ppsieve OpenCL *Mac* testing

Join PrimeGrid

Read our rules and policies.
Download, install and run the BOINC software used by PrimeGrid.
When prompted, enter the URL:
http://www.primegrid.com/
Select which subprojects to run via the PrimeGrid Preferences.

If you are not using graphics enabled BOINC client, you may use manual registration form.

Author

Message

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Hi all,

I've built the OpenCL version of ppsieve and tpsieve for Mac. If you are able to test the apps please get them from http://www.pyramid-productions.net/downloads/ppsieve-cl.zip

Unlike the Windows/Linux apps, which are build on the ATI/Stream SDK, the Mac apps should be able to target both NVidia and ATI GPUs. I have tested the app on my Nvidia 9600M GT using the following range:

./ppsieve-cl-x86_64-mac -p20070e9 -P20070020e6 -k 1201 -K 9999 -N 2000000 -c 60
ppsieve version cl-0.2.3b (testing)
Compiled Nov 25 2010 with GCC 4.2.1 (Apple Inc. build 5664)
nstart=74, nstep=31
ppsieve initialized: 1201 <= k <= 9999, 74 <= n < 2000000
Sieve started: 20070000000000 <= p < 20070020000000
Thread 0 starting
Detected 64 multiprocessors (320 SPUs) on device 0.
nstep changed to 22
CL setup complete.
cthread_count = 4096
20070000475957 | 4995*2^1822738+1
20070001146497 | 4977*2^626298+1
20070001163929 | 3765*2^461308+1
20070001302811 | 7669*2^725426+1
20070001425977 | 5821*2^1775248+1
20070002245151 | 1221*2^646983+1
20070002606341 | 4809*2^497683+1
20070004816819 | 6699*2^1215561+1
20070005914001 | 9847*2^1634140+1
20070006187837 | 9923*2^287853+1
20070006875981 | 1645*2^965954+1
20070007170259 | 3889*2^49730+1
20070008329039 | 9065*2^832569+1
20070012507137 | 7061*2^1730371+1
20070014814419 | 6923*2^1257277+1
20070014902183 | 6981*2^1674047+1
20070014977063 | 4809*2^470009+1
20070015687191 | 5973*2^502229+1
20070016440869 | 9577*2^34776+1
20070016765411 | 1421*2^1766083+1
20070016992617 | 7001*2^1177373+1
20070017505563 | 7627*2^401848+1
20070017763227 | 6625*2^1332032+1
20070018281791 | 4685*2^1510525+1
20070019870763 | 8189*2^2803+1

Thread 0 completed
Waiting for threads to exit
Sieve complete: 20070000000000 <= p < 20070020000000
Found 25 factors
count=652993,sum=0xb5e05bf97b0cfbd5
Elapsed time: 203.03 sec. (0.02 init + 203.01 sieve) at 99431 p/sec.
Processor time: 51.55 sec. (0.01 init + 51.53 sieve) at 391683 p/sec.
Average processor utilization: 0.69 (init), 0.25 (sieve)

Please post your results here - I'd be especially interested to see any results from ATI - firstly to check the app works, and secondly, so we can compare the speed of the app with the windows/linux version.

System requirements are Mac OS 10.6.x (for OpenCL support), and an Nvidia or ATI card with OpenCL-supporting driver.

Have fun!

- Iain

ID: 28347 |

thetedinator
Send message
Joined: 29 Oct 10
Posts: 6
ID: 71569
Credit: 842,201
RAC: 0

./ppsieve-cl-x86_64-mac -p20070e9 -P20070020e6 -k 1201 -K 9999 -N 2000000 -c 60
ppsieve version cl-0.2.3b (testing)
Compiled Nov 25 2010 with GCC 4.2.1 (Apple Inc. build 5664)
nstart=74, nstep=31
ppsieve initialized: 1201 <= k <= 9999, 74 <= n < 2000000
Sieve started: 20070000000000 <= p < 20070020000000
Thread 0 starting
Detected 480 multiprocessors (2400 SPUs) on device 0.
nstep changed to 22
CL setup complete.
cthread_count = 30720
20070000475957 | 4995*2^1822738+1
20070001146497 | 4977*2^626298+1
20070001163929 | 3765*2^461308+1
20070001302811 | 7669*2^725426+1
20070001425977 | 5821*2^1775248+1
20070002245151 | 1221*2^646983+1
20070002606341 | 4809*2^497683+1
20070004816819 | 6699*2^1215561+1
20070005914001 | 9847*2^1634140+1
20070006187837 | 9923*2^287853+1
20070006875981 | 1645*2^965954+1
20070007170259 | 3889*2^49730+1
20070008329039 | 9065*2^832569+1
20070012507137 | 7061*2^1730371+1
20070014814419 | 6923*2^1257277+1
20070014902183 | 6981*2^1674047+1
20070014977063 | 4809*2^470009+1
20070015687191 | 5973*2^502229+1
20070016440869 | 9577*2^34776+1
20070016765411 | 1421*2^1766083+1
20070016992617 | 7001*2^1177373+1
20070017505563 | 7627*2^401848+1
20070017763227 | 6625*2^1332032+1
20070018281791 | 4685*2^1510525+1
20070019870763 | 8189*2^2803+1

Thread 0 completed
Waiting for threads to exit
Sieve complete: 20070000000000 <= p < 20070020000000
Found 25 factors
count=652993,sum=0xb5e05bf97b0cfbd5
Elapsed time: 18.87 sec. (0.01 init + 18.86 sieve) at 1069980 p/sec.
Processor time: 16.84 sec. (0.01 init + 16.83 sieve) at 1199635 p/sec.
Average processor utilization: 1.55 (init), 0.89 (sieve)

____________

ID: 28348 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Can you post what type of card that was run with?

Cheers

- Iain

ID: 28357 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

John suggested it would be better to test with tpsieve, since that's what we'll be using on BOINC, so for the same range of p (note the different max N):

./tpsieve-cl-x86_64-mac -p20070e9 -P20070020e6 -k5 -K9999 -N 3000000 -T -M2 -c 60
tpsieve version cl-0.2.3b (testing)
Compiled Nov 25 2010 with GCC 4.2.1 (Apple Inc. build 5664)
nstart=74, nstep=31
tpsieve initialized: 5 <= k <= 9999, 74 <= n < 3000000
Sieve started: 20070000000000 <= p < 20070020000000
Thread 0 starting
Detected 64 multiprocessors (320 SPUs) on device 0.
nstep changed to 22
CL setup complete.
cthread_count = 4096
20070000059159 | 2205*2^2472339-1
20070000105761 | 733*2^974106+1
...<snip>...
20070019838257 | 8025*2^2120157-1
20070019870763 | 8189*2^2803+1

Thread 0 completed
Waiting for threads to exit
Sieve complete: 20070000000000 <= p < 20070020000000
Found 109 factors
count=652993,sum=0xb5e05bf97b0cfbd5
Elapsed time: 547.60 sec. (0.01 init + 547.59 sieve) at 36862 p/sec.
Processor time: 98.75 sec. (0.02 init + 98.73 sieve) at 204447 p/sec.
Average processor utilization: 1.42 (init), 0.18 (sieve)

Core 2 Due 2.66 GHz, Nvidia GeForce 9600M GT.

Remember, please post your system specs...

- Iain

ID: 28363 |

spamguy

Send message
Joined: 21 Jan 06
Posts: 51
ID: 2225
Credit: 29,719,255
RAC: 0

Model Name: MacBook Pro
Model Identifier: MacBookPro5,1
Processor Name: Intel Core 2 Duo
Processor Speed: 2.4 GHz
Number Of Processors: 1
Total Number Of Cores: 2
L2 Cache: 3 MB
Memory: 2 GB
Bus Speed: 1.07 GHz
Graphics Card: NVIDIA GeForce 9600M GT

---

spamguy:ppsieve-cl spamguy$ ./ppsieve-cl-x86-mac -p20070e9 -P20070020e6 -k 1201 -K 9999 -N 2000000 -c 60
ppsieve version cl-0.2.3b (testing)
Compiled Nov 25 2010 with GCC 4.2.1 (Apple Inc. build 5664)
nstart=74, nstep=31
ppsieve initialized: 1201 <= k <= 9999, 74 <= n < 2000000
Sieve started: 20070000000000 <= p < 20070020000000
Thread 0 starting
Detected 64 multiprocessors (320 SPUs) on device 0.
nstep changed to 22
CL setup complete.
cthread_count = 4096
20070000475957 | 4995*2^1822738+1
20070001146497 | 4977*2^626298+1
20070001163929 | 3765*2^461308+1
20070001302811 | 7669*2^725426+1
20070001425977 | 5821*2^1775248+1
20070002245151 | 1221*2^646983+1
20070002606341 | 4809*2^497683+1
20070004816819 | 6699*2^1215561+1
p=20070006029313, 100.5K p/sec, 0.29 CPU cores, 30.1% done. ETA 28 Nov 11:33
20070005914001 | 9847*2^1634140+1
20070006187837 | 9923*2^287853+1
20070006875981 | 1645*2^965954+1
20070007170259 | 3889*2^49730+1
20070008329039 | 9065*2^832569+1
p=20070012058625, 100.5K p/sec, 0.30 CPU cores, 60.3% done. ETA 28 Nov 11:33
20070012507137 | 7061*2^1730371+1
20070014814419 | 6923*2^1257277+1
20070014902183 | 6981*2^1674047+1
20070014977063 | 4809*2^470009+1
20070015687191 | 5973*2^502229+1
20070016440869 | 9577*2^34776+1
20070016765411 | 1421*2^1766083+1
20070016992617 | 7001*2^1177373+1
p=20070017825793, 96.12K p/sec, 0.30 CPU cores, 89.1% done. ETA 28 Nov 11:33
20070017505563 | 7627*2^401848+1
20070017763227 | 6625*2^1332032+1
20070018281791 | 4685*2^1510525+1
20070019870763 | 8189*2^2803+1

Thread 0 completed
Waiting for threads to exit
Sieve complete: 20070000000000 <= p < 20070020000000
Found 25 factors
count=652993,sum=0xb5e05bf97b0cfbd5
Elapsed time: 204.80 sec. (0.07 init + 204.74 sieve) at 98591 p/sec.
Processor time: 60.35 sec. (0.02 init + 60.33 sieve) at 334566 p/sec.
Average processor utilization: 0.29 (init), 0.29 (sieve)

ID: 28493 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Thanks. Anybody got an ATI card in their Mac?

