My Wish List - part 3.

Message boards : BOINC Manager : My Wish List - part 3.
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 . . . 8 · Next

AuthorMessage
Profile Gundolf Jahn

Send message
Joined: 20 Dec 07
Posts: 1069
Germany
Message 31610 - Posted: 16 Mar 2010, 9:09:18 UTC - in response to Message 31606.  
Last modified: 16 Mar 2010, 9:10:51 UTC

Looking at your sample messages with SETI...

It really is an example I found by searching the boards, as I don't own a CUDA-capable device myself. :-)

So, the only advice I can give is the "dreaded" reduce resolution bit, and don't forget to reboot regularly :-) as "stuck" memory seems only to be cleared by a power cycle.

Gruß,
Gundolf
PS: Perhaps the SETI CUDA (Questions and Answers) and Number crunching (Message boards) fora can provide more advice for you.
ID: 31610 · Report as offensive
bburwood

Send message
Joined: 15 Mar 10
Posts: 4
Australia
Message 31613 - Posted: 16 Mar 2010, 12:54:34 UTC - in response to Message 31610.  

You were right Gundolf. It's definitely the amount of free GPU RAM. I made a cc_config.xml file and saw the dreaded 231MB < 238MB message. So I dropped to 16 bit colour ... 236MB < 238 MB! Gggrrrr! 8 bit ... 238MB < 238MB!? I guess that one was borderline. I didn't try dropping the resolution, but that probably would have done the trick ... at the expense of a blurry image for everyday stuff - I don't think my mother would appreciated that! As for rebooting - it gets turned off (shutdown) every day so that shouldn't be an issue.

As you've suggested I'll continue the quest for a few extra MB GPU RAM on the SETI CUDA board.

That only leaves a new suggestion for the wish list (instead of my original misguided suggestion, which is probably already implemented anyway): that the client check the GPU RAM available against what the project will need *before* downloading a work unit, as seems to happen with normal work unit RAM requirements. Not sure how, or if it can be done in the client, but maybe when a project has a CUDA version available it might need to be required to say how much GPU RAM a work unit needs before that unit actually gets downloaded and assigned. IT seems to work ok for normal RAM requirements - perhaps some better checking is needed for CUDA RAM requirements? (so that thousands of workunits [~1050 in this case] don't get delayed by pc's downloading them that ultimately can't crunch them because they just miss out on having enough GPU RAM available.)

bb from Oz.
ID: 31613 · Report as offensive
skgiven
Avatar

Send message
Joined: 19 Aug 08
Posts: 87
United Kingdom
Message 31622 - Posted: 16 Mar 2010, 16:38:02 UTC - in response to Message 31613.  

Regarding Boinc 6.10.37
The new layout in Boinc Manager is confusing – CPU Tab.
Why have “While processor usage is less than 25%” set to 25% by default? This causes lots of Stop/Starts in GPU tasks, and slows GPU tasks way down – I don’t like it.
Why not list the changes made during the installation (changes from the previous installation that you can call with code) so people know what they have to configure?
There should be a separate Configuration Tab for GPUs, rather than mixing up GPU and CPU configurations under one Tab and with vague subsections such as “Use GPU while computer is in use”, then that “While processor usage is less than 25%” effort. Some default!
ID: 31622 · Report as offensive
Profile Jord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15477
Netherlands
Message 31623 - Posted: 16 Mar 2010, 17:00:46 UTC - in response to Message 31622.  
Last modified: 17 Oct 2010, 20:42:35 UTC

Why have “While processor usage is less than 25%” set to 25% by default?

This preference is a CPU-usage is for non-BOINC programs. It will suspend BOINC when another program (such as anti-virus, word-processor, browser, you name it... anything but for BOINC or one of its processes) will use that many CPU cycles. If you do not want to use it, set it to zero.

