Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Topics - wicked

Pages: [1]
1
Questions / Scheduler RPC pending reasons
« on: August 23, 2014, 02:07:42 pm »
Hello,

Looks like BoincTasks doesn't always show correct info on Status-column of the Projects-tab. For example, at the moment WUProp@Home scheduler is not responding (incorrect DNS resolution) and BoincTasks shows:

WUProp@Home ... Deferred for: 00:53:50, In progress

Looking boinccmd --get_project_status output for the same project:

1) -----------
   name: WUProp@Home
...
   nrpc_failures: 8
   master_fetch_failures: 0
   master fetch pending: no
   scheduler RPC pending: yes
   trickle upload pending: no
   attached via Account Manager: yes
   ended: no
   suspended via GUI: no
   don't request more work: no
   disk usage: 0.000000
   last RPC: 1408626277.443256
   project files downloaded: 0.000000


And project info in client_state.xml shows:

<project>
    <project_name>WUProp@Home</project_name>
...

    <nrpc_failures>8</nrpc_failures>
    <master_fetch_failures>0</master_fetch_failures>
    <min_rpc_time>1408805030.526604</min_rpc_time>
    <next_rpc_time>1408629877.443256</next_rpc_time>
    <sched_rpc_pending>7</sched_rpc_pending>
    <non_cpu_intensive/>


Now, the BT Deferred status is correct but RPC is not "In progress" but rather RPC is "Pending" and the reason code 7 means "Project Requested" as defined in http://boinc.berkeley.edu/gitweb/?p=boinc-v2.git;a=blob;f=lib/common_defs.h;hb=HEAD#l193.  I've also sometimes seen "sRPCp 3" status, which might mean BT doesn't also understand the reason code 3, which is explained in the same header. (RPC for this project is pending because "Need work".) Could you make sure these defines are correctly taken into account for BT Status-column? Thanks!

BTW, there's also upload pending and it's status is shown correctly on Transfers tab:

WUProp@Home ... Upload pending (Project backoff: 00:46:21)

2
Hi,

I'm test TThrottle 2.10.1 on a system that has one ATI Radeon HD5870 based card installed. It also has two nVidia GeForce GTX285 based cards. Whenever I connect my two monitors to the ATI card, TThrottle incorrectly detects two ATI cards and reports two, almost identical, temperatures in addition to the different nVidia temperatures. If I remove the second monitor (actually move it to one of the nVidia cards) TThrottle detectes correctly only one ATI card.

Here's a log excerpt from TThrottle, when I start it with two monitors connected:

Code: [Select]
29 August 2010 - 14:29:32 Driver installed properly. Driver Version: 2.0

Program version: 2.10 64Bit
Microsoft  (build 7600)

nvidia: found 2 logical devices
nvidia: found 2 physical devices
nvidia: Temperature 64 C, max Temperature 191 C
nvidia: Temperature 74 C, max Temperature 191 C

nvidia: GeForce GTX 285, GeForce GTX 285
Amd/Ati: found 2 devices
Amd/Ati: ASUS EAH5870 Series P (0-2)
29 August 2010 - 14:29:32 AdapterInfo: 0-2, Index: 2, ASUS EAH5870 Series
Pos: 0, Temperature: 90
Amd/Ati: ASUS EAH5870 Series P (1-3)
29 August 2010 - 14:29:32 AdapterInfo: 1-3, Index: 3, ASUS EAH5870 Series
Pos: 1, Temperature: 90

And here's a log excerpt when I start TThrottle with only one monitor connected (simply removed the monitor, no reboot, only restarted TThrottle to make it detect things again):

Code: [Select]
29 August 2010 - 14:40:10 Driver installed properly. Driver Version: 2.0

Program version: 2.10 64Bit
Microsoft  (build 7600)

Language: User: 1035 FIN ,System: 1035 FIN
Language: User: 1033 ENU ,System: 1033 ENU

nvidia: found 2 logical devices
nvidia: found 2 physical devices
nvidia: Temperature 64 C, max Temperature 191 C
nvidia: Temperature 74 C, max Temperature 191 C

nvidia: GeForce GTX 285, GeForce GTX 285
Amd/Ati: found 1 devices
Amd/Ati: ASUS EAH5870 Series P (0-2)
29 August 2010 - 14:40:10 AdapterInfo: 0-2, Index: 2, ASUS EAH5870 Series
Pos: 0, Temperature: 88

3
Wish List / Use field
« on: July 23, 2010, 02:15:51 pm »
Hello,

1) A while back I complained about numerous non-CPU intensive FreeHAL WUs taking too much space. There was a change later to make FreeHAL WUs say "Use 0.025C" because previously they didn't say anything. Now, they have added CPU-intensive WUs and actually have one WU limit. So no even the CPU-intensive jobs are marked as "Use 0.025C" in BoincTasks. :( One could detect non-CPU intensive WUs from Name that starts with fh_nci_ and CPU intensive that start with fh_ci_. Not sure if it's a good idea to try detect these or just remove any special code dealing with FreeHAL.. They really should fix their reporting errors.

