V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Moderators: Site Moderators, PandeGroup

Re: V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Postby Fallen-Angel » Sun Nov 20, 2011 1:13 pm

I just made some benchmarks and found out that the results are pretty the same as they were a few weeks ago. That means, in my case it's not a hardware issue.
My Computer also has finished both 0x11 WUs without problems this night. Now I try to set one GPU to advanced, in order to see what happens. Afterwards, I update the driver to 11.11 and try again with 0x11 and 0x16 FahCores. If they fail again, I roll back to 11.8 and check the results again. The thing is, I'm sure I wont finish the tests today, and on Monday I work till 10pm so I will have some results on Thursday.

Best,

FA
Image
Image
User avatar
Fallen-Angel
 
Posts: 35
Joined: Sat Sep 24, 2011 11:41 am
Location: Germany

Re: V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Postby Eagle » Sun Nov 20, 2011 11:03 pm

Even updated to 11.11, but still no luck:
Code: Select all
23:01:48:FahCore 0x16 started
23:01:48:Unit 02:
23:01:48:Unit 02:*------------------------------*
23:01:48:Unit 02:Folding@Home GPU Core
23:01:48:Unit 02:Version 2.11 (Thu Dec 9 15:00:14 PST 2010)
23:01:48:Unit 02:
23:01:48:Unit 02:Compiler  : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 15.00.30729.01 for 80x86
23:01:48:Unit 02:Build host: user-f6d030f24f
23:01:48:Unit 02:Board Type: AMD/OpenCL
23:01:48:Unit 02:Core      : x=16
23:01:48:Unit 02: Window's signal control handler registered.
23:01:48:Unit 02:Preparing to commence simulation
23:01:48:Unit 02:- Ensuring status. Please wait.
23:01:58:Unit 02:- Looking at optimizations...
23:01:58:Unit 02:- Working with standard loops on this execution.
23:01:58:Unit 02:- Previous termination of core was improper.
23:01:58:Unit 02:- Files status OK
23:01:58:Unit 02:sizeof(CORE_PACKET_HDR) = 512 file=<>
23:01:58:Unit 02:- Expanded 44935 -> 169787 (decompressed 377.8 percent)
23:01:58:Unit 02:Called DecompressByteArray: compressed_data_size=44935 data_size=169787, decompressed_data_size=169787 diff=0
23:01:58:Unit 02:- Digital signature verified
23:01:58:Unit 02:
23:01:58:Unit 02:Project: 11294 (Run 15, Clone 24, Gen 116)
23:01:58:Unit 02:
23:01:58:Unit 02:Entering M.D.
23:02:00:Unit 02:Tpr hash 02/wudata_01.tpr:  2516529971 3488723389 3998212329 2758843288 2621262205
23:02:00:Unit 02:Working on PEPTIDE (1-42)
23:02:00:Unit 02:Client config unavailable.
23:02:02:FahCore, running Unit 02, returned: UNKNOWN_ENUM (-1073740777 = 0xc0000417)

And the "client config unavailable"-line is there, too.. :(
Michael Jordan: “I can accept failure — But I can’t accept not trying.”
Image
User avatar
Eagle
 
Posts: 157
Joined: Sun Feb 17, 2008 1:06 am
Location: » Earth » Europe » Germany

Re: V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Postby Fallen-Angel » Mon Nov 21, 2011 5:57 am

Do you have only one GPU?
This night I folded 2 advanced WUs with my first GPU and one normal WU on SMP while the second GPU was paused.
Now I go to work and let the second GPU run 0x11 FahCores, if both GPUs success, I try 0x16 FahCore WUs on both GPUs (advanced)

I also had the line with "unavailable client info".... did you try to reinstall the client?
I did it in order to get the stock setting for the whole client.
User avatar
Fallen-Angel
 
Posts: 35
Joined: Sat Sep 24, 2011 11:41 am
Location: Germany

Re: V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Postby Eagle » Mon Nov 21, 2011 6:52 am

Yes, there's only one GPU within the system. No, I didn't re-install. Might try that - seems like it worked on your end?!
User avatar
Eagle
 
Posts: 157
Joined: Sun Feb 17, 2008 1:06 am
Location: » Earth » Europe » Germany

Re: V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Postby Fallen-Angel » Mon Nov 21, 2011 8:08 pm

Yeah, for now. I have the Catalyst v11.10 driver installed and had no problems to fold normal WUs on SMP, the advanced (core 0x16) WUs on the first and the normal ones on the second GPU. Now I try both GPUs to fold advanced WUs at the same time.
User avatar
Fallen-Angel
 
Posts: 35
Joined: Sat Sep 24, 2011 11:41 am
Location: Germany

Re: V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Postby Eagle » Mon Nov 21, 2011 9:00 pm

