Welcome to MilkyWay@home

credit comparison to other projects

Message boards : Number crunching : credit comparison to other projects
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 12 · 13 · 14 · 15

AuthorMessage
Profile Travis
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 30 Aug 07
Posts: 2046
Credit: 26,480
RAC: 0
Message 9631 - Posted: 3 Feb 2009, 22:16:19 UTC - in response to Message 9610.  

What I'm thinking we're going to do -- depending on how much we can optimize on our side. Is provide our own "stock" optimized applications that will be made available for download here.

That way, since the project will be providing optimized applications, we'll be able to drop the credit down to a reasonable rate, and lift the credit limit.

The basic stock will still have to be the most accessible and not optimized with the best SSE + whatever, so therefore run a bit slower, but since users will have the opportunity to upgrade to an optimized sock app, hopefully it wont be too big of a deal.
ID: 9631 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile GalaxyIce
Avatar

Send message
Joined: 6 Apr 08
Posts: 2018
Credit: 100,142,856
RAC: 0
Message 10278 - Posted: 10 Feb 2009, 23:32:20 UTC

I've been taking a look at some of my credits returned recently.
I'm not sure if this makes for any valid comparisons, but these
are my calculations for credits returned per CPU sec (higher is better);

Project.......CPU sec.....Credit granted.....Credits per sec

GPU grid.....4,071.28.......2,478.99..........0.61

MilkyWay......509.75...........15.29...........0.03 (using optimized apps)

Cosmology...32,653.59.......140.00..........0.004

CPDN........1,067,728.33...5,184.00.........0.005



ID: 10278 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Cluster Physik

Send message
Joined: 26 Jul 08
Posts: 627
Credit: 94,940,203
RAC: 0
Message 10279 - Posted: 11 Feb 2009, 0:31:02 UTC - in response to Message 10278.  

I've been taking a look at some of my credits returned recently.
I'm not sure if this makes for any valid comparisons, but these
are my calculations for credits returned per CPU sec (higher is better);
[..]

Which CPU?
To have some more samples I add some values for a 45nm Core2@3GHz:
POEM: 1,940.47s, 32.48 credits, 0.017cr/s (average WU, varies wildly between 1,400 and 3,000s)
Spinhenge: 1,772.81s, 13.0 credits, 0.0073cr/s (quite on the low end of the spectrum)
QMC: 6,126.11s, 63.93 credits, 0.0104 cr/s
MW: 258.14s, 7,74 credits, 0.03 cr/s with optimized app (SSE4.1)

When the credits are lowered like I suggested, MW would still give 6.93credits/258.14s = 0.0268 cr/s with an optimized app. Quite healthy if you ask me. And the good thing is you can abandon the credit limit so a faster CPU will finally get more credits then a slower one.
I have to test, how a generic x87 compile would do. I guess it would be somewhere in the range between QMC or POEM then, so completely okay.
ID: 10279 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Stevea

Send message
Joined: 14 Jul 08
Posts: 50
Credit: 8,398,033
RAC: 0
Message 10280 - Posted: 11 Feb 2009, 2:29:23 UTC
Last modified: 11 Feb 2009, 2:51:28 UTC

You know what the project has been barely up for 24hs without a problem, why don't we wait a while before adjusting the credits again.

Between the optimized apps, the GPU app, and the stock app, there is no way of being able to adjust the credit to a fair amount for all (hardware) systems, and Operating Systems just yet.

If you all are so gun ho for the cross project pairity thing, then may I suggest that you go hound the 95% of these projects that are giving out less credit than the seti stock app.I have yet to see even one time that the credit police have gotten a project to raise their credit. They just go around and pound on a project until they are beaten into submission and the list below is the results.

When you can get even 50% of the 90% that give less to give a reasonable amount then I will agree with you that this project is too high. Until then, I use this project to even the field, and gives me the opportunity to crunch other projects and get around the same credit as if there was parity.