ID: 28499 |

spamguy

Send message
Joined: 21 Jan 06
Posts: 51
ID: 2225
Credit: 29,719,255
RAC: 0

I imagine ATIs will be hard to come by since Apple rules computer configurations with an iron fist. Hopefully we have hardcore Mac Pro owners lurking!

ID: 28505 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Yes, a quick scroll through the list of hosts on PG and I couldn't find any. There is hope though, since the new iMacs also have ATI cards...

Cheers

- Iain

ID: 28510 |

Phroon
Send message
Joined: 27 Sep 10
Posts: 10
ID: 68760
Credit: 70,711,324
RAC: 0

Not looking good on my ATI.

Model Name: iMac
Model Identifier: iMac11,1
Processor Name: Intel Core i7
Processor Speed: 2.8 GHz
Number Of Processors: 1
Total Number Of Cores: 4
L2 Cache (per core): 256 KB
L3 Cache: 8 MB
Memory: 8 GB
System Version: Mac OS X 10.6.5 (10H574)
Chipset Model: ATI Radeon HD 4850

./ppsieve-cl-x86_64-mac -p20070e9 -P20070020e6 -k 1201 -K 9999 -N 2000000 -c 60
ppsieve version cl-0.2.3b (testing)
Compiled Nov 25 2010 with GCC 4.2.1 (Apple Inc. build 5664)
nstart=74, nstep=31
ppsieve initialized: 1201 <= k <= 9999, 74 <= n < 2000000
Sieve started: 20070000000000 <= p < 20070020000000
Thread 0 starting
Detected 160 multiprocessors (800 SPUs) on device 0.
nstep changed to 22
CL setup complete.
cthread_count = 10240

Thread 0 completed
Waiting for threads to exit
Sieve complete: 20070000000000 <= p < 20070020000000
Found 0 factors
count=652993,sum=0xb5e05bf97b0cfbd5
Elapsed time: 34.74 sec. (0.01 init + 34.74 sieve) at 581080 p/sec.
Processor time: 3.28 sec. (0.01 init + 3.27 sieve) at 6168111 p/sec.
Average processor utilization: 1.42 (init), 0.09 (sieve)

ID: 28816 |

hobold
Send message
Joined: 22 Nov 10
Posts: 1
ID: 74047
Credit: 757,891
RAC: 0

No luck on a new Mac Pro 2010 (Model 5,1) with a Radeon 5870. The machine became unresponsive when the test program was started, although the mouse cursor could still be moved. Final output were the lines "CL setup complete" and "cthread_count = 20480". Nothing but a power cycle would bring it back to life (not very helpful for debugging, I realize). There were even a few graphical glitches after the power cycle, but another soft restart cured those.

Time for a bug report to Apple?

ID: 29039 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Thanks for the reports - I'll look into it!

ID: 29043 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Hi,

Could one of you guys with the ATI cards please try this new build - I still hae no idea what the problem is but this has more debugging turned on - http://www.pyramid-productions.net/downloads/tpsieve-cl.zip

Please run the range:

./tpsieve-cl-x86_64-mac -p20070e9 -P20070020e6 -k5 -K9999 -N 3000000 -T -M2 -c 60

and let me know how you get on.

Cheers

- Iain

ID: 29890 |

Phroon
Send message
Joined: 27 Sep 10
Posts: 10
ID: 68760
Credit: 70,711,324
RAC: 0

Hope this helps:

Model Name: iMac
Model Identifier: iMac11,1
Processor Name: Intel Core i7
Processor Speed: 2.8 GHz
Number Of Processors: 1
Total Number Of Cores: 4
L2 Cache (per core): 256 KB
L3 Cache: 8 MB
Memory: 8 GB
System Version: Mac OS X 10.6.5 (10H574)
Chipset Model: ATI Radeon HD 4850

$ ./tpsieve-cl-x86_64-mac -p20070e9 -P20070020e6 -k5 -K9999 -N 3000000 -T -M2 -c 60
tpsieve version cl-0.2.3b (testing)
Compiled Dec 23 2010 with GCC 4.2.1 (Apple Inc. build 5664)
create_sieve_data: Using 318 odd primes, max=2113, lim=2116
create_sieve_data: Using 318 odd primes, max=2113, lim=2116
init_sieve_primes: Generated 314623 odd primes, max=4479941, lim=4479958
nstart=74, nstep=31
tpsieve initialized: 5 <= k <= 9999, 74 <= n < 3000000
Sieve started: 20070000000000 <= p < 20070020000000
create_sieve: Using 314623 odd primes, max=4479941, lim=4479957
chunk_size=2048 (16384b), block_size=65536 (524288b), chunks=32, blocks=2
Thread 0 starting
Detected 160 multiprocessors (800 SPUs) on device 0.
Device 0 is a AMD Radeon HD 4850.
nstep changed to 22
CL setup complete.
cthread_count = 10240
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...
Retrieve successful...

Thread 0 completedthread_cleanup: 0

Waiting for threads to exit
Sieve complete: 20070000000000 <= p < 20070020000000
Found 0 factors
count=652993,sum=0xb5e05bf97b0cfbd5
Elapsed time: 45.86 sec. (0.01 init + 45.85 sieve) at 440198 p/sec.
Processor time: 4.70 sec. (0.01 init + 4.69 sieve) at 4308255 p/sec.
Average processor utilization: 1.44 (init), 0.10 (sieve)

ID: 29967 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Thanks, there is a new build at http://www.pyramid-productions.net/downloads/tpsieve-cl.zip.

This uses the original opencl kernels from the linux/ATI version - these wouldn't compile for my nvidia card without some changes, so it's possible I broke something for the ATI version. Can you try again with this one?

Cheers

- Iain

ID: 30102 |

Phroon
Send message
Joined: 27 Sep 10
Posts: 10
ID: 68760
Credit: 70,711,324
RAC: 0

That version doesn't build on ATI Either:
(Looks like the board eats leading spaces, even when inside a code block)