Interesting.. :)
User avatar
Eagle
 
Posts: 157
Joined: Sun Feb 17, 2008 1:06 am
Location: » Earth » Europe » Germany

Re: V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Postby Fallen-Angel » Mon Nov 21, 2011 10:03 pm

@Eagle, tomorrow, try to delete your client (including the data files) and delete the whole catalyst center, than in the hardware manager the whole GPU driver. Then reinstall the Catalyst v11.11 ( or some else if you want to) and install the FAH client. Leave all settings on stock. So we will know how your system or better to say, your GPU handles the FahCore 0x11 WUs. I think, if any error appears, it's the driver fault. If not, try to set the advanced flag only on the GPU and run it further (for FahCore 0x16 WUs). If it fails, then we know that the driver is not compatible or has problems with the core 0x16 WUs. Then I think it's time the FAH team contacts the ATI (AMD) Team. I other case, when everything is fine, congratulation, you may fold further ^^

I go sleep now, so the next update is tomorrow ^^

@bruce
I've read the whole topic again and just wanted to mention, that if it were the unstable over-clocking (on the GPU), the system freezes or the blue screen appears, before the client responds (throw a line message which describes the error). That happened in my case when I oced the GPU too far or just too bad. When ocing the CPU, it may happen different things. Here I expect the client's reaction first, followed by a blue screen.

P.S.: I found out that in my case it may be a memory error which causes the fail. It also says that reinstalling the driver may help.... so hope I!
C0000005

Windows and F@H GUI Specific

Code: Select all
CoreStatus = C0000005 (-1073741819)
Client-core communications error: ERROR 0xc0000005
Deleting current work unit & continuing...


This is a known Windows memory error, while running the v5.x GUI client with the GUI open while finishing and uploading a work unit. Workarounds include updating the video driver (doesn't always help), keeping the GUI closed near the end of a work unit, or switching to the console client and using a 3rd party utility to see the pretty pictures and monitor the client's progress.

It can also be caused by faulty memory or a bad memory controller, so you should consider both possibilities.


P.P.S.: @Eagle, I also googled your error code (-1073740777 = 0xc0000417) further and found out that this error appears by many users when they try to compile a C++ (or sometimes even aq PHP) program (source code). [I just can't go sleep without doing anything :D That topic is too interesting!] The clue is, that nobody knows what causes it. I would try the steps above first. When it doesn't help, we mus go further and try different settings for different clients and drivers ... a lot of work ;)
Last edited by Fallen-Angel on Mon Nov 21, 2011 10:47 pm, edited 1 time in total.
User avatar
Fallen-Angel
 
Posts: 35
Joined: Sat Sep 24, 2011 11:41 am
Location: Germany

Re: V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Postby Eagle » Mon Nov 21, 2011 10:42 pm

Unfortunately, I can't do that until Sunday/Monday. I'm on a business trip until Saturday and not at home till Sunday evening.

Additionally, I can hardly imagine that removing the driver would help. It never did in the past, because all necessary files get overridden while updating the Catalyst-stuff.
Reinstalling the FAH-client to get stock-config (Maybe FAH could provide kinda "config-sample.xml" in the future?) is an option, though.
User avatar
Eagle
 
Posts: 157
Joined: Sun Feb 17, 2008 1:06 am
Location: » Earth » Europe » Germany

Re: V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Postby Fallen-Angel » Mon Nov 21, 2011 10:53 pm

Yeah I know what you think of deleting the driver and and and, but I did that and I don't know why, but I have a good feeling about that.... Just a try.

I understand, it's okay. Then I will experiment and search for any solved (-1073740777 = 0xc0000417)-error till Monday.
User avatar
Fallen-Angel
 
Posts: 35
Joined: Sat Sep 24, 2011 11:41 am
Location: Germany

Re: V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Postby bruce » Mon Nov 21, 2011 11:54 pm

Fallen-Angel wrote:I've read the whole topic again and just wanted to mention, that if it were the unstable over-clocking (on the GPU), the system freezes or the blue screen appears, before the client responds (throw a line message which describes the error). That happened in my case when I oced the GPU too far or just too bad. When ocing the CPU, it may happen different things. Here I expect the client's reaction first, followed by a blue screen.


Unstable overclocking USUALLY behaves the way you describe, but not always. In the cases of GPUs, an occasional error might produce an incorrect result one out of (large number) of times with nothing but a pixel of the the wrong shade. Your eye wouldn't care, but FAH would -- and the problem still falls into the things that are properly called unstable overclocks even though no "crash" has occurred. When a crash actually occurs the type of error depends on exactly where it happened in the code. The system might freeze or might BSOD or might even issue another type of error report.

Things that are stable produce predictable results. Things that are unstable produce unpredictable results even if they produce mostly repeatable results 99% of the time.
bruce
 
Posts: 22369
Joined: Thu Nov 29, 2007 10:13 pm
Location: So. Cal.

