Index  | Recent Threads  | Unanswered Threads  | Who's Active  | Guidelines  | Search
 
World Community Grid Forums
  Category: Beta Testing
  Forum: Beta Test Support Forum
  Thread: HCC1 Beta Test for Multiple Images in a work unit (Issues Thread)

Quick Go »
No member browsing this thread
Thread Status: Active
Total posts in this thread: 59
Posts: 59   Pages: 6   [ Previous Page | 1 2 3 4 5 6 ]
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 7646 times and has 58 replies Next Thread
yoro42
Ace Cruncher
United States
Joined: Feb 19, 2011
Post Count: 8976
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: HCC1 Beta Test for Multiple Images in a work unit (Issues Thread)
Picked up three. All completed normally.

BETA_ X0960075630557200610031058-- 201211082324070_ 2 Coltrane Valid 11/9/12 05:44:23 11/9/12 08:58:13 0.06 / 0.14 63.4 / 202.5
BETA_ X0960075130729200609122126-- 201211082324180_ 0 Thelonious Valid 11/9/12 02:13:05 11/9/12 05:04:41 0.34 / 0.36 58.5 / 60.6
BETA_ X0960075630963200610031051-- 201211082324070_ 0 Thelonious Valid 11/9/12 01:54:02 11/9/12 05:27:59 0.35 / 0.37 59.7 / 59.0
----------------------------------------
[Nov 9, 2012 9:07:44 PM]   Link   Report threatening or abusive post: please login first  Go to top 
cristipurdel
Senior Cruncher
Joined: Dec 13, 2008
Post Count: 158
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: HCC1 Beta Test for Multiple Images in a work unit (Issues Thread)
When does the checkpoint occur? After every batch (there are 5 of them) or after each image?
----------------------------------------


How to Recycle Android Phones for BOINC Rig
[Nov 9, 2012 11:50:53 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: HCC1 Beta Test for Multiple Images in a work unit (Issues Thread)
Did you notice if the process went into zombie mode? Also, have you had issues with stuck work units in the past (even on other projects)?


I don't know "zombie mode"
The other 3 cores on that CPU were all making progress (and checkpointing) on work units from the HCC task here and from the Einstein Project.

Yes, I have seen stuck work units before... that's why I try to check BOINCTasks at least once/day... looking back through its History, I see very few work units with more than 12% slippage time (CPU over elapsed times). So far, it appears 7.05 is about 1 percentage point more efficient than 6.56 was... 7.05's ranging from 90.5% to 92% efficient, while the typical CPU/elapsed efficiency for 6.56 on that machine was typically 87.9% to 91%, with CPU times of about an hour (1:00:00) to 1:25:00.

The fastest time on the 7.05 application so far has been 59 minutes and the slowest 1:38:13 (the 7.05 BETA work unit that report was about eventually ended up with 2:06:58 CPU time).

Rarely is there an outlier/flyer... e.g. out of the last 425 HCC results on it, there were 9 that dropped to 74% or 75% efficiency.
There were also 6 out of the 425 that went to Computational Error with very-low efficiencies (e.g. X0960073620699200608081015_1 - 2:00:48 elapsed, 0:19:17 CPU time; X0960072730885200608091643_0 - 1:56:33 elapsed, 0:24:32 CPU time; X0930072770325200607191301_0 - 1:29:11 elapsed, 0:00:51 CPU time).
[Nov 10, 2012 4:23:26 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: HCC1 Beta Test for Multiple Images in a work unit (Issues Thread)
Hi cristipurdel,
When does the checkpoint occur? After every batch (there are 5 of them) or after each image?

It is not possible to checkpoint during an image operation. Therefore, a checkpoint is written after an image is processed. If a work unit has n images, then it can have at most (n - 1) checkpoints, since there is no reason to write a checkpoint after the final image.

Lawrence
[Nov 10, 2012 5:33:43 AM]   Link   Report threatening or abusive post: please login first  Go to top 
LAZA74
Advanced Cruncher
Germany
Joined: Sep 28, 2008
Post Count: 56
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
confused Re: HCC1 Beta Test for Multiple Images in a work unit (Issues Thread)
[
Just add the beta entry into ur app_info and you should be good to go.

    <app>
        <name>beta3</name>
		<user_friendly_name>Help Conquer Cancer</user_friendly_name>
    </app>
    <file_info>
        <name>wcg_beta3_img_7.05_windows_intelx86__ati_hcc1</name>
        <executable/>
    </file_info>
    <file_info>
        <name>hcckernel.cl.7.05</name>
        <executable/>
    </file_info>
    <app_version>
        <app_name>beta3</app_name>
        <version_num>705</version_num>
        <platform>windows_intelx86</platform>
		<plan_class>ati_hcc1</plan_class>
        <avg_ncpus>1.0</avg_ncpus>
        <max_ncpus>1.0</max_ncpus>
		<coproc>
            <type>ATI</type>
            <count>1</count>
        </coproc>
        <file_ref>
            <file_name>wcg_beta3_img_7.05_windows_intelx86__ati_hcc1</file_name>
            <main_program/>
         </file_ref>
        <file_ref>
	       <file_name>hcckernel.cl.7.05</file_name>
		   <open_name>hcckernel.cl</open_name>
        </file_ref>
    </app_version>



So this is the problem for 'normal' people like me, living at the edge of the world! crying

I'm not at home on weekdays an don't want to hack this workarounds to get some betas - and my wife will kill me, if i sit any longer on those 2 machines here!