$ ./tpsieve-cl-x86_64-mac -p20070e9 -P20070020e6 -k5 -K9999 -N 3000000 -T -M2 -c 60 > out.txt
Sieve started: 20070000000000 <= p < 20070020000000
Thread 0 starting
Detected 160 multiprocessors (800 SPUs) on device 0.
Device 0 is a AMD Radeon HD 4850.
Error: Building Program (clBuildProgram): Program build failure
<program source>:47:8: error: used type 'long __attribute__((ext_vector_type(2)))' where arithmetic or pointer type is required
rcx += (rax!=(VLONG)VECTORIZE(0))?((VLONG)VECTORIZE(1)):((VLONG)VECTORIZE(0));
^
<program source>:50:7: error: used type 'long __attribute__((ext_vector_type(2)))' where arithmetic or pointer type is required
rax = (rax>N)?rcx:rax;
^
<program source>:57:7: error: used type 'long __attribute__((ext_vector_type(2)))' where arithmetic or pointer type is required
rax = (rax>N)?rcx:rax;
^
<program source>:69:8: error: used type 'long __attribute__((ext_vector_type(2)))' where arithmetic or pointer type is required
rcx += (rax!=(VLONG)VECTORIZE(0))?((VLONG)VECTORIZE(1)):((VLONG)VECTORIZE(0));
^
<program source>:72:7: error: used type 'long __attribute__((ext_vector_type(2)))' where arithmetic or pointer type is required
rax = (rax>N)?rcx:rax;
^
<program source>:77:1: error: used type 'long __attribute__((ext_vector_type(2)))' where arithmetic or pointer type is required
SHIFTMOD_REDCX(21)
^
<program source>:76:163: note: instantiated from:
#define SHIFTMOD_REDCX(NSTEP) VLONG shiftmod_REDC##NSTEP (const VLONG a, const VLONG N, VLONG rax) { VLONG rcx; rax = rax << (64-NSTEP); rcx = a >> NSTEP; rcx += (rax!=(VLONG)VECTORIZE(0))?((VLONG)VECTORIZE(1)):((VLONG)VECTORIZE(0)); rax = mad_hi(rax, N, rcx); rcx = rax - N; rax = (rax>N)?rcx:rax; return rax; }
^
<program source>:77:1: error: used type 'long __attribute__((ext_vector_type(2)))' where arithmetic or pointer type is required
SHIFTMOD_REDCX(21)
^
<program source>:76:283: note: instantiated from:
#define SHIFTMOD_REDCX(NSTEP) VLONG shiftmod_REDC##NSTEP (const VLONG a, const VLONG N, VLONG rax) { VLONG rcx; rax = rax << (64-NSTEP); rcx = a >> NSTEP; rcx += (rax!=(VLONG)VECTORIZE(0))?((VLONG)VECTORIZE(1)):((VLONG)VECTORIZE(0)); rax = mad_hi(rax, N, rcx); rcx = rax - N; rax = (rax>N)?rcx:rax; return rax; }
^
<program source>:78:1: error: used type 'long __attribute__((ext_vector_type(2)))' where arithmetic or pointer type is required
SHIFTMOD_REDCX(42)
^
<program source>:76:163: note: instantiated from:
#define SHIFTMOD_REDCX(NSTEP) VLONG shiftmod_REDC##NSTEP (const VLONG a, const VLONG N, VLONG rax) { VLONG rcx; rax = rax << (64-NSTEP); rcx = a >> NSTEP; rcx += (rax!=(VLONG)VECTORIZE(0))?((VLONG)VECTORIZE(1)):((VLONG)VECTORIZE(0)); rax = mad_hi(rax, N, rcx); rcx = rax - N; rax = (rax>N)?rcx:rax; return rax; }
^
<program source>:78:1: error: used type 'long __attribute__((ext_vector_type(2)))' where arithmetic or pointer type is required
SHIFTMOD_REDCX(42)
^
<program source>:76:283: note: instantiated from:
#define SHIFTMOD_REDCX(NSTEP) VLONG shiftmod_REDC##NSTEP (const VLONG a, const VLONG N, VLONG rax) { VLONG rcx; rax = rax << (64-NSTEP); rcx = a >> NSTEP; rcx += (rax!=(VLONG)VECTORIZE(0))?((VLONG)VECTORIZE(1)):((VLONG)VECTORIZE(0)); rax = mad_hi(rax, N, rcx); rcx = rax - N; rax = (rax>N)?rcx:rax; return rax; }
^
<program source>:80:142: error: used type 'long __attribute__((ext_vector_type(2)))' where arithmetic or pointer type is required
VLONG shiftmod_REDCX (const VLONG a, const VLONG N, VLONG rax, const uint step) { VLONG rcx; rax = rax << (64-step); rcx = a >> step; rcx += (rax!=(VLONG)VECTORIZE(0))?((VLONG)VECTORIZE(1)):((VLONG)VECTORIZE(0)); rax = mad_hi(rax, N, rcx); rcx = rax - N; rax = (rax>N)?rcx:rax; return rax; }
^
<program source>:80:262: error: used type 'long __attribute__((ext_vector_type(2)))' where arithmetic or pointer type is required
VLONG shiftmod_REDCX (const VLONG a, const VLONG N, VLONG rax, const uint step) { VLONG rcx; rax = rax << (64-step); rcx = a >> step; rcx += (rax!=(VLONG)VECTORIZE(0))?((VLONG)VECTORIZE(1)):((VLONG)VECTORIZE(0)); rax = mad_hi(rax, N, rcx); rcx = rax - N; rax = (rax>N)?rcx:rax; return rax; }
^
<program source>:87:6: error: used type 'long __attribute__((ext_vector_type(2)))' where arithmetic or pointer type is required
r += ((r&((VLONG)VECTORIZE(1))) == (VLONG)VECTORIZE(1))?N:((VLONG)VECTORIZE(0));
^
<program source>:122:1:{122:1-122:7}: error: no matching overload found for arguments of type 'int, uint, uint __attribute__((address_space(1)))*'
VSTORE(0, i, factor_found_arr);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
<program source>:97:16: note: instantiated from:
#define VSTORE VECSIZEIT(vstore,VECSIZE)
^
<program source>:15:24: note: instantiated from:
#define VECSIZEIT(x,y) VECSIZEIT2(x,y)
^
<scratch space>:83:1: note: instantiated from:
vstore2
^
/System/Library/Frameworks/OpenCL.framework/Resources/cl_kernel.h:483:29: note: instantiated from:
#define vstore2( X,Y,Z ) __builtin_overload( 3, X, Y, Z, __vstore2_i8, __vstore2_u8, __vstore2_i16, __vstore2_u16, __vstore2_i32, __vstore2_u32, __vstore2_i64, __vstore2_u64, __vstore2_f, __vstore2_d, \
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
<program source>:208:1: error: used type 'long __attribute__((ext_vector_type(2)))' where arithmetic or pointer type is required
TWIN_CHOOSE_EVEN_K0
^
<program source>:166:36: note: instantiated from:
#define TWIN_CHOOSE_EVEN_K0 kpos = (((k0) & (VLONG)VECTORIZE(1)) == (VLONG)VECTORIZE(1))?(my_P - k0):k0;
^
<program source>:221:1: error: used type 'long __attribute__((ext_vector_type(2)))' where arithmetic or pointer type is required
TWIN_CHOOSE_EVEN
^
<program source>:167:33: note: instantiated from:
#define TWIN_CHOOSE_EVEN kpos = (((kpos) & (VLONG)VECTORIZE(1)) == (VLONG)VECTORIZE(1))?(my_P - kpos):kpos;
^
<program source>:226:1: error: used type 'long __attribute__((ext_vector_type(2)))' where arithmetic or pointer type is required
TWIN_CHOOSE_EVEN
^
<program source>:167:33: note: instantiated from:
#define TWIN_CHOOSE_EVEN kpos = (((kpos) & (VLONG)VECTORIZE(1)) == (VLONG)VECTORIZE(1))?(my_P - kpos):kpos;
^

ID: 30117 |

oldDirty

Send message
Joined: 30 Jun 10
Posts: 21
ID: 63083
Credit: 8,191,526
RAC: 0

I have a hd5870 on a 2x Xeon Sys, waiting for openCL work, MAC OS X 10.6.6.
____________

Public Energy -Crunch da Power-

ID: 30953 |

[AF>Le_Pommier>MacBidouille.com]Prof
Send message
Joined: 16 Dec 09
Posts: 4
ID: 51927
Credit: 345,837,834
RAC: 103,056

I have a hd5870 on a 2x Xeon Sys, waiting for openCL work, MAC OS X 10.6.6.

Hi!! Where do you find the app for mac??
Thx

ID: 31003 |

Martin P.
Send message
Joined: 10 Mar 10
Posts: 3
ID: 56691
Credit: 234,315,265
RAC: 0

I have a hd5870 on a 2x Xeon Sys, waiting for openCL work, MAC OS X 10.6.6.

Hi!! Where do you find the app for mac??
Thx

Same question.

Modellname: Mac Pro
Modell-Identifizierung: MacPro4,1
Prozessortyp: Quad-Core Intel Xeon
Prozessorgeschwindigkeit: 2,66 GHz
Anzahl der Prozessoren: 2
Gesamtzahl der Kerne: 8
Chipsatz-Modell: ATI Radeon HD 4870

____________

ID: 31224 |

pabliedung

Volunteer tester
Send message
Joined: 2 Jan 11
Posts: 87
ID: 79535
Credit: 266,553,765
RAC: 33,258

I also have an iMac i7 2.93Ghz with an ATI 5750 waiting for work.

I am willing to test the app if someone can post a link.

thanks!

ID: 31239 |

rroonnaalldd

Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 3 Jul 09
Posts: 1213
ID: 42893
Credit: 34,634,263
RAC: 0

I also have an iMac i7 2.93Ghz with an ATI 5750 waiting for work.

I am willing to test the app if someone can post a link.

thanks!

"Iain Bethune" wrote:

http://www.pyramid-productions.net/downloads/tpsieve-cl.zip

____________
Best wishes. Knowledge is power. by jjwhalen

ID: 31240 |

pabliedung

Volunteer tester
Send message
Joined: 2 Jan 11
Posts: 87
ID: 79535
Credit: 266,553,765
RAC: 33,258

I also have an iMac i7 2.93Ghz with an ATI 5750 waiting for work.

I am willing to test the app if someone can post a link.

thanks!

"Iain Bethune" wrote:
http://www.pyramid-productions.net/downloads/tpsieve-cl.zip

unfortunately that link is dead: "file not found"

ID: 31245 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Hi, no more testing needed right now - I know there are some OpenCL compilers bugs, I need to find workarounds for some of them. Don't worry, I know many of you want this app, and I'll try and get it working, but I have other projects on the go at the moment too...

ID: 31291 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

As it happens, just got an email from apple re: one of the bugs I'd filed, suggesting it was fixed in OS 10.6.6, but this seems not to be the case, so right now we're no closer to a solution...

ID: 31363 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

One of the two outstanding bugs was fixed in 10.6.7, but the other still exists, so still no working OpenCL mac app I'm afraid...

ID: 34395 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

I have posted a new build http://www.pyramid-productions.net/downloads/tpsieve-cl-x86_64-mac.zip which contains (hopefully) workarounds for the various bugs that still exist. Could someone with an ATI card please download and test the app (see earlier in the thread for the command line to test with).

Please note, for this to work you will need to have Mac OS 10.6.8 which contained new graphics card drivers.

As always, please post your results to the thread.

Cheers

- Iain
____________
Twitter: IainBethune
Proud member of team "Aggie The Pew". Go Aggie!
3073428256125*2^1290000-1 is Prime!

ID: 38189 |

jpranke2011
Send message
Joined: 1 Jan 11
Posts: 7
ID: 79419
Credit: 16,167,917
RAC: 0

Got the following output if it helps:

ID: 38192 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Thanks, clearly the new build still does not work correctly on the ATIs (but does on my Nvidia card...) I'll try to hassle Apple for fixes to the outstanding compiler bugs, as it may be the case that my 'fixes' don't work on ATI hardware.

Cheers

- Iain
____________
Twitter: IainBethune
Proud member of team "Aggie The Pew". Go Aggie!
3073428256125*2^1290000-1 is Prime!

ID: 38193 |

Bernd Schnitker
Send message
Joined: 20 Jun 10
Posts: 34
ID: 62631
Credit: 3,722,994
RAC: 0

Any changes with OS 10.7.1 Lion? I have not yet upgraded but am watching with interest. Also any changes/ updates on 10.6.8 Snow Leopard?

