Page 2 of 2

Re: FaHCore_21.exe has stopped working - GTX970, ver. 384.94

Posted: Mon Sep 04, 2017 4:57 pm
by bruce
When Windows detects a serious error ... like a GPU failure ... it only has two choices. 1) Kill the app that was involved or 2) BSOD/Reboot.

By display the popup on you screen, Windows started choice 1, informing you that FAHCore_xx was going to be killed. At that momement, its processing had been suspended but it was still resident in (virtual) memory in case you wanted to examine the code or the status registeres or perform a specialized core-dump. Once you acknowledged the popup, it proceeds to clear out the "offending" program and reports the error to FAHClient.

That explains why no progress was being made ... because FAHCore_xx had been suspended by the OS.

Perhaps error 0xc0000374 is one of the ways WIndows reports an out-of-memory condition. If you've been leaving FAHControl running, the known memory leak might be the only problem that Windows detected. In any case, the event viewer report would be soon after the last progress message from that FAHCore.

Re: FaHCore_21.exe has stopped working - GTX970, ver. 384.94

Posted: Sun Sep 10, 2017 8:47 pm
by matitaccia
Dear all,

thank you very much for all the answers.

I might have found the culprit: almost by mistake I stumbled in to device manager and discovered that I hadn't installed everything properly. After having installed Intel Chipset Drivers it seems to be working better and with no crashes (cross fingers!).

Out of the two unidentified devices there is still one which I don't know how what it is or how to identify. Even the Intel Utility is not helping at all (I any case I don't like this utility too much, as it's not the first time where it fails to identify the correct driver).

In any case, for the moment being, FAH hasn't crashed yet.

Thank you very much!

matitaccia

P.S.
I have tried to give a look at Windows Event viewer... it's not that easy to read and to understand!

Re: FaHCore_21.exe has stopped working - GTX970, ver. 384.94

Posted: Sun Sep 10, 2017 10:30 pm
by bruce
Try running FAHClient --lspci

Re: FaHCore_21.exe has stopped working - GTX970, ver. 384.94

Posted: Mon Sep 11, 2017 5:46 am
by matitaccia
Here you are:

Code: Select all

VendorID:DeviceID:Vendor Name:Description
0x10de:0x0fbb:NVIDIA Corporation:High Definition Audio Controller
0x10de:0x13c2:NVIDIA Corporation:NVIDIA GeForce GTX 970
0x14e4:0x43b1:Broadcom Corporation:ASUS PCE-AC56 802.11ac Network Ad
0x8086:0x0c00:Intel Corporation:Intel(R) Xeon(R) processor E3 - 1200 Core processor DRAM Controller - 0C00
0x8086:0x0c01:Intel Corporation:Intel(R) Xeon(R) processor E3 - 1200 Core processor PCI Express x16 Controller - 0C01
0x8086:0x153b:Intel Corporation:Intel(R) Ethernet Connection I217-V
0x8086:0x8c02:Intel Corporation:Standard AHCI 1.0 Serial ATA Control
0x8086:0x8c10:Intel Corporation:Intel(R) 8 Series/C220 Series PCI Ex ort #1 - 8C10
0x8086:0x8c18:Intel Corporation:Intel(R) 8 Series/C220 Series PCI Ex ort #5 - 8C18
0x8086:0x8c20:Intel Corporation:High Definition Audio Controller
0x8086:0x8c22:Intel Corporation:Intel(R) 8 Series/C220 Series SMBus 8C22
0x8086:0x8c26:Intel Corporation:Standard Enhanced PCI to USB Host Co
0x8086:0x8c2d:Intel Corporation:Standard Enhanced PCI to USB Host Co
0x8086:0x8c31:Intel Corporation:Intel(R) USB 3.0 eXtensible Host Con
0x8086:0x8c3a:Intel Corporation:Intel(R) Management Engine Interface
0x8086:0x8c44:Intel Corporation:Intel(R) Z87 LPC Controller - 8C44

Re: FaHCore_21.exe has stopped working - GTX970, ver. 384.94

Posted: Mon Sep 11, 2017 5:31 pm
by bruce
According to that report, your OS recognizes two types of devices which FAH can use:
Intel and Nvidia FAH recognizes your GPU: 0x10de:0x13c2:NVIDIA Corporation:NVIDIA GeForce GTX 970 and your CPUs.

The other devices listed are for general use, including memory, your network device, the PCIe bus, etc. but they're unlikely to have problems related to or reported by FAHClient.

Re: FaHCore_21.exe has stopped working - GTX970, ver. 384.94

Posted: Mon Sep 18, 2017 5:32 pm
by matitaccia
Dear all,

I can confirm. After installing the correct chipset drivers I haven't got any further issue. The computer is folding continuously without any app crash.

