posted 30 Jun 2012 by mpyusko 		I've been using BOINC for years, but I just recently purchased a decent GPU, An AMD Radeon HD 6950. (Sapphire Radeon HD 6950 2GB DDR5 OC) I upgraded from a Gigabyte HD 5670 1GB DDR3. BOINC has never, and still doesn't detect my GPU.

bash-4.1$ fglrxinfo
display: :0  screen: 0
OpenGL vendor string: Advanced Micro Devices, Inc.
OpenGL renderer string: AMD Radeon HD 6900 Series 
OpenGL version string: 4.2.11733 Compatibility Profile Context

bash-4.1$


I'm running Slackware Linux 13.37 64-bit. I just compiled 7.0.30 from source and it is running fine. I tried using a cc_config.xml file, but I still can't seem to get it to work. My goal is to run Seti@Home on my two CPU cores and Milkyway@Home exclusively on my GPU (When idle). I goggled around and I see there are other people out there using HD 6950's with BOINC and turning out great results. But I still can't seem to figure out why the GPU isn't being detected. I'm using Catalyst 12.6 (tried 12.4 too) and I have the OpenCL SDK installed.

Please Help.


Copyright © 2025 University of California.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.