It's included for users who steadfastly complain that BOINC takes over their system, who find their computer to be slow when they start up other processes. And who then complain and deinstall BOINC.
ID: 31623 · Report as offensive
Saul L

Send message
Joined: 23 Dec 07
Posts: 4
Bolivia
Message 31661 - Posted: 19 Mar 2010, 6:43:12 UTC
Last modified: 19 Mar 2010, 6:56:37 UTC

Hello, I'd like to add my wishes for BOINC manager:

- Add a different method to run and WUs on the CPU: don't switch between project every xx minutes, just put all WUs on a single list separating projects with a horizontal line but run the WUs of all projects indistinctly.

- Take the example of Folding@Home GPU client (I'm running it right now alongside with BOINC: WCG) and run GPU projects with a minimum use of the CPU (10% max.): they use this method to configure it, although is not the default config, you have to do it manually, is really easy, but you have to fine tone it to achieve optimal use of both CPU+GPU, so an automatic setting for the environment variables would be great or use internal variables to BOINC client.

The sample config. is good enough for my system
(AMD Athlon II X2 2.8Ghz,
Hyper Transport @ 2Ghz,
ASUS ATI HD4670 1GB DDR3-1600 PCIE@x16,
RAM: G Skill F3-12800 2x2GB DDR3-1600 CL -6 black edition,
MoBo: ASUS M4A79XTD EVO AMD 790 ChipSet):

FLUSH_INTERVAL = 128-256 for a 48x0, 64-96 for a 38x0 (the optimum setting so that the GPU remains at 100% regardless of the WU, without causing too much lag should be within these ranges).

BROOK_YIELD = 2 (to stop utilising 100% of the CPU and therefore allow an additional CPU client to be started)

CAL_PRE_FLUSH = 1

CAL_NO_FLUSH = 1 (but should be reverted to 0 if it causes too many VPU Recovers).
ID: 31661 · Report as offensive
Profile Jord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15477
Netherlands
Message 31662 - Posted: 19 Mar 2010, 8:36:22 UTC - in response to Message 31661.  
Last modified: 17 Oct 2010, 20:42:38 UTC

- Add a different method to run and WUs on the CPU: don't switch between project every xx minutes, just put all WUs on a single list separating projects with a horizontal line but run the WUs of all projects indistinctly.

Not sure what you mean here. Can you explain better?
Do you mean Processor Affinity, whereby a task started on one CPU will run on that CPU until it's done? Or do you mean that you want to set a CPU (core) per project? Then the question becomes, what if you are attached to more projects than you have CPU (cores), with wildly varying deadlines?

- Take the example of Folding@Home GPU client (I'm running it right now alongside with BOINC: WCG) and run GPU projects with a minimum use of the CPU (10% max.): they use this method to configure it

Figure out a way to do that on the Mac and (any) Linux without too much fuss and without requiring to use a Windows specific ATI or Nvidia API and you may have a winner.

But even with this method, you'll find lots of people who won't know what to do. They'll want a button. They will not want to put in things in the environmental variables of their OS. Typos abound.
ID: 31662 · Report as offensive
Nicolas

Send message
Joined: 19 Jan 07
Posts: 1179
Argentina
Message 31664 - Posted: 19 Mar 2010, 15:18:39 UTC - in response to Message 31661.  
Last modified: 19 Mar 2010, 15:19:28 UTC

- Add a different method to run and WUs on the CPU: don't switch between project every xx minutes, just put all WUs on a single list separating projects with a horizontal line but run the WUs of all projects indistinctly.

You think "switching between projects every X minutes" is all BOINC does? The CPU scheduler is one of the most complex parts of the client.
ID: 31664 · Report as offensive
Saul L

Send message
Joined: 23 Dec 07
Posts: 4
Bolivia
Message 31675 - Posted: 19 Mar 2010, 20:17:53 UTC
Last modified: 19 Mar 2010, 20:22:56 UTC

Well, I meant to assign a CPU core to a WU until it ends, of course if you have more projects that CPU Cores then you need short queues and yes, a good way to schedule the start of the WUs per project.

