Kokomiko: looks like your task hangs due to some strange reason. Did you restart BOINC? (not just the manager)

@CPU usage: since on my machine I have both, a few seconds or several minutes, I can clearly correlate the times to interactive usage.

@lower performance: I can clearly say it's not the firewall, since I have it turned off and have no other ones running, except my hardware router.

NaRyan wrote:

Looking at my Boincview logfiles, there used to be just under a 12 hour gap between workunits, now it has gone to 14.5 hours between workunits :(

That's the same magnitude of slow-down which I am seeing on XP32.

MrS

---

Scanning for our furry friends since Jan 2002

The new Windows-application shows since hours the same percentage and the time to complete is rising. 2 hours ago they show 97,294% and a completion time 4:08, now 97,294% and a completion time of 4:25. The shown CPU-time is 2:30:38, the taskmanager shows 3% for the core. Looks like there is no progress. The real total time is running since 8:44 UTC, now it's 17:34 UTC, that's far too long for my GTX280.

Same problem here, my last 3 work units are slower than normal.

Time per step: 32.684 ms
Approximate elapsed time for entire WU: 27781.558 s

Time per step: 44.760 ms
Approximate elapsed time for entire WU: 38046.248 s

Time per step: 33.262 ms
Approximate elapsed time for entire WU: 28272.461 s

My average used to be 26,000s plus or minus 300.

Kokomiko: looks like your task hangs due to some strange reason. Did you restart BOINC? (not just the manager)

Yes, have made a restart. Now the running time is 2:37:29, the percentage is furthermore 97,294%, the time to complete is now 4:36. I will wait one hour more. Should I make a reset on the project? With this behavior I get more credits on MilkyWay for this idle core :D

---

I'd either

1. abort the current WU and see how the next one does

or

2. switch the project to "no new work", abort the current WUs, update the project and then reset it

I'd prefer the 1st option, but if it's OK with your daily quota the 2nd is the more secure one.

MrS

---

Scanning for our furry friends since Jan 2002

I'm don't the only cruncher with problems with this WU, don't know why.

http://www.gpugrid.net/workunit.php?wuid=42214

Had abort the task, the next is running fine. Will have a eye on it, hope, this is running better.

12 hours for the birds ... :(

---

12 hours for the birds ... :(

Nevermind, it's beta after all! The devs will surely notice this WU and may find out something useful.

MrS

---

Scanning for our furry friends since Jan 2002

... and the next WU is broken. Looks like the GTX280 don't like the 6.45 ... :(

WU broken again

... and again I'm not alone with a error on this WU ...

---

Application 6.45 is way to slow on gtx280.

Gone from 7 hours a wu to 9 hours :(

... and the next WU is broken. Looks like the GTX280 don't like the 6.45 ... :(

WU broken again

... and again I'm not alone with a error on this WU ...

Since last sunday most of my WUs are brolen. Only 1 WU ist finished succesfull. The sizuation ist awfully, I become desperate.

Here are my tasks:
http://www.ps3grid.net/results.php?userid=5402

I gone back with the ddriber from 177.98 to 177.84 and U detached the PS3GGRID in Boinc, Whatcan I do else?
I have a Gainward 9800GTX+ without OC.


Some of the errors:
<core_client_version>6.3.10</core_client_version>
<![CDATA[
<message>
Unzul�ssige Funktion. (0x1) - exit code 1 (0x1)
</message>
<stderr_txt>
# Using CUDA device 0
# Device 0: "GeForce 9800 GTX/9800 GTX+"
# Clock rate: 1836000 kilohertz
MDIO ERROR: cannot open file "restart.coor"
Cuda error: Kernel [frc_sum_kernel_angle] failed in file 'force.cu' in line 539 : unknown error.

</stderr_txt>
]]>

Or this one:

<core_client_version>6.3.10</core_client_version>
<![CDATA[
<message>
Unzul�ssige Funktion. (0x1) - exit code 1 (0x1)
</message>
<stderr_txt>
# Using CUDA device 0
# Device 0: "GeForce 9800 GTX/9800 GTX+"
# Clock rate: 1836000 kilohertz
MDIO ERROR: cannot open file "restart.coor"

</stderr_txt>
]]>


Or another error:

<core_client_version>6.3.10</core_client_version>
<![CDATA[
<message>
Unzul�ssige Funktion. (0x1) - exit code 1 (0x1)
</message>
<stderr_txt>
# Using CUDA device 0
# Device 0: "GeForce 9800 GTX/9800 GTX+"
# Clock rate: 1836000 kilohertz
MDIO ERROR: cannot open file "restart.coor"
Cuda error: Kernel [frc_sum_kernel_bond] failed in file 'force.cu' in line 553 : unknown error.

