TPU's WCG/BOINC Team

[SparkyJJO]

Noticed HPF is still having computation errors in vista x64

Using BOINC 6.2.28, is there a different version that doesn't have this issue? I hate the units getting borked

[123bob]

Originally Posted by SparkyJJO

Noticed HPF is still having computation errors in vista x64

Using BOINC 6.2.28, is there a different version that doesn't have this issue? I hate the units getting borked

I could not find a version that HPF2 worked reliably on with Vista 64. As you know, I tried for a LONG while to make it work. Did everything I could think of and everything other folks suggested.

Team, Sparky is referring to a problem with Human Proteome Folding 2 WUs, running on Vista 64 bit. Occaisionally, the WUs will error out usually within the first few seconds of run time. If you run these on V64 bit, you can see errors show up in your "result status" using the filter to view errors. "Result status" is a useful tool to use to find out a lot of stuff. This is located on your WCG account page, on the left side.

Bob

[SparkyJJO]

Yeah some run fine, some don't. Sometimes I'll go a week with nary an error, then all of a sudden it'll have a hissy fit for a few days.

Drove me nuts when I first saw it, I thought my CPU was going unstable again on me

[Hicks]

Milestones Today
Damulta passes 1,000
123bob_TPU passes 200


Milestones Yesterday
TechPowerUp! passed 1,000,000
Chicken Patty passed 70,000
3xploit passed 2,000
Marineborn passed 5,000
123bob_TPU passed 100
paulieg passed 500

Boinc Total Credit
1,062,322
WCG Total Credit
7,628,321



Congrats guys!!

[oily_17]

Hi guys was just checking my Results Status and see I have some results with an "Error" tagged to them.

This is the error

 	
Result Log 	

<core_client_version>6.2.28</core_client_version>
<![CDATA[
<message>
Incorrect function. (0x1) - exit code 1 (0x1)
</message>
<stderr_txt>
ERROR:: Exit at: .\dock_structure.cc line:401

</stderr_txt>
]]>

Any ideas on what is causing this...maybe not as stable as I thought ??

I also see this, in the log for the pc that is giving the errors, it says the output file is absent for some tasks.

10/04/2009 14:53:27|World Community Grid|Restarting task mi947_00060_15 using hpf2 version 603
10/04/2009 14:53:27|World Community Grid|Restarting task mi985_00042_7 using hpf2 version 603
10/04/2009 14:54:31|World Community Grid|Computation for task mi947_00060_15 finished
10/04/2009 14:54:31|World Community Grid|Output file mi947_00060_15_0 for task mi947_00060_15 absent
10/04/2009 14:54:31|World Community Grid|Starting mi986_00014_6
10/04/2009 14:54:31|World Community Grid|Starting task mi986_00014_6 using hpf2 version 603
10/04/2009 14:56:27|World Community Grid|Computation for task mi986_00014_6 finished
10/04/2009 14:56:27|World Community Grid|Output file mi986_00014_6_0 for task mi986_00014_6 absent

This is the client version, processor etc -

10/04/2009 14:51:26||Starting BOINC client version 6.2.28 for windows_intelx86
10/04/2009 14:51:26||log flags: task, file_xfer, sched_ops
10/04/2009 14:51:26||Libraries: libcurl/7.19.0 OpenSSL/0.9.8i zlib/1.2.3
10/04/2009 14:51:26||Data directory: C:\ProgramData\BOINC
10/04/2009 14:51:26||Processor: 2 AuthenticAMD Dual Core AMD Opteron(tm) Processor 165 [x86 Family 15 Model 35 Stepping 2]
10/04/2009 14:51:26||Processor features: fpu tsc pae nx sse sse2 pni 3dnow mmx
10/04/2009 14:51:26||OS: Microsoft Windows Vista: Ultimate x86 Editon, Service Pack 1, (06.00.6001.00)
10/04/2009 14:51:26||Memory: 2.00 GB physical, 7.93 GB virtual
10/04/2009 14:51:26||Disk: 88.79 GB total, 37.49 GB free

[Hicks]

Whats your OC? My 1st guess is ram(most of the time it is).