I use this and it has proven to be effective (no CPU dispute at the seems):

FLUSH_INTERVAL = 128 (128+64= 192 was OK if I don't use the PC, if I use it looks like BOINC doesn't get enough attention of the CPU and restarts WUs)

BROOK_YIELD = 2 (this made the trick to only use CPU ~10%)

CAL_PRE_FLUSH = 1

CAL_NO_FLUSH = 1

And yes, would be better if we could have these parameters calculated according to the hardware.

19/03/2010 10:44:59 World Community Grid Restarting task X0000090131304200707311400_0 using hcc1 version 606
19/03/2010 11:00:18 World Community Grid Restarting task R00607_f81cf9ce2fa93ed602b1ae86c9a01f1b_02_9 using rice version 617
19/03/2010 11:15:19 World Community Grid Restarting task X0000090131304200707311400_0 using hcc1 version 606
19/03/2010 11:33:24 World Community Grid Restarting task R00607_f81cf9ce2fa93ed602b1ae86c9a01f1b_02_9 using rice version 617
19/03/2010 11:48:25 World Community Grid Restarting task X0000090121363200707111304_0 using hcc1 version 606
19/03/2010 12:04:41 World Community Grid Restarting task R00607_f81cf9ce2fa93ed602b1ae86c9a01f1b_02_9 using rice version 617
19/03/2010 12:19:42 World Community Grid Restarting task X0000090131304200707311400_0 using hcc1 version 606
19/03/2010 12:19:42 World Community Grid Restarting task nf128_00035_2 using hpf2 version 603
19/03/2010 12:49:46 World Community Grid Restarting task R00607_f81cf9ce2fa93ed602b1ae86c9a01f1b_02_9 using rice version 617
19/03/2010 13:06:48 World Community Grid Restarting task nf128_00035_2 using hpf2 version 603
19/03/2010 13:17:46 World Community Grid Restarting task R00607_f81cf9ce2fa93ed602b1ae86c9a01f1b_02_9 using rice version 617
19/03/2010 13:19:09 World Community Grid Restarting task X0000090131304200707311400_0 using hcc1 version 606
19/03/2010 13:58:06 World Community Grid Restarting task R00607_f81cf9ce2fa93ed602b1ae86c9a01f1b_02_9 using rice version 617
19/03/2010 14:07:26 World Community Grid Restarting task X0000090121363200707111304_0 using hcc1 version 606
19/03/2010 14:22:27 World Community Grid Restarting task X0000090131304200707311400_0 using hcc1 version 606
19/03/2010 14:35:20 World Community Grid Restarting task faah11276_ZINC32112590_xMut_md20240_01_0 using faah version 607
19/03/2010 14:49:26 World Community Grid Restarting task R00607_f81cf9ce2fa93ed602b1ae86c9a01f1b_02_9 using rice version 617
19/03/2010 15:05:35 World Community Grid Restarting task X0000090131304200707311400_0 using hcc1 version 606
19/03/2010 15:05:35 World Community Grid Restarting task nf128_00035_2 using hpf2 version 603
19/03/2010 15:31:26 World Community Grid Restarting task R00608_1681a3ebd878ad83fd89a7f2fa69d840_01_11 using rice version 617
19/03/2010 16:17:03 World Community Grid Restarting task nf128_00035_2 using hpf2 version 603
ID: 31675 · Report as offensive
Saul L

Send message
Joined: 23 Dec 07
Posts: 4
Bolivia
Message 31677 - Posted: 20 Mar 2010, 0:31:47 UTC - in response to Message 31675.  

My bad, those wasn't boinc restarts because of the F@H client, I set my config to switch between projects every 15 min, now is set to 999, the BOINC Client replaces 0 with 60 everytime I set it that way,well nothing serious...

cheers.
ID: 31677 · Report as offensive
Profile Jord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15477
Netherlands
Message 31678 - Posted: 20 Mar 2010, 0:39:03 UTC - in response to Message 31675.  
Last modified: 17 Oct 2010, 20:42:42 UTC

Well, I meant to assign a CPU core to a WU until it ends, of course if you have more projects that CPU Cores then you need short queues and yes, a good way to schedule the start of the WUs per project.