ID: 39316 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

I've not yet upgraded either - so no idea if there is a new version of OpenCL in Lion, or new card drivers which might help. Still waiting to hear back from Apple from my bug reports but I've waited months before...
____________
Twitter: IainBethune
Proud member of team "Aggie The Pew". Go Aggie!
3073428256125*2^1290000-1 is Prime!

ID: 39317 |

jpranke2011
Send message
Joined: 1 Jan 11
Posts: 7
ID: 79419
Credit: 16,167,917
RAC: 0

Just ran this on v10.7.1, same code as before. Looks more promising on Lion:

./tpsieve-cl-x86_64-mac -p20070e9 -P20070020e6 -k5 -K9999 -N 3000000 -T -M2 -c 60 tpsieve version cl-0.2.3e (testing) Compiled Jul 11 2011 with GCC 4.2.1 (Apple Inc. build 5666) (dot 3) nstart=74, nstep=31 tpsieve initialized: 5 <= k <= 9999, 74 <= n < 3000000 Sieve started: 20070000000000 <= p < 20070020000000 Thread 0 starting Detected 160 multiprocessors (800 SPUs) on device 0. Device 0 is a AMD ATI Radeon HD 5750. nstep changed to 22 clBuildProgram() successful CL setup complete. cthread_count = 10240 20070000059159 | 2205*2^2472339-1 20070000105761 | 733*2^974106+1 20070000382183 | 683*2^2637373+1 20070000475957 | 4995*2^1822738+1 20070000541441 | 3243*2^1584966-1 20070000674041 | 8143*2^1397047-1 20070001141921 | 2787*2^2536178+1 20070001146497 | 4977*2^626298+1 20070001163929 | 3765*2^461308+1 20070001302811 | 7669*2^725426+1 20070001379699 | 5781*2^2459443+1 20070001425977 | 5821*2^1775248+1 20070001823101 | 7647*2^1022532-1 20070001843627 | 4955*2^248864-1 20070002245151 | 1221*2^646983+1 20070002606341 | 4809*2^497683+1 20070002680493 | 6455*2^1778260-1 20070003067151 | 9259*2^1869285-1 20070004531271 | 4755*2^2642856-1 20070004606567 | 5269*2^257879-1 20070004750837 | 525*2^1636405+1 20070004787473 | 1153*2^2819766+1 20070004816819 | 6699*2^1215561+1 20070004848679 | 9911*2^2058269+1 20070004886501 | 6401*2^2784714-1 20070004973621 | 705*2^1329798-1 20070005282267 | 1055*2^1575895+1 20070005348311 | 3791*2^2114525+1 20070005404357 | 7893*2^796719-1 20070005914001 | 9847*2^1634140+1 20070006110513 | 2335*2^2220548+1 20070006186677 | 2413*2^1924519-1 20070006187837 | 9923*2^287853+1 20070006365161 | 569*2^782235+1 20070006603683 | 8229*2^2745282+1 20070006754367 | 5025*2^2568467-1 20070006875981 | 1645*2^965954+1 20070007049707 | 2149*2^153375-1 20070007170259 | 3889*2^49730+1 20070007400801 | 4261*2^2489220+1 20070007462667 | 917*2^1923231+1 20070007529777 | 1367*2^1532230-1 20070007530649 | 599*2^1252359+1 20070008013637 | 2097*2^2286269-1 20070008026721 | 9231*2^2302830-1 20070008213771 | 4041*2^1187467-1 20070008329039 | 9065*2^832569+1 20070008576117 | 153*2^560883-1 20070008674201 | 9713*2^2549724-1 20070008917537 | 6671*2^1260830-1 20070008991223 | 5523*2^101606-1 20070009203669 | 417*2^817661-1 20070009223217 | 3789*2^2853163+1 20070009299779 | 6221*2^2940469+1 20070010348009 | 5177*2^2176044-1 20070010618159 | 2597*2^2763167+1 20070010645547 | 2701*2^1048169-1 20070010903867 | 8493*2^523854-1 20070010984423 | 231*2^2927044+1 20070011210369 | 649*2^2493015-1 20070011245997 | 6565*2^2452502+1 20070011258593 | 2187*2^1101790-1 20070011593871 | 5305*2^2360059-1 20070011888243 | 8371*2^2418889-1 20070012085751 | 1337*2^1688050-1 20070012086261 | 847*2^553088+1 20070012163009 | 901*2^956937-1 20070012229009 | 1445*2^1110320-1 20070012297143 | 4009*2^181277-1 20070012309493 | 2781*2^1034907-1 20070012403537 | 4747*2^733649-1 20070012507137 | 7061*2^1730371+1 20070012579293 | 8145*2^851778-1 20070012797183 | 9213*2^2101908+1 20070012855791 | 5277*2^731353-1 20070012882203 | 4425*2^1603994-1 20070013269221 | 6763*2^1741799-1 20070014434517 | 4051*2^1308881-1 20070014814419 | 6923*2^1257277+1 20070014822999 | 8443*2^243383-1 20070014902183 | 6981*2^1674047+1 20070014977063 | 4809*2^470009+1 20070014982443 | 791*2^1478283+1 20070015187811 | 2685*2^2186147+1 20070015282311 | 4015*2^2902138+1 20070015687191 | 5973*2^502229+1 20070016364921 | 2723*2^1269504-1 20070016440869 | 9577*2^34776+1 20070016546427 | 3945*2^2086463+1 20070016765411 | 1421*2^1766083+1 20070016948067 | 4735*2^2548621-1 20070016973093 | 2857*2^1226993-1 20070016992617 | 7001*2^1177373+1 20070017505563 | 7627*2^401848+1 20070017763227 | 6625*2^1332032+1 20070017775449 | 8415*2^2761332+1 20070017820647 | 6955*2^1347189-1 20070017949379 | 1097*2^757927+1 20070018187309 | 5775*2^2258100-1 20070018281791 | 4685*2^1510525+1 20070018308227 | 6809*2^2911648-1 20070018425833 | 1911*2^1460578-1 20070018515161 | 1515*2^1186401-1 20070018709289 | 1045*2^2655475-1 20070019115281 | 4641*2^1797127-1 20070019117289 | 2339*2^565480-1 20070019307549 | 9835*2^2523382+1 20070019838257 | 8025*2^2120157-1 20070019870763 | 8189*2^2803+1 Thread 0 completed Waiting for threads to exit Sieve complete: 20070000000000 <= p < 20070020000000 Found 109 factors count=652993,sum=0xb5e05bf97b0cfbd5 Elapsed time: 35.19 sec. (0.01 init + 35.19 sieve) at 573675 p/sec. Processor time: 1.83 sec. (0.01 init + 1.82 sieve) at 11094902 p/sec. Average processor utilization: 1.74 (init), 0.05 (sieve)

ID: 39702 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Thanks for the results - that looks encouraging. Can any other Lion/AMD users confirm?

Cheers

- Iain
____________
Twitter: IainBethune
Proud member of team "Aggie The Pew". Go Aggie!
3073428256125*2^1290000-1 is Prime!

ID: 39712 |

pabliedung

Volunteer tester
Send message
Joined: 2 Jan 11
Posts: 87
ID: 79535
Credit: 266,553,765
RAC: 33,258

Thanks for the results - that looks encouraging. Can any other Lion/AMD users confirm?

Cheers

- Iain

Hi Ian,

can you point me to the file I need to download and run? The link that was posted earlier in this thread is not working for me.

I will be able to test it on my iMac (2010 model with ATI Radeon HD 5750) later today.

Cheers,
Philipp

ID: 39714 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Sorry for the delay - I was out of the country with the 'wrong' laptop!

I've posted three slightly different builds - I'd be grateful if you could try all three:

http://www.pyramid-productions.net/downloads/tpsieve-cl-x86_64-mac.zip
http://www.pyramid-productions.net/downloads/tpsieve-cl-x86_64-mac-patched.zip
http://www.pyramid-productions.net/downloads/tpsieve-cl-x86_64-mac-patched2.zip

The 'patched2' version should be the same as posted before, so I'm hopeful it works (on Lion). The 'patched' version has a workaround removed for a bug which Apple have reported has been fixed in Lion, so should also work. The 1st version has another workaround removed, which may be fixed in Lion, although I'm not sure, and may improve performance if it does work.

Thanks for the testing, and let me know how you get on!

Cheers

- Iain
____________
Twitter: IainBethune
Proud member of team "Aggie The Pew". Go Aggie!
3073428256125*2^1290000-1 is Prime!