[alucasa]

If it's just one error, it's nothing to be alarmed about. It's due to computing errors which happen occasionally.

If you are receiving massive amount of errors, then your OC is unstable. If it wasn't an OCed rig, it'd be hardware incompatibility.

Personally, I take up to 3 errors a day per a machine normal behavior.

[oily_17]

Originally Posted by Hicks

Whats your OC? My 1st guess is ram(most of the time it is).

It's running at 2.9 and the ram is underclocked....I just recently replaced it with some DDR500, but you may be right it, I hope not!!

Originally Posted by alucasa

If it's just one error, it's nothing to be alarmed about. It's due to computing errors which happen occasionally.

If you are receiving massive amount of errors, then your OC is unstable. If it wasn't an OCed rig, it'd be hardware incompatibility.

Personally, I take up to 3 errors a day per a machine normal behavior.

There are about 7 errors in 4 days of running.They all seem to be with the Human Proteome Folding tasks, my other tasks seem to be passing OK.

This machine was running the Folding@home SMP client, before I shifted to WCG, and it run with out any problems on it.

[Hicks]

Originally Posted by alucasa

If it's just one error, it's nothing to be alarmed about. It's due to computing errors which happen occasionally.

If you are receiving massive amount of errors, then your OC is unstable. If it wasn't an OCed rig, it'd be hardware incompatibility.

Personally, I take up to 3 errors a day per a machine normal behavior.

What the rack freak said....

[Hicks]

Your not running Vista x64 by chance?

[oily_17]

Not on this machine but I am on the other one I added yesterday.

[Hicks]

HPF & Vista x64 dont get along too well. My advice is to not do those WU's on the x64 rig.

[alucasa]

Originally Posted by oily_17

There are about 7 errors in 4 days of running.They all seem to be with the Human Proteome Folding tasks, my other tasks seem to be passing OK.

This machine was running the Folding@home SMP client, before I shifted to WCG, and it run with out any problems on it.

7 errors in 4 days, sounds acceptable to me. Folding uses GPU. WCG uses CPU. It's different worlds, really... Though their goals are the same. Sometimes, errors are caused by the tasks themselves. WCG forum often announces if there are serious errors in their task units.

[oily_17]

Originally Posted by Hicks

HPF & Vista x64 dont get along too well. My advice is to not do those WU's on the x64 rig.

Thanks I did not know that

Originally Posted by alucasa

7 errors in 4 days, sounds acceptable to me. Folding uses GPU. WCG uses CPU. It's different worlds, really... Though their goals are the same.

The SMP client uses the CPU to fold although I dont know if they are the same tech, but it does stress the CPU alot when folding with it.

[alucasa]

If I remember correctly, in Windows XP, the SMP client used to use a fairly large amount of a core's computing powers.
But, in Vista, I was told that it uses almost zero cpu powers if set up correctly, so that should not be an issue.

What I do know is I do get a few errors even on rigs that I run stock setups.

I don't OC WCGing rigs anymore due to having had a nightmare with power bills once when I was really addicted to WCGing.

[oily_17]

Originally Posted by alucasa

If I remember correctly, in Windows XP, the SMP client used to use a fairly large amount of a core's computing powers.
But, in Vista, I was told that it uses almost zero cpu powers if set up correctly, so that should not be an issue.

That is for the GPU2 client which uses your GPU to fold but the SMP client uses the CPU to fold and uses 100% of the core that it is assigned to.
I think you are mixing the two up But I have run both types of client on this machine at the same time without problems.

What I do know is I do get a few errors even on rigs that I run stock setups.

I don't OC WCGing rigs anymore due to having had a nightmare with power bills once when I was really addicted to WCGing.

I will keep an eye on it I see you can set each machine to use different tasks on the WCG site..I will do that later

[alucasa]

Originally Posted by oily_17

That is for the GPU2 client which uses your GPU to fold but the SMP client uses the CPU to fold and uses 100% of the core that it is assigned to.
I think you are mixing the two up But I have run both types of client on this machine at the same time without problems.

I suppose I am. For one thing, I don't really know GPU folding because I barely buy any video cards to fold with. My passion doesn't lie with video cards apparently.