That's called CPU affinity and has been tested by some people. It hypothetically increases efficiency by a mere 5%, which isn't much and isn't worth the trouble of changing the whole CPU scheduling code to get that effect.

Play around with resource shares between projects and the time that applications may switch. The number isn't finite, it can be anything above zero.
ID: 31678 · Report as offensive
Nicolas

Send message
Joined: 19 Jan 07
Posts: 1179
Argentina
Message 31711 - Posted: 22 Mar 2010, 17:19:47 UTC - in response to Message 31709.  

My wishlist is to make an option to "force applications to 32bit even if it runs on 64bit machine(I mean so-called amd64/intel64). Currently 64bit applications of seti@home enhanced and astropulse are full of bugs and people want to 32bit applications. However recent boinc Linux core downloads 64bit applications even though the core itself is a 32bit application and began crunching!!

Wouldn't it make more sense to fix the bugs in the 64bit app??
ID: 31711 · Report as offensive
Profile Euphoriabuzz
Avatar

Send message
Joined: 6 Apr 10
Posts: 6
United Kingdom
Message 31986 - Posted: 6 Apr 2010, 12:22:05 UTC

Can we have an extra line on the projects tab giving us totals for work done and avg. work done please?
ID: 31986 · Report as offensive
Nicolas

Send message
Joined: 19 Jan 07
Posts: 1179
Argentina
Message 32006 - Posted: 7 Apr 2010, 5:31:48 UTC - in response to Message 31986.  

Can we have an extra line on the projects tab giving us totals for work done and avg. work done please?

What totals?

I have a 118,980 score in Hexic, and 225,200 in Bejeweled. Do I have a "total casual-gaming score" of 344,180?
ID: 32006 · Report as offensive
Profile Jord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15477
Netherlands
Message 32013 - Posted: 7 Apr 2010, 11:11:16 UTC - in response to Message 32006.  
Last modified: 17 Oct 2010, 20:42:45 UTC

'We' don't want Cross-Project-Credit Parity as apparently not all credit is the same all over BOINC, but 'we' do want to be able to add it all up. ;-)

Alas, BOINC is just a managing program, it's not a statistics program per se.

If you want to see all your credit&RAC added up and undergo intricate mathematical equations, please look at places such as BOINCstats, BOINC Synergy Stats, Mundayweb BOINC Stats Counters and such places.
ID: 32013 · Report as offensive
nhelder

Send message
Joined: 10 Apr 10
Posts: 3
United States
Message 32088 - Posted: 10 Apr 2010, 2:44:50 UTC

I'd like to vote for a item from "Wish List - 2" - namely, Part A from:
http://boinc.berkeley.edu/dev/forum_thread.php?id=1943&nowrap=true#22507

Restating that request slightly: it would be nice if you could specify, in your BOINC computation preferences, how long a "snooze" request/enable should last.

Movie playback often doesn't include any mouse/keyboard activity, and if you're playing a large H.264 encode, when BOINC kicks in it tends to degrade playback quite a bit.

Rather than having to exit BOINC entirely (or keep moving the mouse), if I could set a snooze request to last 2 hours and just do that prior to watching such a file, that'd be pretty slick.
ID: 32088 · Report as offensive
mmattel

Send message
Joined: 16 Apr 10
Posts: 1
Austria
Message 32220 - Posted: 16 Apr 2010, 14:19:46 UTC - in response to Message 32088.  

wish regarding internet / proxy access:

I have 3 computers running boinc. one of them is a notebook which is eighter connected via a proxy in the company or direct when I am outside.

the request is, that if there is a proxy set, boinc should try to go direct without proxy if it cannot reach it. this could be done eighter with a extra new setting in the options/proxy tab (like "try to bypass proxy if unreachable") or without extra setting just doing it. the report in the log should reflect the path used.

martin
ID: 32220 · Report as offensive
Nicolas

Send message
Joined: 19 Jan 07
Posts: 1179
Argentina
Message 32222 - Posted: 16 Apr 2010, 15:03:16 UTC - in response to Message 32220.  