ID: 40192 |

jpranke2011
Send message
Joined: 1 Jan 11
Posts: 7
ID: 79419
Credit: 16,167,917
RAC: 0

Hi Ian,

Below are outputs for the 3 files. The first one did not run at all and the other two ran successfully. The patched ran sightly faster than the patched2.

http://www.pyramid-productions.net/downloads/tpsieve-cl-x86_64-mac.zip

./tpsieve-cl-x86_64-mac -p20070e9 -P20070020e6 -k5 -K9999 -N 3000000 -T -M2 -c 60 -q tpsieve version cl-0.2.3e (testing) Compiled Sep 9 2011 with GCC 4.2.1 (Apple Inc. build 5666) (dot 3) nstart=74, nstep=31 tpsieve initialized: 5 <= k <= 9999, 74 <= n < 3000000 Sieve started: 20070000000000 <= p < 20070020000000 Thread 0 starting Detected 160 multiprocessors (800 SPUs) on device 0. Device 0 is a AMD ATI Radeon HD 5750. nstep changed to 22 Error: Building Program (clBuildProgram): Program build failure <program source>:29:1: warning: no previous prototype for function 'invmod2pow_ul' invmod2pow_ul (const VLONG n) ^ <program source>:40:7: warning: no previous prototype for function 'mulmod_REDC' VLONG mulmod_REDC (const VLONG a, const VLONG b, ^ <program source>:53:7: warning: no previous prototype for function 'onemod_REDC' VLONG onemod_REDC(const VLONG N, VLONG rax) { ^ <program source>:60:7: warning: no previous prototype for function 'mod_REDC' VLONG mod_REDC(const VLONG a, const VLONG N, const VLONG Ns) { ^ <program source>:63:7: warning: no previous prototype for function 'shiftmod_REDC' VLONG shiftmod_REDC (const VLONG a, ^ <program source>:77:1: warning: no previous prototype for function 'shiftmod_REDC21' SHIFTMOD_REDCX(21) ^ <scratch space>:94:1: note: instantiated from: shiftmod_REDC21 ^ <program source>:78:1: warning: no previous prototype for function 'shiftmod_REDC42' SHIFTMOD_REDCX(42) ^ <scratch space>:97:1: note: instantiated from: shiftmod_REDC42 ^ <program source>:80:7: warning: no previous prototype for function 'shiftmod_REDCX' VLONG shiftmod_REDCX (const VLONG a, const VLONG N, VLONG rax, const uint step) { VLONG rcx; rax = rax << (64-step); rcx = a >> step; rcx += (rax!=(VLONG)VECTORIZE(0))?((VLONG)VECTORIZE(1)):((VLONG)VECTORIZE(0)); rax = mad_hi(rax, N, rcx); rcx = rax - N; rax = (rax>N)?rcx:rax; return rax; } ^ <program source>:82:1: warning: no previous prototype for function 'invpowmod_REDClr' invpowmod_REDClr (const VLONG N, const VLONG Ns, int bbits, const ulong r0) { ^ <program source>:122:1: error: no matching overload found for arguments of type 'int, uint, uint __attribute__((address_space(1))) *' VSTORE(0, i, factor_found_arr); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ <program source>:97:16: note: instantiated from: #define VSTORE VECSIZEIT(vstore,VECSIZE) ^ <program source>:15:24: note: instantiated from: #define VECSIZEIT(x,y) VECSIZEIT2(x,y) ^ <scratch space>:83:1: note: instantiated from: vstore2 ^ /System/Library/Frameworks/OpenCL.framework/Resources/cl_kernel.h:508:29: note: instantiated from: #define vstore2( X,Y,Z ) __builtin_overload( 3, X, Y, Z, __vstore2_i8, __vstore2_u8, __vstore2_i16, __vstore2_u16, __vstore2_i32, __vstore2_u32, __vstore2_i64, __vstore2_u64, __vstore2_f, __vstore2_d, \ ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

http://www.pyramid-productions.net/downloads/tpsieve-cl-x86_64-mac-patched.zip

./tpsieve-cl-x86_64-mac-patched -p20070e9 -P20070020e6 -k5 -K9999 -N 3000000 -T -M2 -c 60 -q tpsieve version cl-0.2.3e (testing) Compiled Sep 9 2011 with GCC 4.2.1 (Apple Inc. build 5666) (dot 3) nstart=74, nstep=31 tpsieve initialized: 5 <= k <= 9999, 74 <= n < 3000000 Sieve started: 20070000000000 <= p < 20070020000000 Thread 0 starting Detected 160 multiprocessors (800 SPUs) on device 0. Device 0 is a AMD ATI Radeon HD 5750. nstep changed to 22 clBuildProgram() successful CL setup complete. cthread_count = 10240 Thread 0 completed Waiting for threads to exit Sieve complete: 20070000000000 <= p < 20070020000000 Found 109 factors count=652993,sum=0xb5e05bf97b0cfbd5 Elapsed time: 33.41 sec. (0.01 init + 33.40 sieve) at 604268 p/sec. Processor time: 1.50 sec. (0.01 init + 1.49 sieve) at 13546530 p/sec. Average processor utilization: 1.77 (init), 0.04 (sieve)

http://www.pyramid-productions.net/downloads/tpsieve-cl-x86_64-mac-patched2.zip

./tpsieve-cl-x86_64-mac-patched2 -p20070e9 -P20070020e6 -k5 -K9999 -N 3000000 -T -M2 -c 60 -q tpsieve version cl-0.2.3e (testing) Compiled Sep 9 2011 with GCC 4.2.1 (Apple Inc. build 5666) (dot 3) nstart=74, nstep=31 tpsieve initialized: 5 <= k <= 9999, 74 <= n < 3000000 Sieve started: 20070000000000 <= p < 20070020000000 Thread 0 starting Detected 160 multiprocessors (800 SPUs) on device 0. Device 0 is a AMD ATI Radeon HD 5750. nstep changed to 22 clBuildProgram() successful CL setup complete. cthread_count = 10240 Thread 0 completed Waiting for threads to exit Sieve complete: 20070000000000 <= p < 20070020000000 Found 109 factors count=652993,sum=0xb5e05bf97b0cfbd5 Elapsed time: 33.55 sec. (0.01 init + 33.54 sieve) at 601829 p/sec. Processor time: 1.51 sec. (0.01 init + 1.49 sieve) at 13522734 p/sec. Average processor utilization: 1.78 (init), 0.04 (sieve)

ID: 40212 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Thanks, did you cut out the list of found factors in the middle of the output?

And I think I fixed the first error - please try http://www.pyramid-productions.net/downloads/tpsieve-cl-x86_64-mac.zip again and let me know how it goes.

I think I may have to upgrade to Lion soon.

Cheers

- Iain

____________
Twitter: IainBethune
Proud member of team "Aggie The Pew". Go Aggie!
3073428256125*2^1290000-1 is Prime!

ID: 40225 |

jpranke2011
Send message
Joined: 1 Jan 11
Posts: 7
ID: 79419
Credit: 16,167,917
RAC: 0

Hi, I ran with -q (quiet mode) to only show the metrics used to calculate the factors. Looks like the update worked and ran successfully as below.

Anything else I can test (larger factor range, etc)?

I totally recommend Lion. Worth the upgrade!

./tpsieve-cl-x86_64-mac -p20070e9 -P20070020e6 -k5 -K9999 -N 3000000 -T -M2 -c 60 -q tpsieve version cl-0.2.3e (testing) Compiled Sep 11 2011 with GCC 4.2.1 (Apple Inc. build 5666) (dot 3) nstart=74, nstep=31 tpsieve initialized: 5 <= k <= 9999, 74 <= n < 3000000 Sieve started: 20070000000000 <= p < 20070020000000 Thread 0 starting Detected 160 multiprocessors (800 SPUs) on device 0. Device 0 is a AMD ATI Radeon HD 5750. nstep changed to 22 clBuildProgram() successful CL setup complete. cthread_count = 10240 Thread 0 completed Waiting for threads to exit Sieve complete: 20070000000000 <= p < 20070020000000 Found 109 factors count=652993,sum=0xb5e05bf97b0cfbd5 Elapsed time: 33.43 sec. (0.01 init + 33.42 sieve) at 603911 p/sec. Processor time: 1.77 sec. (0.01 init + 1.75 sieve) at 11507468 p/sec. Average processor utilization: 1.82 (init), 0.05 (sieve)

ID: 40232 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

One more test to do - the BOINC-enabled app is http://www.pyramid-productions.net/downloads/tpsieve-cl-boinc-x86_64-mac.zip.

Can you give it a try as before please? Also if anyone else with Lion reads this, any extra testing would be gratefully received (esp. with different ATI cards).

Cheers

- Iain
____________
Twitter: IainBethune
Proud member of team "Aggie The Pew". Go Aggie!
3073428256125*2^1290000-1 is Prime!

ID: 40249 |

jpranke2011
Send message
Joined: 1 Jan 11
Posts: 7
ID: 79419
Credit: 16,167,917
RAC: 0

This time I ran with full output and results look good:

./tpsieve-cl-boinc-x86_64-mac -p20070e9 -P20070020e6 -k5 -K9999 -N 3000000 -T -M2 -c 60 tpsieve version cl-0.2.3e (testing) Compiled Sep 12 2011 with GCC 4.2.1 (Apple Inc. build 5666) (dot 3) nstart=74, nstep=31 tpsieve initialized: 5 <= k <= 9999, 74 <= n < 3000000 nstep changed to 22 CL setup complete. cthread_count = 10240 20070000059159 | 2205*2^2472339-1 20070000105761 | 733*2^974106+1 20070000382183 | 683*2^2637373+1 20070000475957 | 4995*2^1822738+1 20070000541441 | 3243*2^1584966-1 20070000674041 | 8143*2^1397047-1 20070001141921 | 2787*2^2536178+1 20070001146497 | 4977*2^626298+1 20070001163929 | 3765*2^461308+1 20070001302811 | 7669*2^725426+1 20070001379699 | 5781*2^2459443+1 20070001425977 | 5821*2^1775248+1 20070001823101 | 7647*2^1022532-1 20070001843627 | 4955*2^248864-1 20070002245151 | 1221*2^646983+1 20070002606341 | 4809*2^497683+1 20070002680493 | 6455*2^1778260-1 20070003067151 | 9259*2^1869285-1 20070004531271 | 4755*2^2642856-1 20070004606567 | 5269*2^257879-1 20070004750837 | 525*2^1636405+1 20070004787473 | 1153*2^2819766+1 20070004816819 | 6699*2^1215561+1 20070004848679 | 9911*2^2058269+1 20070004886501 | 6401*2^2784714-1 20070004973621 | 705*2^1329798-1 20070005282267 | 1055*2^1575895+1 20070005348311 | 3791*2^2114525+1 20070005404357 | 7893*2^796719-1 20070005914001 | 9847*2^1634140+1 20070006110513 | 2335*2^2220548+1 20070006186677 | 2413*2^1924519-1 20070006187837 | 9923*2^287853+1 20070006365161 | 569*2^782235+1 20070006603683 | 8229*2^2745282+1 20070006754367 | 5025*2^2568467-1 20070006875981 | 1645*2^965954+1 20070007049707 | 2149*2^153375-1 20070007170259 | 3889*2^49730+1 20070007400801 | 4261*2^2489220+1 20070007462667 | 917*2^1923231+1 20070007529777 | 1367*2^1532230-1 20070007530649 | 599*2^1252359+1 20070008013637 | 2097*2^2286269-1 20070008026721 | 9231*2^2302830-1 20070008213771 | 4041*2^1187467-1 20070008329039 | 9065*2^832569+1 20070008576117 | 153*2^560883-1 20070008674201 | 9713*2^2549724-1 20070008917537 | 6671*2^1260830-1 20070008991223 | 5523*2^101606-1 20070009203669 | 417*2^817661-1 20070009223217 | 3789*2^2853163+1 20070009299779 | 6221*2^2940469+1 20070010348009 | 5177*2^2176044-1 20070010618159 | 2597*2^2763167+1 20070010645547 | 2701*2^1048169-1 20070010903867 | 8493*2^523854-1 20070010984423 | 231*2^2927044+1 20070011210369 | 649*2^2493015-1 20070011245997 | 6565*2^2452502+1 20070011258593 | 2187*2^1101790-1 20070011593871 | 5305*2^2360059-1 20070011888243 | 8371*2^2418889-1 20070012085751 | 1337*2^1688050-1 20070012086261 | 847*2^553088+1 20070012163009 | 901*2^956937-1 20070012229009 | 1445*2^1110320-1 20070012297143 | 4009*2^181277-1 20070012309493 | 2781*2^1034907-1 20070012403537 | 4747*2^733649-1 20070012507137 | 7061*2^1730371+1 20070012579293 | 8145*2^851778-1 20070012797183 | 9213*2^2101908+1 20070012855791 | 5277*2^731353-1 20070012882203 | 4425*2^1603994-1 20070013269221 | 6763*2^1741799-1 20070014434517 | 4051*2^1308881-1 20070014814419 | 6923*2^1257277+1 20070014822999 | 8443*2^243383-1 20070014902183 | 6981*2^1674047+1 20070014977063 | 4809*2^470009+1 20070014982443 | 791*2^1478283+1 20070015187811 | 2685*2^2186147+1 20070015282311 | 4015*2^2902138+1 20070015687191 | 5973*2^502229+1 20070016364921 | 2723*2^1269504-1 20070016440869 | 9577*2^34776+1 20070016546427 | 3945*2^2086463+1 20070016765411 | 1421*2^1766083+1 20070016948067 | 4735*2^2548621-1 20070016973093 | 2857*2^1226993-1 20070016992617 | 7001*2^1177373+1 20070017505563 | 7627*2^401848+1 20070017763227 | 6625*2^1332032+1 20070017775449 | 8415*2^2761332+1 20070017820647 | 6955*2^1347189-1 20070017949379 | 1097*2^757927+1 20070018187309 | 5775*2^2258100-1 20070018281791 | 4685*2^1510525+1 20070018308227 | 6809*2^2911648-1 20070018425833 | 1911*2^1460578-1 20070018515161 | 1515*2^1186401-1 20070018709289 | 1045*2^2655475-1 20070019115281 | 4641*2^1797127-1 20070019117289 | 2339*2^565480-1 20070019307549 | 9835*2^2523382+1 20070019838257 | 8025*2^2120157-1 20070019870763 | 8189*2^2803+1 Found 109 factors

ID: 40254 |

pabliedung

Volunteer tester
Send message
Joined: 2 Jan 11
Posts: 87
ID: 79535
Credit: 266,553,765
RAC: 33,258

One more test to do - the BOINC-enabled app is http://www.pyramid-productions.net/downloads/tpsieve-cl-boinc-x86_64-mac.zip.

Can you give it a try as before please? Also if anyone else with Lion reads this, any extra testing would be gratefully received (esp. with different ATI cards).

Cheers

- Iain

Hi Ian,

below is the output from my 2010 iMac with an ATI Radeon 5750:

iMac-i7:opencl test philipp$ ./tpsieve-cl-boinc-x86_64-mac -p20070e9 -P20070020e6 -k5 -K9999 -N 3000000 -T -M2 -c 60
tpsieve version cl-0.2.3e (testing)
Compiled Sep 12 2011 with GCC 4.2.1 (Apple Inc. build 5666) (dot 3)
nstart=74, nstep=31
tpsieve initialized: 5 <= k <= 9999, 74 <= n < 3000000
nstep changed to 22
CL setup complete.
cthread_count = 10240
20070000059159 | 2205*2^2472339-1
20070000105761 | 733*2^974106+1
20070000382183 | 683*2^2637373+1
20070000475957 | 4995*2^1822738+1
20070000541441 | 3243*2^1584966-1
20070000674041 | 8143*2^1397047-1
20070001141921 | 2787*2^2536178+1
20070001146497 | 4977*2^626298+1
20070001163929 | 3765*2^461308+1
20070001302811 | 7669*2^725426+1
20070001379699 | 5781*2^2459443+1
20070001425977 | 5821*2^1775248+1
20070001823101 | 7647*2^1022532-1
20070001843627 | 4955*2^248864-1
20070002245151 | 1221*2^646983+1
20070002606341 | 4809*2^497683+1
20070002680493 | 6455*2^1778260-1
20070003067151 | 9259*2^1869285-1
20070004531271 | 4755*2^2642856-1
20070004606567 | 5269*2^257879-1
20070004750837 | 525*2^1636405+1
20070004787473 | 1153*2^2819766+1
20070004816819 | 6699*2^1215561+1
20070004848679 | 9911*2^2058269+1
20070004886501 | 6401*2^2784714-1
20070004973621 | 705*2^1329798-1
20070005282267 | 1055*2^1575895+1
20070005348311 | 3791*2^2114525+1
20070005404357 | 7893*2^796719-1
20070005914001 | 9847*2^1634140+1
20070006110513 | 2335*2^2220548+1
20070006186677 | 2413*2^1924519-1
20070006187837 | 9923*2^287853+1
20070006365161 | 569*2^782235+1
20070006603683 | 8229*2^2745282+1
20070006754367 | 5025*2^2568467-1
20070006875981 | 1645*2^965954+1
20070007049707 | 2149*2^153375-1
20070007170259 | 3889*2^49730+1
20070007400801 | 4261*2^2489220+1
20070007462667 | 917*2^1923231+1
20070007529777 | 1367*2^1532230-1
20070007530649 | 599*2^1252359+1
20070008013637 | 2097*2^2286269-1
20070008026721 | 9231*2^2302830-1
20070008213771 | 4041*2^1187467-1
20070008329039 | 9065*2^832569+1
20070008576117 | 153*2^560883-1
20070008674201 | 9713*2^2549724-1
20070008917537 | 6671*2^1260830-1
20070008991223 | 5523*2^101606-1
20070009203669 | 417*2^817661-1
20070009223217 | 3789*2^2853163+1
20070009299779 | 6221*2^2940469+1
20070010348009 | 5177*2^2176044-1
20070010618159 | 2597*2^2763167+1
20070010645547 | 2701*2^1048169-1
20070010903867 | 8493*2^523854-1
20070010984423 | 231*2^2927044+1
20070011210369 | 649*2^2493015-1
20070011245997 | 6565*2^2452502+1
20070011258593 | 2187*2^1101790-1
20070011593871 | 5305*2^2360059-1
20070011888243 | 8371*2^2418889-1
20070012085751 | 1337*2^1688050-1
20070012086261 | 847*2^553088+1
20070012163009 | 901*2^956937-1
20070012229009 | 1445*2^1110320-1
20070012297143 | 4009*2^181277-1
20070012309493 | 2781*2^1034907-1
20070012403537 | 4747*2^733649-1
20070012507137 | 7061*2^1730371+1
20070012579293 | 8145*2^851778-1
20070012797183 | 9213*2^2101908+1
20070012855791 | 5277*2^731353-1
20070012882203 | 4425*2^1603994-1
20070013269221 | 6763*2^1741799-1
20070014434517 | 4051*2^1308881-1
20070014814419 | 6923*2^1257277+1
20070014822999 | 8443*2^243383-1
20070014902183 | 6981*2^1674047+1
20070014977063 | 4809*2^470009+1
20070014982443 | 791*2^1478283+1
20070015187811 | 2685*2^2186147+1
20070015282311 | 4015*2^2902138+1
20070015687191 | 5973*2^502229+1
20070016364921 | 2723*2^1269504-1
20070016440869 | 9577*2^34776+1
20070016546427 | 3945*2^2086463+1
20070016765411 | 1421*2^1766083+1
20070016948067 | 4735*2^2548621-1
20070016973093 | 2857*2^1226993-1
20070016992617 | 7001*2^1177373+1
20070017505563 | 7627*2^401848+1
20070017763227 | 6625*2^1332032+1
20070017775449 | 8415*2^2761332+1
20070017820647 | 6955*2^1347189-1
20070017949379 | 1097*2^757927+1
20070018187309 | 5775*2^2258100-1
20070018281791 | 4685*2^1510525+1
20070018308227 | 6809*2^2911648-1
20070018425833 | 1911*2^1460578-1
20070018515161 | 1515*2^1186401-1
20070018709289 | 1045*2^2655475-1
20070019115281 | 4641*2^1797127-1
20070019117289 | 2339*2^565480-1
20070019307549 | 9835*2^2523382+1
20070019838257 | 8025*2^2120157-1
20070019870763 | 8189*2^2803+1
Found 109 factors

I am unable to download any of the first three builds you posted. Let me know if the links changed or if I can find them somewhere else.