Thank you all for your support.

matitaccia

Re: FaHCore_21.exe has stopped working - GTX970, ver. 384.94

Posted: Sun Nov 05, 2017 5:18 pm
by matitaccia
Dear all,

it has happened once more. Maybe this time it is Nvidia related (and not to the chipset as I was mentioning in the previous posts) as there is a reference to a ntdll.dll... Please have a look at the event viewer log and tell me what you think about it.

Code: Select all

Log Name:      Application
Source:        Application Error
Date:          04.11.2017 04:12:37
Event ID:      1000
Task Category: (100)
Level:         Error
Keywords:      Classic
User:          N/A
Computer:      Echo
Description:
Faulting application name: FahCore_21.exe, version: 0.0.0.0, time stamp: 0x588257cc
Faulting module name: ntdll.dll, version: 6.1.7601.23915, time stamp: 0x59b94ee4
Exception code: 0xc0000374
Fault offset: 0x00000000000bf3e2
Faulting process id: 0x1008
Faulting application start time: 0x01d3551ac38f226c
Faulting application path: C:\Users\matih\AppData\Roaming\FAHClient\cores\fahwebx.stanford.edu\cores\Win32\AMD64\NVIDIA\Fermi\Core_21.fah\FahCore_21.exe
Faulting module path: C:\Windows\SYSTEM32\ntdll.dll
Report Id: 0237ff29-c10e-11e7-8d88-e03f497e41ab
Event Xml:
<Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
  <System>
    <Provider Name="Application Error" />
    <EventID Qualifiers="0">1000</EventID>
    <Level>2</Level>
    <Task>100</Task>
    <Keywords>0x80000000000000</Keywords>
    <TimeCreated SystemTime="2017-11-04T03:12:37.000000000Z" />
    <EventRecordID>5662</EventRecordID>
    <Channel>Application</Channel>
    <Computer>Echo</Computer>
    <Security />
  </System>
  <EventData>
    <Data>FahCore_21.exe</Data>
    <Data>0.0.0.0</Data>
    <Data>588257cc</Data>
    <Data>ntdll.dll</Data>
    <Data>6.1.7601.23915</Data>
    <Data>59b94ee4</Data>
    <Data>c0000374</Data>
    <Data>00000000000bf3e2</Data>
    <Data>1008</Data>
    <Data>01d3551ac38f226c</Data>
    <Data>C:\Users\matih\AppData\Roaming\FAHClient\cores\fahwebx.stanford.edu\cores\Win32\AMD64\NVIDIA\Fermi\Core_21.fah\FahCore_21.exe</Data>
    <Data>C:\Windows\SYSTEM32\ntdll.dll</Data>
    <Data>0237ff29-c10e-11e7-8d88-e03f497e41ab</Data>
  </EventData>
</Event>
Thank you very much!
matitaccia

Re: FaHCore_21.exe has stopped working - GTX970, ver. 384.94

Posted: Tue Nov 07, 2017 5:03 am
by bruce
Ntdll.dll is not part of FAH; it's essentially basic Windows. The error code 0xc0000374 indicates that Windows has detected a "heap corruption."

This can be caused by hardware errors (defective hardware, overheating, overclocking etc.) or simply by running out of memory. Which version of FAHClient are you running?

As I suggested above, the beta version of FAHControl has a known memory leak bug. Until it's fixed, we recommend that you use FAHControl whenever you need it but you should shut it down when you're not using it rather than leaving it running continuously. There are unconfirmed reports of a memory leak in FAHCore_21.

What are your Windows virtual memory settings?

Re: FaHCore_21.exe has stopped working - GTX970, ver. 384.94

Posted: Fri Nov 10, 2017 4:13 pm
by matitaccia
Client version: 7.4.4

I remember about the memory leak. I keep the FAH Control application closed. Still sometimes the error rarely happens (it hasn't has happened since my last post for example, dated Nov 5th).

Windows Virtual Memory Settings Image

Is that print screen enough?

Thank you all!

matitaccia

Re: FaHCore_21.exe has stopped working - GTX970, ver. 384.94

Posted: Mon Nov 13, 2017 4:10 pm
by bollix47
I'm not totally convinced this is a FAHControl memory leak ... yes, there is a default theme problem in Windows but that's not necessarily related to the memory leak.

Most of my setups are using Ubuntu and Xubuntu with one using win 10 pro and they rarely, if ever, run FAHControl. All are using 7.4.16. Yet, the memory usage for FAHCient grows over time and I have seen it as high as 20 GB. Needless to say that computer with 8 GB of ram was running a bit sluggish! So I went around to all my computers and they were all using extremely high amounts of memory, many in the gigabyte range.

If the user reboots once every week or two due to updates etc. the memory usage would probably not reach some of the extremes I've seen and they may not even notice but it's there and it appears to be in FAHClient. Perhaps something to do with the logs continually growing and memory not being expunged when the logs rotate??? Just a wild guess and probably a wrong one!