</stderr_txt>
]]>

... next broken WU on a GTX280:

Logfile (10:00 MESZ = 8:00 UTC):

17.09.2008 10:01:09|PS3GRID|Computation for task pY11683-GPUTEST3-0-10-acemd_0 finished
17.09.2008 10:01:09|PS3GRID|Output file pY11683-GPUTEST3-0-10-acemd_0_1 for task pY11683-GPUTEST3-0-10-acemd_0 absent
17.09.2008 10:01:09|PS3GRID|Output file pY11683-GPUTEST3-0-10-acemd_0_2 for task pY11683-GPUTEST3-0-10-acemd_0 absent
17.09.2008 10:01:09|PS3GRID|Output file pY11683-GPUTEST3-0-10-acemd_0_3 for task pY11683-GPUTEST3-0-10-acemd_0 absent
17.09.2008 10:01:11|PS3GRID|Started upload of pY11683-GPUTEST3-0-10-acemd_0_0
17.09.2008 10:01:15|PS3GRID|Finished upload of pY11683-GPUTEST3-0-10-acemd_0_0


Link to WU

<core_client_version>6.3.10</core_client_version>
<![CDATA[
<message>
Unzul�ssige Funktion. (0x1) - exit code 1 (0x1)
</message>
<stderr_txt>
# Using CUDA device 0
# Device 0: "GeForce GTX 280"
# Clock rate: 1296000 kilohertz
MDIO ERROR: cannot open file "restart.coor"
Cuda error: Kernel [frc_sum_kernel_impr] failed in file 'force.cu' in line 583 : unknown error.

</stderr_txt>
]]>

Same problem: File restart.coor is missing.

Remark: Since the 6.45 is running, the PC has under Vista 64 bit make at 3 times a reboot without any hint in the logfile.

btw: The other machine (also Vista 64bit) with the 8800GT is running without problems.

---

... next broken WU on a GTX280:

<core_client_version>6.3.10</core_client_version>
<![CDATA[
<message>
Unzul�ssige Funktion. (0x1) - exit code 1 (0x1)
</message>
<stderr_txt>
# Using CUDA device 0
# Device 0: "GeForce GTX 280"

In all my broken WUs I have the reason "Unzulässige Funktion" or in english "unsopported function" and I think it is a problem in the 6.45 code. Coulf you locate zhe error im your code?

These result from 6.45 still confuse me

Work #1 177.84 8800GT 600MHz-512MB P4-HT 3.0GHz 2GB Windows XP SP3
TskID CPUs GPUsec ms/Step
58017 285s 68463s 80.545
58579 203s Error
59231 623s 68343s 80.404

Work #2 177.84 8800GT 600MHz-512MB P4-HT 3.0GHz 2GB Windows XP SP3
TskID CPU-s GPUsec ms/Step
58264 3456s 61631s 72.508ms
59024 3540s 61432s 72.274ms
59183 3513s 61384s 72.217ms

Home #1 177.92 8800GT 640MHz-512MB Intel P4-HT 3.8GHz 2GB Windows XP SP3 Media Center
TskID CPU-s GPUsec ms/Step
58254 3723s 58289s 68.576ms
58428 3436s 59270s 69.730ms
58997 3545s 59464s 69.958ms
59533 3703s 60286s 70.925ms
Home #1 is faster on GPU becuase it is a slightly faster clock on the GPU, factory overclocked model. All three are same brand. This matches previous apps, in that it was faster by about 2000s. Which it still is compared to work #2.


Home #1 still has similar CPU times to work #2. I cannot find any software running that would be common to home #1 and work #2. Most is common to all three or common between work #1 and work #2.

About the only common element between home #1 and work #2 is MS SQL Server Service Manager, I believe this is what they call SQLExpress. Home is Windows Media Center Edition, It came installed with that, I think it would only be in use when running media center, which I don't on this system. Work #2 has one piece of software installed, but not in use which installed it there. I do not know what else really uses it, so basically it sits dormant on both. I do not really see how that would interfere.