ID: 40256 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Thanks guys, I took down the first three builds as they were just works-in-progress. The BOINC one looks like it will be the one we will go with, unless anyone finds any further issues with it.

Cheers

- Iain
____________
Twitter: IainBethune
Proud member of team "Aggie The Pew". Go Aggie!
3073428256125*2^1290000-1 is Prime!

ID: 40262 |

Phroon
Send message
Joined: 27 Sep 10
Posts: 10
ID: 68760
Credit: 70,711,324
RAC: 0

Looks good here on Lion:

Model Name: iMac
Model Identifier: iMac11,1
Processor Name: Intel Core i7
Processor Speed: 2.8 GHz
Number Of Processors: 1
Total Number Of Cores: 4
L2 Cache (per core): 256 KB
L3 Cache: 8 MB
Memory: 8 GB
System Version: Mac OS X 10.7.1 (11B26)
Chipset Model: ATI Radeon HD 4850

./tpsieve-cl-boinc-x86_64-mac -p20070e9 -P20070020e6 -k5 -K9999 -N 3000000 -T -M2 -c 60
tpsieve version cl-0.2.3e (testing)
Compiled Sep 12 2011 with GCC 4.2.1 (Apple Inc. build 5666) (dot 3)
nstart=74, nstep=31
tpsieve initialized: 5 <= k <= 9999, 74 <= n < 3000000
nstep changed to 22
CL setup complete.
cthread_count = 10240
20070000059159 | 2205*2^2472339-1
20070000105761 | 733*2^974106+1
20070000382183 | 683*2^2637373+1
20070000475957 | 4995*2^1822738+1
20070000541441 | 3243*2^1584966-1
20070000674041 | 8143*2^1397047-1
20070001141921 | 2787*2^2536178+1
20070001146497 | 4977*2^626298+1
20070001163929 | 3765*2^461308+1
20070001302811 | 7669*2^725426+1
20070001379699 | 5781*2^2459443+1
20070001425977 | 5821*2^1775248+1
20070001823101 | 7647*2^1022532-1
20070001843627 | 4955*2^248864-1
20070002245151 | 1221*2^646983+1
20070002606341 | 4809*2^497683+1
20070002680493 | 6455*2^1778260-1
20070003067151 | 9259*2^1869285-1
20070004531271 | 4755*2^2642856-1
20070004606567 | 5269*2^257879-1
20070004750837 | 525*2^1636405+1
20070004787473 | 1153*2^2819766+1
20070004816819 | 6699*2^1215561+1
20070004848679 | 9911*2^2058269+1
20070004886501 | 6401*2^2784714-1
20070004973621 | 705*2^1329798-1
20070005282267 | 1055*2^1575895+1
20070005348311 | 3791*2^2114525+1
20070005404357 | 7893*2^796719-1
20070005914001 | 9847*2^1634140+1
20070006110513 | 2335*2^2220548+1
20070006186677 | 2413*2^1924519-1
20070006187837 | 9923*2^287853+1
20070006365161 | 569*2^782235+1
20070006603683 | 8229*2^2745282+1
20070006754367 | 5025*2^2568467-1
20070006875981 | 1645*2^965954+1
20070007049707 | 2149*2^153375-1
20070007170259 | 3889*2^49730+1
20070007400801 | 4261*2^2489220+1
20070007462667 | 917*2^1923231+1
20070007529777 | 1367*2^1532230-1
20070007530649 | 599*2^1252359+1
20070008013637 | 2097*2^2286269-1
20070008026721 | 9231*2^2302830-1
20070008213771 | 4041*2^1187467-1
20070008329039 | 9065*2^832569+1
20070008576117 | 153*2^560883-1
20070008674201 | 9713*2^2549724-1
20070008917537 | 6671*2^1260830-1
20070008991223 | 5523*2^101606-1
20070009203669 | 417*2^817661-1
20070009223217 | 3789*2^2853163+1
20070009299779 | 6221*2^2940469+1
20070010348009 | 5177*2^2176044-1
20070010618159 | 2597*2^2763167+1
20070010645547 | 2701*2^1048169-1
20070010903867 | 8493*2^523854-1
20070010984423 | 231*2^2927044+1
20070011210369 | 649*2^2493015-1
20070011245997 | 6565*2^2452502+1
20070011258593 | 2187*2^1101790-1
20070011593871 | 5305*2^2360059-1
20070011888243 | 8371*2^2418889-1
20070012085751 | 1337*2^1688050-1
20070012086261 | 847*2^553088+1
20070012163009 | 901*2^956937-1
20070012229009 | 1445*2^1110320-1
20070012297143 | 4009*2^181277-1
20070012309493 | 2781*2^1034907-1
20070012403537 | 4747*2^733649-1
20070012507137 | 7061*2^1730371+1
20070012579293 | 8145*2^851778-1
20070012797183 | 9213*2^2101908+1
20070012855791 | 5277*2^731353-1
20070012882203 | 4425*2^1603994-1
20070013269221 | 6763*2^1741799-1
20070014434517 | 4051*2^1308881-1
20070014814419 | 6923*2^1257277+1
20070014822999 | 8443*2^243383-1
20070014902183 | 6981*2^1674047+1
20070014977063 | 4809*2^470009+1
20070014982443 | 791*2^1478283+1
20070015187811 | 2685*2^2186147+1
20070015282311 | 4015*2^2902138+1
20070015687191 | 5973*2^502229+1
20070016364921 | 2723*2^1269504-1
20070016440869 | 9577*2^34776+1
20070016546427 | 3945*2^2086463+1
20070016765411 | 1421*2^1766083+1
20070016948067 | 4735*2^2548621-1
20070016973093 | 2857*2^1226993-1
20070016992617 | 7001*2^1177373+1
20070017505563 | 7627*2^401848+1
20070017763227 | 6625*2^1332032+1
20070017775449 | 8415*2^2761332+1
20070017820647 | 6955*2^1347189-1
20070017949379 | 1097*2^757927+1
20070018187309 | 5775*2^2258100-1
20070018281791 | 4685*2^1510525+1
20070018308227 | 6809*2^2911648-1
20070018425833 | 1911*2^1460578-1
20070018515161 | 1515*2^1186401-1
20070018709289 | 1045*2^2655475-1
20070019115281 | 4641*2^1797127-1
20070019117289 | 2339*2^565480-1
20070019307549 | 9835*2^2523382+1
20070019838257 | 8025*2^2120157-1
20070019870763 | 8189*2^2803+1
Found 109 factors

ID: 40306 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Thanks, I've sent the app to the admins, so hopefully we'll see it on BOINC soon.
____________
Twitter: IainBethune
Proud member of team "Aggie The Pew". Go Aggie!
3073428256125*2^1290000-1 is Prime!

ID: 40307 |

John

Honorary cruncher
Avatar

Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0

Thanks, I've sent the app to the admins, so hopefully we'll see it on BOINC soon.

Rytis has implemented the OpenCL app for the Mac OS in the PPS (Sieve) project. Mac OS 10.7 is required. The App is 64-bit only since 10.7 only supports 64 bit.

A special thanks to Iain for making this possible.
____________

ID: 40312 |

pabliedung

Volunteer tester
Send message
Joined: 2 Jan 11
Posts: 87
ID: 79535
Credit: 266,553,765
RAC: 33,258

How can I get the BOINC Manager to download the OpenCL app?

I selected the AMD(ATI) option for PPS (Sieve) in my preferences but since my BOINC Manager reports "No usable GPUs found" no app or wu is being downloaded.
Instead I get:
Thu Sep 15 23:20:07 2011 | PrimeGrid | Scheduler request completed: got 0 new tasks
Thu Sep 15 23:20:07 2011 | PrimeGrid | No work sent
Thu Sep 15 23:20:07 2011 | PrimeGrid | Message from server: No work available for the applications you have selected. Please check your project preferences on the web site.

I know there is no ATI Stream SDK for OSX since the OpenCL drivers should be included with the OS but how can I get BOINC to recognize the GPU?

I am running BOINC Manager 6.12.33 and OSX Lion 10.7.1 with an ATI Radeon HD 5750 GPU.

Thanks,
Phil

ID: 40337 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Can you try upgrading to BOINC 6.13.1 (http://boinc.berkeley.edu/download_all.php) and see if that fixes it?

If not, check you have the latest ATI card drivers in the 'ATI Displays' section of System Preferences.

I note that none of the folk who have posted in this thread have computers where the ATI card shows up to BOINC - but let's see if this fixes it.
____________
Twitter: IainBethune
Proud member of team "Aggie The Pew". Go Aggie!
3073428256125*2^1290000-1 is Prime!

ID: 40340 |

pabliedung

Volunteer tester
Send message
Joined: 2 Jan 11
Posts: 87
ID: 79535
Credit: 266,553,765
RAC: 33,258

I upgraded to 6.13.1 but I still get "No usable GPU found".

Under System Preferences I don't see anything specific to ATI cards/drivers. On my Macbook Pro, which has an Nvidia card, I have a separate section for the CUDA drivers... unfortunately there is nothing like that on the ATI side.

I looks like it is a limitation in the BOINC Manager rather than PG?

ID: 40344 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Thanks for trying - I have posted to BOINC forums http://boinc.berkeley.edu/dev/forum_thread.php?id=6939#40209 so will see if they have any ideas.