2) Any zero digits seems to be removed in Use field. This is good, because "0.05C + 1NV" is much nicer. However, it seems it doesn't work when you have more than one required resources. For example, I can see "0.75C + 2.00NV" or "1C + 3.00NV". Hopefully it also supports ATI cards as I'm about to add such beasts to my arsenal. :)

4
Beta Testing / BT 0.46 crashed after returning from sleep
« on: March 25, 2010, 05:23:43 am »
Hi,

I just noticed that BoincTask64 had crashed when the computer automatically returned from sleep this morning. :o
Maybe this event information helps you track down the reason and possibly fix something in the code.

Quote

Log Name:      Application
Source:        Application Error
Date:          25.3.2010 5:59:42
Event ID:      1000
Task Category: (100)
Level:         Error
Keywords:      Classic
User:          N/A
Description:
Faulting application name: boinctasks64.exe, version: 0.4.6.0, time stamp: 0x4ba555a7
Faulting module name: boinctasks64.exe, version: 0.4.6.0, time stamp: 0x4ba555a7
Exception code: 0xc0000005
Fault offset: 0x0000000000001c5e
Faulting process id: 0xab0
Faulting application start time: 0x01cacab4e70ae87a
Faulting application path: C:\Program Files\eFMer\BoincTasks\boinctasks64.exe
Faulting module path: C:\Program Files\eFMer\BoincTasks\boinctasks64.exe
Report Id: d6d536de-37c2-11df-889c-00224369f427

5
Wish List / Initializing status in Projects
« on: March 13, 2010, 07:38:07 pm »
Hi,

I see lot of "Initializing" statuses in Projects (in v0.44). These seem to happen also to projects that have been in the computer for sometime already. To me these look more like "Scheduler request in progress" status when comparing to Messages for the computer. Especially since they happen to all after Account Manager sync that trigger scheduler request to all projects.

Maybe this is related to the following change in v0.43?

Quote
-(FIXED) Project view code 5 = initializing, 7 = in progress.

6
Hi,

Even though most of my problems with FreeHAL project tasks were fixed (thanks for that!) I'd still like to hide it's numerous tasks. So I would like an option in Tasks tab to filter out all tasks for a specific project (or projects). Other option is to allow combining on Running status for one or more projects.

I'd like to have this feature because FreeHAL project has upto 25 running tasks per computer. Display gets crowded pretty fast and I'm usually not interested on this project but rather see how tasks for all the other projects are progressing. Add in my hesitation to touching my mouse to scroll the long task list because that would suspend my GPUs and you might understand me better. :)

BTW, please, could you update your donation status and wishes? Thank you, and keep up the good work! (But enjoy your vacation.)

7
Hello!

It's so awesome to find an active project after trying BoincView and getting excited only to realize it's development has stopped. After using BoincTasks for a day I have few wishes that'll make it a perfect tool for me use. Some of these I miss after trying BoincView. Here goes:

    1. Better handling for non-CPU intensive projects like FreeHAL (no, it doesn't use GPU either). Now the CPU % and Elapsed Time seem totally off (jumping between 0.xx% and ~80%) and they keep triggering the Warning color due to the default CPU rule. (Yes, I already disabled the warning by deleting the rule.) These should be marked as non-CPU intensive (on Use column?) and excluded somehow from the low CPU warnings (rather warn when they ARE using much CPU).
    2. Quick fix for the numerous non-CPU intensive tasks would be to allow exclude (filter out) one or more projects from the lists (especially Tasks). This also helps with the clutter since these add lots of running jobs..
    3. I miss the current speed (for example GFlops per second) display of the running tasks that BoincView had. This allowed me to compare effectiveness of different computers running same project. Then I could decide if I should prefer running some project on some specific computer and save the other computer running something else.
    4. Sorting with Use column on the Tasks tab doesn't seem to work. Even if I set it up as primary sort criteria it doesn't list blank lines as a group but intermingles them with GPU tasks. Yes, I tried this to have GPU and CPU tasks listed separately. Also, same problem seems to prevent me to sort Projects tab first with Project and then Computer.
    5. I have two computers that seems to corrupt their stored password every time BoincTasks is closed and started again. Looking at the XML file there doesn't seem to anything strange there but still the passwords won't survive intact. :(
    6. Projects list in Statistics graphics window seems to gather lots of blank entries that multiply every time it's refreshed.
    7. I really like the idea of adding Computers as a "normal tab" that's on the task list. After that's done it can be enhanced to add measured speed of host and other details like BoincView has in it's Hosts tab. :)
    8. Some projects have blank name in the Projects tab. These are probably ones that haven't been able to contact project servers to fetch their official name. BoincTasks should detect this condition and switch to project URL or some other identification string for these.. This is how Boinc Manager handles these projects.

That's probably enough wishes for now. ;) Thanks for listening!

Pages: [1]