Now looking at the times, why (?) on work #1 is the CPU faster and the GPU slower than work # 2 ? These two computers have identical hardware, CPU speed, GPU's, daily usage. Basically I use those 8 hours a day, then they are untouched for 16 hours. At home this computer gets maybe 5 hours max useage. So if using the GPU for video slows down the cpu and app, why would this not be shown across all hosts. Certainly 1 task would be running while I'm using host, but then 1 would be able to run undisturbed during the time I'm not using it, so as if suggested by some some use of the gpu by other programs is causing the slowdown, then each host would have fast and slow cpu times.

Work #1 and #2 get about the same usage, They will be unused with screens on standby for 16 hours, then for about 8 hours both are awake. I would use work #1 slightly more as that is my primary. I do email on it, document scanning and some of our daily operations and invoicing. It has the printers attached to it. Work #2 I primarily do invoicing for shipments, while I read email on #1 and on #3 process the shipment in UPS software. Work #2 also is used for some internet browsing while on #1 I would be entering information viewed on #2. Yet work #1 shows the faster CPU times over work #2.

Do you have any screensaver activated or simply a "blank screen" ?

Are the cards "really" identical ? SP (shaders), frequency matters more than the GPU frequency for GPUgrid.
You can check that with GPU-Z .

edit: spelling

...
Same problem: File restart.coor is missing.
...

That's not a problem, restart.coor is the checkpoint file. It only means there was no checkpoint file because it ran from the beginning until it errored out.

But please don't ask me what causes the other errors... ;)

---

pixelicious.at - my little photoblog

Do you have any screensaver activated or simply a "blank screen" ?

Are the cards "really" identical ? SP (shaders), frequency matters more than the GPU frequency for GPUgrid.
You can check that with GPU-Z .

edit: spelling

No screen saver (none), windows powers down the monitor after 15 minutes of non-use.

Yes work #1 and work #2 are identical, bought at same time, same brand, same frequency 600MHz, 512MB at 900MHz and stock settings, no tinkering or overclocking, same driver installed from same downlaod. Yes they have same number of shaders, 112 at 1500MHz. Same everything. The CPUs are the same, same frequency and memory in each host. Same ahrd drives. Only some small software differences.

Yes I already have GPUz, Just ran again, all info reported on both is identical down to the bios verison. I do not have a way at the moment to upload pictures (screenshots). I work on that and psot those later.

In all my broken WUs I have the reason "Unzulässige Funktion" or in english "unsopported function"

Wolfram, did you reboot your machine? Maybe also tried a project reset, in case some file on your hdd became corrupted.

And btw, "Unzulässige Funktion" should be more like "invalid function" ;)

MrS

---

Scanning for our furry friends since Jan 2002

@Keith: your situation puzzles me as well.

Your data may suggest a correlation: high cpu usage - lower computation time. I checked with my WUs: 3 of them have 62.5 ms/step and CPU usages of 14, 14 and 291s. One has CPU 851 and "only" 60.1 ms/step. That's not really a clear-cut picture either.

MrS

---

Scanning for our furry friends since Jan 2002

Looks like I have solved the problem for my machine, don't know, if it is relevant for others.

I use a Gigabyte MA790FX-DS5 board, a Phenom 9950 BE with 4 x 2 GB OCZ RAM 1066 and a GTX280 from XFX. My RAM was running in DUAL and UNGANGED mode. Since I switched to GANGED mode, the PC is still running without crashing the PS3Grid-WU with the 6.45. Maybe there is a problem with the 2 x 64 bit RAM access simultaneous for different cores in the UNGANGED mode, so I use now the 1 x 128 bit access of the GANGED mode.

---

In all my broken WUs I have the reason "Unzulässige Funktion" or in english "unsopported function"

Wolfram, did you reboot your machine? Maybe also tried a project reset, in case some file on your hdd became corrupted.

And btw, "Unzulässige Funktion" should be more like "invalid function" ;)

MrS

Thank you for your help. Yes I rebooted and reset the project. I had 1 option in the cc_config.xml file. I deleted this file now. rebooted and started a new WU nearly 3 hours ago and has now 23,7% finished. I hope, I have a good night.

Thx again for your help.

Looks like I have solved the problem for my machine... Since I switched to GANGED mode, the PC is still running without crashing the PS3Grid-WU with the 6.45

Although it's possible it seems very random. I would wait a few more WUs to see if it's stable. Then switch back to unganged and I wouldn't be surprised if it was stable as well ;)

MrS

---

Scanning for our furry friends since Jan 2002