Cheers

- Iain
____________
Twitter: IainBethune
Proud member of team "Aggie The Pew". Go Aggie!
3073428256125*2^1290000-1 is Prime!

ID: 40345 |

Jaska*

Send message
Joined: 1 Oct 10
Posts: 256
ID: 69151
Credit: 114,560,325
RAC: 0

I recently got my hands on a [url=http://www.primegrid.com/show_host_detail.php?hostid=220017
]work computer[/url] for testing purposes.

It is rather old (close to 4 years I am told) and it's got a HD 2400 or similar in it. I know that is not DP-capable, but then again I don't recall if ppsieve requires that on ATi cards.

I'm not even entirely sure it's OpenCL capable. Seems to be no way to check. Certainly no specific prefpane.

Granted also, I cannot go installing large amounts of software and driver packages on this machine, which I don't technically own. I simply administrate it. I work as a computer tech, too, so this is definitely something I can play around with in my free time.
____________

ID: 40357 |

Scott Brown

Volunteer moderator
Project administrator
Volunteer tester
Project scientist
Avatar

Send message
Joined: 17 Oct 05
Posts: 2394
ID: 1178
Credit: 18,681,406,695
RAC: 6,903,600

It is rather old (close to 4 years I am told) and it's got a HD 2400 or similar in it...I'm not even entirely sure it's OpenCL capable.

Must be a HD 4xxx card or higher for OpenCL...first ones in that series came out in Aug/Sept 2008.
____________
141941*2^4299438-1 is prime!

ID: 40358 |

Phroon
Send message
Joined: 27 Sep 10
Posts: 10
ID: 68760
Credit: 70,711,324
RAC: 0

Would it be possible to run the OpenCL app to crunch WUs as an anonymous platform if the app_info.xml lied to BOINC and told it that it was a CPU application?

Or is there better way to crunch work units with it until BOINC itself is fixed?

ID: 40450 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Would it be possible to run the OpenCL app to crunch WUs as an anonymous platform if the app_info.xml lied to BOINC and told it that it was a CPU application?

Probably yes, since BOINC can't run anything else on the GPU. Give it a try if you feel up to it!
____________
Twitter: IainBethune
Proud member of team "Aggie The Pew". Go Aggie!
3073428256125*2^1290000-1 is Prime!

ID: 40461 |

Phroon
Send message
Joined: 27 Sep 10
Posts: 10
ID: 68760
Credit: 70,711,324
RAC: 0

I can report validated work units using the OpenCL application as an anonymous platform:
http://www.primegrid.com/result.php?resultid=301978299

My hacked together app_info.xml only allows the OpenCL application to run and it tries to run as many instances as I have CPU cores (A quick client setting to only use 12% of my multi-processors fixed that). But it does do work.

WUs take under 2 hours to finish on my ATI Radeon HD 4850.

Many thanks! Lets hope that BOINC gets fixed so more people can OpenCL crunch with convenience.

Here's the cobbled together app_info.xml that I wouldn't recommend for anything serious for those who care to try. You'll have to play around with the ownership of the files so that BOINC doesn't complain, of course:

<app_info> <app> <name>pps_sr2sieve</name> </app> <file_info> <name>tpsieve-cl-boinc-x86_64-mac</name> <executable/> </file_info> <app_version> <app_name>pps_sr2sieve</app_name> <version_num>150</version_num> <file_ref> <file_name>tpsieve-cl-boinc-x86_64-mac</file_name> <main_program/> </file_ref> </app_version> </app_info>

ID: 40528 |

jpranke2011
Send message
Joined: 1 Jan 11
Posts: 7
ID: 79419
Credit: 16,167,917
RAC: 0

Looks like the new pre-release version of BOINC (6.13.6) can detect Mac ATI GPUs. Unfortunately still can't crunch due to the following message:

Thu Oct 6 08:13:13 2011 | PrimeGrid | Sending scheduler request: To fetch work.
Thu Oct 6 08:13:13 2011 | PrimeGrid | Requesting new tasks for ATI
Thu Oct 6 08:13:15 2011 | PrimeGrid | Scheduler request completed: got 0 new tasks
Thu Oct 6 08:13:15 2011 | PrimeGrid | Message from server: ATI GPU: Upgrade to the latest driver to process tasks using your computer's GPU
Thu Oct 6 08:13:15 2011 | PrimeGrid | No work sent
Thu Oct 6 08:13:15 2011 | PrimeGrid | Message from server: No work available for the applications you have selected. Please check your project preferences on the web site.
Thu Oct 6 08:13:15 2011 | PrimeGrid | This computer has reached a limit on tasks in progress

Maybe on the PrimeGrid side it's running a driver version check which is not applicable to Macs. Any way to debug this further?

ID: 41392 |

pabliedung

Volunteer tester
Send message
Joined: 2 Jan 11
Posts: 87
ID: 79535
Credit: 266,553,765
RAC: 33,258

I am getting the same error message on my iMac. The ATI GPU is now detected by BOINC 6.13.6 but I can not download any ATI wu's from PG. I have the ATI GPU option and PPS Sieve project (AMD/ATI) selected in my preferences.

Sun Oct 9 22:45:16 2011 | | Starting BOINC client version 6.13.6 for x86_64-apple-darwin
Sun Oct 9 22:45:16 2011 | | OS: Mac OS X 10.7.1 (Darwin 11.1.0)
Sun Oct 9 22:45:16 2011 | | ATI GPU is OpenCL-capable
Sun Oct 9 22:45:16 2011 | | Version change (6.13.1 -> 6.13.6)
...
Sun Oct 9 22:46:12 2011 | PrimeGrid | work fetch resumed by user
Sun Oct 9 22:46:13 2011 | PrimeGrid | update requested by user
Sun Oct 9 22:46:16 2011 | PrimeGrid | Sending scheduler request: Requested by user.
Sun Oct 9 22:46:16 2011 | PrimeGrid | Requesting new tasks for ATI
Sun Oct 9 22:46:18 2011 | PrimeGrid | Scheduler request completed: got 0 new tasks
Sun Oct 9 22:46:18 2011 | PrimeGrid | Message from server: ATI GPU: Upgrade to the latest driver to process tasks using your computer's GPU
Sun Oct 9 22:46:18 2011 | PrimeGrid | No work sent
Sun Oct 9 22:46:18 2011 | PrimeGrid | Message from server: No work available for the applications you have selected. Please check your project preferences on the web site.

Edit: PG now lists my ATI GPU under "computer information" as having 32768MB when it should be closer to 1024MB...
http://www.primegrid.com/show_host_detail.php?hostid=178104

ID: 41586 |

Iain Bethune

Honorary cruncher
Send message
Joined: 28 Jan 09
Posts: 1588
ID: 34775
Credit: 194,056,043
RAC: 0

Just to keep you all up-to-date on this, the issue had been discussed with the BOINC developers, and there are a few things to be done:

1) Updates to the client to support Mac OS X / OpenCL (in 6.13.x) - looks like this might be working or at least partly working

2) Some server-side updates to add a new plan class for Mac/OpenCL apps - don't know when Rytis plans to do this.

So there is no point doing further testing on PG/Boinc until at least these two steps have been completed.
____________
Twitter: IainBethune
Proud member of team "Aggie The Pew". Go Aggie!
3073428256125*2^1290000-1 is Prime!

ID: 41588 |

Phroon
Send message
Joined: 27 Sep 10
Posts: 10
ID: 68760
Credit: 70,711,324
RAC: 0

Just to keep you all up-to-date on this, the issue had been discussed with the BOINC developers, and there are a few things to be done:

1) Updates to the client to support Mac OS X / OpenCL (in 6.13.x) - looks like this might be working or at least partly working

2) Some server-side updates to add a new plan class for Mac/OpenCL apps - don't know when Rytis plans to do this.

So there is no point doing further testing on PG/Boinc until at least these two steps have been completed.

Looks like #1 has been completed, I've edited my custom app_info.xml to be the following and BOINC Manager now handles my ATI GPU as a GPU:

However, without that I still get the dreaded "Message from server: ATI GPU: Upgrade to the latest driver to process tasks using your computer's GPU" message.

Perhaps now that the client seems to be working it may be worthwhile to approach #2? Or it quite a large server-side problem?

ID: 57648 |

Brook Harste

Send message
Joined: 30 Nov 08
Posts: 1351
ID: 32401
Credit: 807,949,303
RAC: 73

I got it working in boinc with your app_info.xml file.
It's running right now.

Boinc Log

Wed Dec 19 18:41:47 2012 | PrimeGrid | Requesting new tasks for ATI Wed Dec 19 18:41:49 2012 | PrimeGrid | Scheduler request completed: got 1 new tasks Wed Dec 19 18:41:49 2012 | PrimeGrid | Message from server: ATI GPU: Upgrade to the latest driver to process tasks using your computer's GPU Wed Dec 19 18:41:51 2012 | PrimeGrid | File stat_primegrid.png exists already, skipping download Wed Dec 19 18:41:51 2012 | PrimeGrid | File primegrid_slideshow_00.png exists already, skipping download Wed Dec 19 18:41:51 2012 | PrimeGrid | Starting task pps_sr2sieve_55384454_0 using pps_sr2sieve version 150 in slot 0

Activity Monitor

56956 tpsieve-cl-boinc-x86_64-mac boinc_project 9.2 7 201.2 MB Intel (64 bit) 197.5 MB 9.7 MB

Edit: A little info on my hardware. I have the first power mac (1.1) with an ATI 5770. My first task is still running, so it looks like it will take 1 hour 20ish minutes per task.
____________

ID: 60630 |

Toggle Menu

Join PrimeGrid

Returning Participants

Community

Leader Boards

Results

Other