I have 3 computers running boinc. one of them is a notebook which is eighter connected via a proxy in the company or direct when I am outside.

the request is, that if there is a proxy set, boinc should try to go direct without proxy if it cannot reach it. this could be done eighter with a extra new setting in the options/proxy tab (like "try to bypass proxy if unreachable") or without extra setting just doing it. the report in the log should reflect the path used.

That could be generalized to "I have different proxy settings and change between them" (which could mean proxy vs no proxy, or two different proxies). It would be better to have a solution that can support that too.

ID: 32222 · Report as offensive
dick

Send message
Joined: 15 May 08
Posts: 2
Netherlands
Message 32454 - Posted: 29 Apr 2010, 12:40:34 UTC

All.

Some years ago I posted a request to make Boinc IPv6 compatible (for Microsoft Windows).
The reason was that IPv4 addresses would be scarce in some years.

Now the number of IPv4 addresses available for new assignments has decreased to less than 10% (see Arin, Ripe, etc). The forecast is that in 2 years all IPv4 addresses will be assigned. Additional Internet addresses will be IPv6.

The newer Microsoft OSes are IPv6 enabled and operate fine in my environment.
Many applications are IPv6 compl






ient too.

So I repeat my question: When will we make Boinc IPv6 compliant?

Regards,

Dick.


ID: 32454 · Report as offensive
Schaduwtje

Send message
Joined: 1 May 10
Posts: 1
Netherlands
Message 32510 - Posted: 1 May 2010, 9:14:08 UTC - in response to Message 31109.  

I would very much like the option to prioritize the CPU (and GPU) workload. For example. If i leave my computer untouched for a minute, it will switch to full processorusage, or what ever i set it to, but then when i start to use my computer, it will slow down to about 50% or 25% usage.

As of now, the only option is workload while untouched or no workload.
Or it could be, that it stoppet using one of the cores, when you start to use the computer. Ofcourse, this would only work on multicore systems.
Or you could choose between the two.

Sometimes you dont need the full computer workload, when you use your computer, and in that case, this feature would be handy. In that way you could use your computer, have some processorpower at hand and meanwhile, still compute.

Regards
Hansen1982


+1 to this here.

I'd really love to be able to tell my computer to use 25% of its power (one of the 4 CPUs, say, or all 4 CPUs at 25% workload) when the computer's not idle, and 100% when it is idle. Instead, I have to set it to suspend all work when the computer's not idle. (I'm running v6.10.18 -- the newest one, I believe, and I see no options to change what I'm talking about. So please feel free to correct me if I'm wrong.)

Additionally, if I could (for example) tell it to use one GPU when the computer is not idle, and all GPUs when it is idle, that'd be fantastic too.

+1

This perfectly sums up my wishes. When I work on my laptop I don't want it to get hot and noisy, because it is working on 100%. However, when I am not using my laptop I am perfectly fine with this. So I would like different cpu-usage settings depending on if I am using the computer or not and not only the all or nothing "when I am using my computer" option.
ID: 32510 · Report as offensive
Anthony Chapman

Send message
Joined: 3 May 10
Posts: 2
United States
Message 32543 - Posted: 3 May 2010, 4:38:23 UTC

I would like to be able to tell my client:
When I am attached to SETI and MilkyWay, if SETI has no work available and my computers have finished all the WU's from SETI;
I would like my client to then receive1 WU per cpu or gpu from the next project on my list. The client should keep checking with SETI to see if it has new work. If it does, then it should finish the WU from MilkyWay, and then start working on the SETI projects again.
ID: 32543 · Report as offensive
Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 . . . 8 · Next

Message boards : BOINC Manager : My Wish List - part 3.

Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.