I am for low-voltage computing where performance per watt matters.

Originally Posted by oily_17

I will keep an eye on it I see you can set each machine to use different tasks on the WCG site..I will do that later

Yes..., do watch your power bill. When I had 8 rigs running... along with AMD quad-FX 4x4 rig..., my power bill... doubled ...

[123bob]

@oily

This is what Sparky and I were talking about in posts 427-429 in regards to HPF2 and Vista.

IIRC, this has been known to affect 32 bit Vista machines also. Here is a thread from the WCG forum that goes into it. We never did come up with a full solution. It seems to be "hit-or-miss" for a fix.

If you don't like the errors, you can un-check the HPF2 project on your profile page at WCG. This will stop you from getting HPF2 WUs.

Regards,
Bob

[oily_17]

Originally Posted by alucasa

I suppose I am. For one thing, I don't really know GPU folding because I barely buy any video cards to fold with. My passion doesn't lie with video cards apparently.

I am for low-voltage computing where performance per watt matters.



Yes..., do watch your power bill. When I had 8 rigs running... along with AMD quad-FX 4x4 rig..., my power bill... doubled ...

No probs...and my power bill has risen sharply with folding on three rigs with GPU's..the low voltage performance may be the way to go.

Originally Posted by 123bob

@oily

This is what Sparky and I were talking about in posts 427-429 in regards to HPF2 and Vista.

IIRC, this has been known to affect 32 bit Vista machines also. Here is a thread from the WCG forum that goes into it. We never did come up with a full solution. It seems to be "hit-or-miss" for a fix.

If you don't like the errors, you can un-check the HPF2 project on your profile page at WCG. This will stop you from getting HPF2 WUs.

Regards,
Bob

Thanks for the link Bob....I was just looking at the site and may change my projects on these couple of machines.

[SparkyJJO]

Originally Posted by alucasa

If I remember correctly, in Windows XP, the SMP client used to use a fairly large amount of a core's computing powers.
But, in Vista, I was told that it uses almost zero cpu powers if set up correctly, so that should not be an issue.

What I do know is I do get a few errors even on rigs that I run stock setups.

I don't OC WCGing rigs anymore due to having had a nightmare with power bills once when I was really addicted to WCGing.

That was GPU. And the CPU usage with GPU in XP is fixed since driver 181.20 which is awesome

[alucasa]

God, we need a dedicated forum, but anyway -

I've been looking at Casetronic C287. It's a new case from them. I've been thinking of stacking 3 or 4 of them together with Intel Penryn or even Atom 330 for the kicks.

Casetronic C287 http://www.casetronic.com/product_d.php?id=0000000057

I've had to reduce my farm to 2 machines from 6 but I'm keeping them on TPU's team indefinitely. It's only about 20k WCG points per day, but it should help.

I hope I will be able to finally get a new machine up with a Powercolor 4850 up and running, but I don't know about doing any folding. I shut down the other machines because my electric bill was over $300/mo even with oil heat. I'm obsessive-compulsive but I'm not stupid - although since I'd been doing it for over a year I guess that's debatable. Anyway, I'll have to see how much the 4850 draws at load and if it interferes with video operations.

Plus, I might give some of my "old" machines to friends, in which case I'll insist they run boinc - on my account of course - hehehe.

[Marineborn]

hey guys, stopped folding for awhile, base got hit by a tornado last night, she got pretty beat up, lost power, lots of trees down, tree almost crushed my car, it was strong enough to snap millitary grade gates. hahahah BUT IM BACK UP AND FOLDING NOW, sorry for the pause in folding

[Chicken Patty]

Originally Posted by Marineborn

hey guys, stopped folding for awhile, base got hit by a tornado last night, she got pretty beat up, lost power, lots of trees down, tree almost crushed my car, it was strong enough to snap millitary grade gates. hahahah BUT IM BACK UP AND FOLDING NOW, sorry for the pause in folding

folding or crunching?

anyhow, sorry to hear bout that, good everything came out ok and that you are ok

[Marineborn]

crunching sorry...dont think my laptops would fold very well HAHAHAHA. and thanks chicken...its was crazy