So i will NEVER got any beta more, and this is truly sad.

One machine is crunching 6.56 and 7.05 without problems on an HD 5570...
----------------------------------------
NAS - Eigenbau
Xiaomi Mi 10T
[Nov 11, 2012 7:48:52 AM]   Link   Report threatening or abusive post: please login first  Go to top 
mikey
Veteran Cruncher
Joined: May 10, 2009
Post Count: 821
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: HCC1 Beta Test for Multiple Images in a work unit (Issues Thread)
As an aside, how about a link to more about the app_info.xml file in the FAQ ?
I searched it for "anonymous" and "app_info" but got no hits, and followed a couple of the GPU links but there was no mention of app_info.xml, either.


The PROBLEMS with an app_info file is that if can be tweaked to add additional options and then DESTROY your gpu! Heat is the biggest problem to a pc, tweaking your gpu usually means making it do more than the defaults and this usually means more heat! MOST warranties do NOT cover heat related damages and this can get VERY expensive VERY quickly!! People using a default app_info file should NOT have any problems, but tweaking it CAN cause LOTS of problems, including invalid units! Just as overclocking a cpu can, this is USUALLY why most projects do not tell you how to do this, they let you discover and take your own chances.
----------------------------------------
[Nov 11, 2012 3:04:49 PM]   Link   Report threatening or abusive post: please login first  Go to top 
ThreadRipper
Veteran Cruncher
Sweden
Joined: Apr 26, 2007
Post Count: 1320
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: HCC1 Beta Test for Multiple Images in a work unit (Issues Thread)
As an aside, how about a link to more about the app_info.xml file in the FAQ ?
I searched it for "anonymous" and "app_info" but got no hits, and followed a couple of the GPU links but there was no mention of app_info.xml, either.


The PROBLEMS with an app_info file is that if can be tweaked to add additional options and then DESTROY your gpu! Heat is the biggest problem to a pc, tweaking your gpu usually means making it do more than the defaults and this usually means more heat! MOST warranties do NOT cover heat related damages and this can get VERY expensive VERY quickly!! People using a default app_info file should NOT have any problems, but tweaking it CAN cause LOTS of problems, including invalid units! Just as overclocking a cpu can, this is USUALLY why most projects do not tell you how to do this, they let you discover and take your own chances.


Any retail GPU board should be able to handle 100% load without voiding the warraty. If you just increase the utilization via ap_info by running multiple WUs at a time it will not void your warranty and it should not ruin your GPU. Just check your GPU temps and if they are fine - good! If not, then you chould be able to file an RMA request to the manufacturer of the board since it shoul be able to run at 100% load.

Ok, now, if you overclock your GPU, your're on your own with no warranty to bak you up, but as long as your temps are cool enough you should be good to go.

EDIT: And if your PC hangs/blue-screens after an OC of GPU/CPU, or you start getting invalid results, simply back down your clocks a bit until you reach 100% stability and no invalids. Then you have found your sweet spot.
----------------------------------------

Join The International Team: https://www.worldcommunitygrid.org/team/viewTeamInfo.do?teamId=CK9RP1BKX1

AMD TR2990WX @ PBO, 64GB Quad 3200MHz 14-17-17-17-1T, RX6900XT @ Stock
AMD 3800X @ PBO
AMD 2700X @ 4GHz
----------------------------------------
[Edit 2 times, last edit by flodisar at Nov 11, 2012 9:56:28 PM]
[Nov 11, 2012 9:53:49 PM]   Link   Report threatening or abusive post: please login first  Go to top 
armstrdj
Former World Community Grid Tech
Joined: Oct 21, 2004
Post Count: 695
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: HCC1 Beta Test for Multiple Images in a work unit (Issues Thread)
Checkpoints for the new GPU version occur after each image is processed.
Thanks,
armstrdj
[Nov 12, 2012 3:05:07 PM]   Link   Report threatening or abusive post: please login first  Go to top 
nanoprobe
Master Cruncher
Classified
Joined: Aug 29, 2008
Post Count: 2998
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: HCC1 Beta Test for Multiple Images in a work unit (Issues Thread)
As an aside, how about a link to more about the app_info.xml file in the FAQ ?
I searched it for "anonymous" and "app_info" but got no hits, and followed a couple of the GPU links but there was no mention of app_info.xml, either.


The PROBLEMS with an app_info file is that if can be tweaked to add additional options and then DESTROY your gpu! Heat is the biggest problem to a pc, tweaking your gpu usually means making it do more than the defaults and this usually means more heat! MOST warranties do NOT cover heat related damages and this can get VERY expensive VERY quickly!! People using a default app_info file should NOT have any problems, but tweaking it CAN cause LOTS of problems, including invalid units! Just as overclocking a cpu can, this is USUALLY why most projects do not tell you how to do this, they let you discover and take your own chances.

1st. You can lower your temps and power consumption for this project by down clocking your GPU memory clocks. It will not effect how the HCC GPU tasks run. FWIW I down clock mine by at least 50%. Depending on the model of your card that can be 15c or more less heat and 30+/- watts less power.
2nd. Most if not all the newer models of GPUs automatically down clock the core themselves if they get too hot.
3rd. What is a default app_info file? confused
----------------------------------------
In 1969 I took an oath to defend and protect the U S Constitution against all enemies, both foreign and Domestic. There was no expiration date.
[Nov 13, 2012 4:38:07 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Posts: 59   Pages: 6   [ Previous Page | 1 2 3 4 5 6 ]
[ Jump to Last Post ]
Post new Thread