Re: V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Postby Fallen-Angel » Tue Nov 22, 2011 5:47 am

bruce wrote:
Fallen-Angel wrote:I've read the whole topic again and just wanted to mention, that if it were the unstable over-clocking (on the GPU), the system freezes or the blue screen appears, before the client responds (throw a line message which describes the error). That happened in my case when I oced the GPU too far or just too bad. When ocing the CPU, it may happen different things. Here I expect the client's reaction first, followed by a blue screen.


Unstable overclocking USUALLY behaves the way you describe, but not always. In the cases of GPUs, an occasional error might produce an incorrect result one out of (large number) of times with nothing but a pixel of the the wrong shade. Your eye wouldn't care, but FAH would -- and the problem still falls into the things that are properly called unstable overclocks even though no "crash" has occurred. When a crash actually occurs the type of error depends on exactly where it happened in the code. The system might freeze or might BSOD or might even issue another type of error report.

Things that are stable produce predictable results. Things that are unstable produce unpredictable results even if they produce mostly repeatable results 99% of the time.


Hm, ok I agree.

This night I was able to run both GPUs with advanced flag without fails and now I set the SMP to advanced, too. We'll see how it goes together, but I think my ENUM issue is kind of solved. Nevertheless, I'll try some different beta configuration, today.
Last edited by Fallen-Angel on Tue Nov 22, 2011 1:13 pm, edited 1 time in total.
User avatar
Fallen-Angel
 
Posts: 35
Joined: Sat Sep 24, 2011 11:41 am
Location: Germany

Re: V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Postby bruce » Tue Nov 22, 2011 6:15 am

If we assume that the advanced projects that are being assigned today produce more heat, then you might see an instability that would disappear if you got different assignments. The presence or absence of the "advanced" flag is more likely unrelated, at least on a long-term basis. Advanced project turn into standard FAH projects after some time passes and the next batch of adv projects might run cooler.

I'm assuming that this is a heat-related issue, and there's really no foundation for that to necessarily be true. It could just as well be bad WUs or something totally unrelated.
bruce
 
Posts: 22369
Joined: Thu Nov 29, 2007 10:13 pm
Location: So. Cal.

Re: V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Postby Eagle » Tue Nov 22, 2011 6:59 am

Well, I left everything as it was - except the advanced-flag being set to "normal" before I fell asleep.

Result: GPU is still folding, no UNKNOWN_ENUM error so far - so, my guess is that it's the WU not being compatible with the driver(s)..
User avatar
Eagle
 
Posts: 157
Joined: Sun Feb 17, 2008 1:06 am
Location: » Earth » Europe » Germany

Re: V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Postby MtM » Tue Nov 22, 2011 10:29 am

Bruce, the original problem in this thread relates to the detection problem tied to ticket 752.

It's a laptop, and has both a hd4500 and a rv710? That's just as my card ( hd4500 is higher clocked rv710/HD4350 ).
MtM
 
Posts: 3054
Joined: Fri Jun 27, 2008 2:20 pm
Location: The Netherlands

Re: V7.1.38 cause UNKNOWN_ENUM (-10 = 0xfffffff6) on ATI

Postby Fallen-Angel » Tue Nov 22, 2011 1:32 pm

bruce wrote:If we assume that the advanced projects that are being assigned today produce more heat, then you might see an instability that would disappear if you got different assignments. The presence or absence of the "advanced" flag is more likely unrelated, at least on a long-term basis. Advanced project turn into standard FAH projects after some time passes and the next batch of adv projects might run cooler.

I'm assuming that this is a heat-related issue, and there's really no foundation for that to necessarily be true. It could just as well be bad WUs or something totally unrelated.


I can't imagine that advanced WUs turn to normal WUs on GPU, because I folded the advanced WUs, which have the 0x16cores for two months without stopping and didn't got "normal" WUs. Here under normal I understand 0x11cores because if you don't change the flag to beta or advanced, you only get 0x11 ones (at least for AMD GPUs). About the heat, unfortunately I didn't record the temps so I can't tell you what it was during the fails. Nevertheless if the temps were too high, my system would leave a message for me... I don't think that was a cause in my case.

@Eagle, can you post your CPU-Z stats? But everything, from CPU till Graphics tab. And which FahCores did your GPU get?

Best,

FA

P.S.: It seems to be stable now. I don't have any fails and that with three advanced clients (2xGPU, SMP) ^^
User avatar
Fallen-Angel
 
Posts: 35
Joined: Sat Sep 24, 2011 11:41 am
Location: Germany

PreviousNext

Return to V7.1.52 Windows/Linux

Who is online

Users browsing this forum: No registered users and 1 guest

cron