The machine runs SETI ATI work without any errors,
(i am unable to get work from them atm)
the machine runs win 7 home 64 with ATI 7970 GPUs ccc 12:4, running one task per GPU
I have no idea what has caused the new APP to fail so badly.
Well. here i am back again,
i don`t give up.
The problem is gone/fixed in an unusual way.
The computer is all the same other than a swap of cpu.
From what i can see the problem was that the 1:28 app considers a 3.6ghz P4 with HT 64 bit (prescot 660) that is doing no other work but feed the two 7970 GPUs with one work unit each to be so patheticaly slow that it can not wait for it to do it`s bit of the job and errors the task at hand!!!
Using process explorer to look at any other programs running there is more than 90% cpu time free for boinc to use.
Uninstaled every program possible, just the minimal win7 os, no anti virus, or anything that could get in the way.
I found this fix by finaly swaping the Q6600 from another PC of mine and when run out of SETI work again giving Einstein a spin to see what happened.
No errors was the result and all work units compleated in less than 30 minutes.
I now have boinc set to only run cpu tasks on two cpu cores so as to get the best possible out of the gpu`s.
I may get around to setting up a venue for this machine so it can run more than one task at a time,
for now i am glad the problem of errored tasks is gone.
The P4 is now in my other long term Einstein crunching machine and is doing very nicely,
The Q6600 was doing fore cpu jobs and feeding two low speck gpu`s,
the odd thing is that now the P4 is in that other pc and only running one cpu task the HT part for feeding the gpu`s (long story short) the task run times are a little lower.
I accept that a P4, even a skt 775 one, is not fast by todays standards and was going to be a bottle neck in the system, my seti throughput has increased as well, to be the source of the problem does seem a bit odd.
I have no idea what the difference is between the 1:24 app and the 1:28 app as far as its demand for cpu time is other than it did my head in for a while.
I tried to do some tasks
)
I tried to do some tasks today but they errored out after 20-30 seconds. Here's my message log.
9/12/2012 5:15:35 PM | Einstein@Home | Output file p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1_0 for task p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1 absent
9/12/2012 5:15:35 PM | Einstein@Home | Output file p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1_1 for task p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1 absent
9/12/2012 5:15:35 PM | Einstein@Home | Output file p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1_2 for task p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1 absent
9/12/2012 5:15:35 PM | Einstein@Home | Output file p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1_3 for task p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1 absent
9/12/2012 5:15:35 PM | Einstein@Home | Output file p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1_4 for task p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1 absent
9/12/2012 5:15:35 PM | Einstein@Home | Output file p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1_5 for task p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1 absent
9/12/2012 5:15:35 PM | Einstein@Home | Output file p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1_6 for task p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1 absent
9/12/2012 5:15:35 PM | Einstein@Home | Output file p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1_7 for task p2030.20110123.G176.20-02.80.N.b6s0g0.00000_496_1 absent
Here's the Stderr file.
Stderr output
7.0.31
There was an error while deleting the color transform. (0x7e3) - exit code 2019 (0x7e3)
Activated exception handling...
[17:20:24][3084][INFO ] Starting data processing...
[17:20:25][3084][INFO ] Using OpenCL platform provided by: Advanced Micro Devices, Inc.
[17:20:25][3084][INFO ] Using OpenCL device "Cypress" by: Advanced Micro Devices, Inc.
[17:20:26][3084][INFO ] Checkpoint file unavailable: status.cpt (No such file or directory).
------> Starting from scratch...
[17:20:26][3084][INFO ] Header contents:
------> Original WAPP file: ./p2030.20110123.G176.20-02.80.N.b6s0g0.00000_DM350.40
------> Sample time in microseconds: 65.4762
------> Observation time in seconds: 274.62705
------> Time stamp (MJD): 55585.080850181192
------> Number of samples/record: 0
------> Center freq in MHz: 1214.289551
------> Channel band in MHz: 0.33605957
------> Number of channels/record: 960
------> Nifs: 1
------> RA (J2000): 52502.0467
------> DEC (J2000): 303639.793598
------> Galactic l: 0
------> Galactic b: 0
------> Name: G176.20-02.80.N
------> Lagformat: 0
------> Sum: 1
------> Level: 3
------> AZ at start: 0
------> ZA at start: 0
------> AST at start: 0
------> LST at start: 0
------> Project ID: --
------> Observers: --
------> File size (bytes): 0
------> Data size (bytes): 0
------> Number of samples: 4194304
------> Trial dispersion measure: 350.4 cm^-3 pc
------> Scale factor: 0.00173531
[17:20:29][3084][INFO ] Seed for random number generator is 1165762014.
[17:20:38][3084][INFO ] Derived global search parameters:
------> f_A probability = 0.08
------> single bin prob(P_noise > P_thr) = 1.32531e-008
------> thr1 = 18.139
------> thr2 = 21.241
------> thr4 = 26.2686
------> thr8 = 34.6478
------> thr16 = 48.9581
[17:20:43][3084][ERROR] Error during OpenCL kernel setup: PS_R3 (error: -5)
[17:20:43][3084][ERROR] Demodulation failed (error: 2019)!
17:20:43 (3084): called boinc_finish
]]>
Any idea what's going on?
RE: With the previous ATI
)
Well. here i am back again,
i don`t give up.
The problem is gone/fixed in an unusual way.
The computer is all the same other than a swap of cpu.
From what i can see the problem was that the 1:28 app considers a 3.6ghz P4 with HT 64 bit (prescot 660) that is doing no other work but feed the two 7970 GPUs with one work unit each to be so patheticaly slow that it can not wait for it to do it`s bit of the job and errors the task at hand!!!
Using process explorer to look at any other programs running there is more than 90% cpu time free for boinc to use.
Uninstaled every program possible, just the minimal win7 os, no anti virus, or anything that could get in the way.
I found this fix by finaly swaping the Q6600 from another PC of mine and when run out of SETI work again giving Einstein a spin to see what happened.
No errors was the result and all work units compleated in less than 30 minutes.
I now have boinc set to only run cpu tasks on two cpu cores so as to get the best possible out of the gpu`s.
I may get around to setting up a venue for this machine so it can run more than one task at a time,
for now i am glad the problem of errored tasks is gone.
The P4 is now in my other long term Einstein crunching machine and is doing very nicely,
The Q6600 was doing fore cpu jobs and feeding two low speck gpu`s,
the odd thing is that now the P4 is in that other pc and only running one cpu task the HT part for feeding the gpu`s (long story short) the task run times are a little lower.
I accept that a P4, even a skt 775 one, is not fast by todays standards and was going to be a bottle neck in the system, my seti throughput has increased as well, to be the source of the problem does seem a bit odd.
I have no idea what the difference is between the 1:24 app and the 1:28 app as far as its demand for cpu time is other than it did my head in for a while.