10% of active hosts (last 2 weeks) w/ GPU seems to provide 50% of all work

Fred J. Verster
Fred J. Verster
Joined: 27 Apr 08
Posts: 118
Credit: 22451438
RAC: 0

RE: RE: RE: Also GPUs

Quote:
Quote:
Quote:

Also GPUs only benefit the pulsar search (BRP4); they aren't used in the gravity wave work (SL6V1) where it's only CPU power that counts.

Which let me wonder why I get often no new work unit (within last 1-2 weeks) for that project (I deselected the BRP4 project fwiw)

Have a look at http://einsteinathome.org/node/196689

Well, maybe you missed the attemp to achieve the 1 PetaFLOP barrier as
processing power. {1+e9 x MFLOPS}! (1000 PetaFLOP)
(If I'm not terribly mistaken ;-))
This message appeared in the BOINC managers message service the last week.

These extra BRP4CUDA/ATI WUs can interfere with the bandwidth available
toalso send Gravity Wave work (SL6V1), or focus on BRP4GPU?!

99% of the available work is PULSAR search, done by (fast) GPUs, on my hosts.
And I increased Einsteins resources to 50% to contribute to this effort ;-)
(BOINC 7.0.28, x86 & x64; WIN XP & WIN7)

Well, just have a look how SETI struggles with their available bandwidth :-/

DanNeely
DanNeely
Joined: 4 Sep 05
Posts: 1364
Credit: 3562358667
RAC: 1580

RE: RE: ...we still

Quote:
Quote:
...we still support something like a 8800 GT (which is actually not that slow!!).

Works on an 8400GS 512Mb too (slowly), and I'm using the 9600 GT in some hosts (3,500s in a fast host, 5,000s in a slow one) which is a slower card than an 8800 GT (312 vs. 504GFLOPS, according to this list on Wikipedia).

Each GeForce range seems to overlap several other ranges above and below. For example, according to that list, the 8800GT still outranks everything up to a GT 640 in the current range.

There's massive overlap between generations. There's a 30:1 spread between the fastest and slowest cards in nVidia's 6xx series This is an average spread; it's ranged from as little as 10:1 to as much as 60:1, with the bottom end card generally defined by what Intel's IGP is capable of. Meanwhile the top end grows by an average of ~50%/generation (although there's a lot of variation here too).

http://en.wikipedia.org/wiki/Comparison_of_Nvidia_graphics_processing_units#GeForce_600_Series

mikey
mikey
Joined: 22 Jan 05
Posts: 11889
Credit: 1828213831
RAC: 201898

RE: Several

Quote:

Several thoughts:

BOINC was supossed to use the power of iddle computers to do something more usefull than stay there showing flying toasters while the user was bussy doing something else away of the keyboard...
So with that in mind any project should have at least one "not highly optimized" version of their apps which should run on almost any hardware...

But also, there is the matter of energy consumption and the effects on the planet, in this case it would be good to discourage old and inefficient hardware in same way... I mean if (for example) an old P4 were not supported then the owners will not think about using it for crunching 24/7 and most probably they will be saving energy without affecting too much the amount of work the project gets done...

On the other hand, having very optimized apps specifically designed to use the most of every different harware could work as a big motivation for heavy crunchers and credit hunters, which at the end could give the project that elusive petaflop...

From the point of view of the project my questions, knowing that it is not possible to handle thousands of versions, would be: from where it comes the work done? is the amount provided by the xx% top users/hardware really significative? Is it worth to give them more optimized apps loosing support for the legacy/botton hardware/users?

Of course, Ill be happy with ultra optimized versions specially designed for my specific hardware, but if all those optimizations are going to give me just a 1% extra my happiness wont be very noticed... :b

I understand what you are saying but disagree based on a persons money...a person buys a pc for x amount of dollars, they spent hard earned money on that pc and want it to last as long as it can, not replace it in two years because something else came out that is faster and better. Now let's add in that the person found Boinc and is crunching...should a project essentially FORCE that person to quit crunching or replace their pc to keep crunching, in my opinion NO they should not. Filling the landfills with old unusable pc's is just silly and wasteful and I think a better use is to use them to crunch with to solve some of the Worlds problems and/or questions. I would like to see project maintain workunits that support the older pc's as long as possible, including even potentially splitting a workunit into two workunits just for them, as opposed to only supporting the latest and greatest technology. YES I would love to ALSO see the same project make larger workunits for the newer pc's, supporting both as long as possible. Maybe even going so far as to recognize the processor being used and send the best units for that pc. We know they split LARGE units into smaller one to make workunits for our pc's, Seti does anyway and I am guessing others do too. Why not split it one or two more times and support the older pc's and keep them from the landfills. Split them a couple more times and maybe a smartphone type device can crunch too! YES absolutely it wastes electricity, but compared to landfills we have plenty of electricity. Newer pc's are more efficient and faster at the same time, how else would they get us to buy them? But unless you can still use that older pc you are replacing, where does it go besides the landfill. If you say 'handed down to someone else', then why can't they also crunch, the pc is running anyway!

Claggy....THANKS for setting me straight and correcting my mistakes!!!

tullio
tullio
Joined: 22 Jan 05
Posts: 2118
Credit: 61407735
RAC: 0

At SETI@home people with GPUs

At SETI@home people with GPUs produce thousands of invalid results since they are not using the right driver and since they do not visit the forums they are not aware of this. There is a thread dedicated to people who PM the owners of PCs producing invalid results, but if the owner is anonymous how can you contact him?
Tullio

Claggy
Claggy
Joined: 29 Dec 06
Posts: 560
Credit: 2694028
RAC: 0

RE: At SETI@home people

Quote:
At SETI@home people with GPUs produce thousands of invalid results since they are not using the right driver and since they do not visit the forums they are not aware of this.


There are many factors involved with those hosts producing invalids, not running a 'right driver' isn't really one of them (at least for the Main project, Beta is different)
take your pick of: Faulty GPU, overheating CPU or GPU, inadequate PSU for their GPU, incorrect app for their GPU, etc.
Running the wrong driver for Keplar GPUs produces errored tasks, not invalids, but it could be argued that they are running the wrong app there and not the wrong driver,
(because the project hasn't supplied a keplar compatible Cuda app yet, and are using a Cuda 3.0 app that requires a environment variable to be manually set to enable compatiblity)

Claggy

Horacio
Horacio
Joined: 3 Oct 11
Posts: 205
Credit: 80557243
RAC: 0

RE: I understand what you

Quote:
I understand what you are saying but disagree based on a persons money...a person buys a pc for x amount of dollars, they spent hard earned money on that pc and want it to last as long as it can, not replace it in two years because something else came out that is faster and better...


But Ive not said that... what I mean with old and inefficient was not just one generation older, neither has anything to do with their speed, my point about inefficient was aimed at power consumption per work done... So, if a 10 years old computer is still good for somebody needs, he wont need to change it... my point was to discourage its usage as a 24/7 cruncher... Off course, setting the line to split what is old and inefficient from what is not might become in an endless discussion... And could give a lot of frustration to those having that hardware that is first after the "no-go" line...

But also, this thing about discouraging old computers was just one line of thinking oppossed to the other line in which Ive said that due to the nature of BOINC, projects should support almost any hardware...
And in other lines Ive said that the project's point of view about where the results come from should be weighted when choosing to give up "support" for legacy hardware...

In any case, your disagreement about discouraging inefficient hardware based on people's economies, is still another valid point of view.

mikey
mikey
Joined: 22 Jan 05
Posts: 11889
Credit: 1828213831
RAC: 201898

RE: At SETI@home people

Quote:
At SETI@home people with GPUs produce thousands of invalid results since they are not using the right driver and since they do not visit the forums they are not aware of this. There is a thread dedicated to people who PM the owners of PCs producing invalid results, but if the owner is anonymous how can you contact him?
Tullio

I thought they were using software that slowly reduced the pc's ability to get new work based on the idea that they were only producing junk anyway? Did they stop using it?

There was a study once saying that only like 4% of crunchers ever even visit the forums.

As for contacting the user the Admins should be able to, at the very least they can send the user an email, and if it continues they can stop their ability to get any more units. The project must perceive it as a problem though, and if it is only 1% of the total users it may not be worth the time to worry about it. Although it does make for extra work for the Server.

tullio
tullio
Joined: 22 Jan 05
Posts: 2118
Credit: 61407735
RAC: 0

SETI@home is managed by three

SETI@home is managed by three people working part time. One of them, Matt, is a traveling musician. They don't have the time to write to people. This is done by volunteers, but they cannot PM the anonymous.
Tullio

Tonyina
Tonyina
Joined: 9 Nov 12
Posts: 1
Credit: 4727373
RAC: 0

RE: One of the success of a

Quote:
One of the success of a project is the number of users who allready joint the project, not the power of his hosts,

I believe the success is about getting the job done, the more and the quicker the better, not about having a club with lots of members.

It's great to have millions of people doing the same stuff as you do, but efficiency rules. Since the sooner the job is done the sooner the project will reach its real goal.

transient
transient
Joined: 3 Jun 05
Posts: 62
Credit: 115835369
RAC: 0

RE: RE: One of the

Quote:
Quote:
One of the success of a project is the number of users who allready joint the project, not the power of his hosts,

I believe the success is about getting the job done, the more and the quicker the better, not about having a club with lots of members.

It's great to have millions of people doing the same stuff as you do, but efficiency rules. Since the sooner the job is done the sooner the project will reach its real goal.

True, but not having the money and/or the people to do the job you want to do, does not really help. And essentially that is the situation SETI@Home is in, IMO.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.