Here's the link: http://boinc.netsoft-online.com/e107_plugins/boinc/get_cpcs.php

Here's the list : Go sick'em

The almighty SETI.........1.000

Einstein@Home.............0.842
climateprediction.net.....0.838
LHC@home...................0.619
Rosetta@home.............0.707
BURB.............................0.439
PrimeGrid......................0.709
QMC@HOME..................0.833
SZTAKI Desktop Grid......0.913 .... .9 omg, you better go beat this one down
World Community Grid...0.517
uFluids...........................0.507
malariacontrol.net.........0.559
The Lattice Project........0.457
SIMAP............................0.750
Pirates@Home...............0.402
SETI@home Beta...........1.387 .... Of course this is OK, because it's SETI
Leiden Classical.............0.474
RALPH@Home................0.557
cpdn seasonal...............1.355 ... omg, better get the police over there now
vtu@home.....................0.444 ... much better
Nano-Hive@Home..........0.737
Spinhenge@home..........0.563
Docking@Home..............0.360 ... this is a joke right? No good science here
BOINC alpha test...........0.403
ABC@home....................0.595
WEP-M+2.......................0.796
orbit@home...................0.338
PS3GRID........................1.655 ... can't have this, go get'em
Cosmology@Home.........1.123 ... 1.123 can't have this
BRaTS@Home................0.304
Hydrogen@Home...........0.274 ... lol . just pathetic
Superlink@Technion......0.615
Enigma@Home..............0.495
yoyo@home..................0.684
Eternity2.net.................0.370 ... much better
SHA-1 Collision
Search Graz...................0.626
Sudoku..........................0.609
NQueens Project...........0.449
POEM@HOME.................0.609
Intelligence Realm.........0.771
MilkyWay@home...........1.107 ... omg the sky is falling
MindModeling@Home....0.411 ... that's more like it
3x+1@home.................1.085 ... can't have this now can we
Cels@Home..................0.707
Magnetism@home........0.411
Genetic Life..................0.438
pPot Tables..................0.677
AQUA@home................0.663

43 projects giving less credit than Seti

6 giving more and one of them is seti beta, and we are not even talking about the regular SETI with the opto app here either, which would bring the number of SETI projects giving more than the regular app to 2

If you really fell outraged it should be because of this Hydrogen@Home 0.274 ... lol thats just pathetic

So go and get even 10 of those to raise their credit and we can talk about parity to the almighty SETI

Until then just leave this (ALPHA) project alone!

If not there are some projects there that are giving out .7 or .8 might want to go beat them down some more. That way everyone will only crunch SETI, and we know how well there servers are handling the load right now, lets see what happens if everyone only crunches the same SETI results 150 each times just to keep everyone with work.
ID: 10280 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Kevint
Avatar

Send message
Joined: 22 Nov 07
Posts: 285
Credit: 1,076,786,368
RAC: 0
Message 10281 - Posted: 11 Feb 2009, 2:55:56 UTC - in response to Message 10280.  
Last modified: 11 Feb 2009, 2:57:56 UTC

Credit comparisons should be based on the stock apps only.

All these stats using the optimized app to compare credits with is well,,,, retarded.


If you want to compare then compare stock app to stock app.

And all this crazy talk about cross project parity, - well, it simply does not work.
.
ID: 10281 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Stevea

Send message
Joined: 14 Jul 08
Posts: 50
Credit: 8,398,033
RAC: 0
Message 10282 - Posted: 11 Feb 2009, 3:02:02 UTC

I believe that was my point.

And until I see the police get even one project to raise credit, the parity discussion is OVER.
ID: 10282 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Gavin Shaw
Avatar

Send message
Joined: 16 Jan 08
Posts: 98
Credit: 1,371,299
RAC: 0
Message 10283 - Posted: 11 Feb 2009, 3:03:00 UTC

Stevea,

Thanks for the list. I had always wondered how all the projects compared to Seti. Of course I could have gone looking at BoincStats but I guess I'm just a little bit lazy.

Though I would have thought Rosetta was lower than 0.7 or at least that is what it feels like when I crunch for it.

But you have a point. High projects are hounded about lowering to be roughly in line, but I never here of low paying projects being 'hounded' or encouraged to raise their level to come in line...

Never surrender and never give up. In the darkest hour there is always hope.

ID: 10283 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Cappy [Team Musketeers]
Avatar

Send message
Joined: 3 Oct 07
Posts: 71
Credit: 33,212,009
RAC: 0
Message 10285 - Posted: 11 Feb 2009, 3:11:26 UTC - in response to Message 10281.  

Credit comparisons should be based on the stock apps only.

All these stats using the optimized app to compare credits with is well,,,, retarded.


If you want to compare then compare stock app to stock app.

And all this crazy talk about cross project parity, - well, it simply does not work.


hear hear !!!!!!!!!!!!!!!!!!!!!

how can you compare the projects when the sciences behind the projects

are TOTALY diffrent from 1 to another. i feel if your machines can do the work

faster then soemone elses then you DAMN right you should be rewarded for it and not punished because some of the others are still running slow a$$ p3's or p4 single cores.... why we bigger better machines have to be downgraded to their level is beyond my comprehension. i worked Heavy construction all my life. now if a job was bidded for 1 month to complete and i could get it done in 2 weeks i got a BONUS for saving the company time and money.....

but now and here we have a SPEED LIMIT so us faster guys cant go as fast as we can why????? becasue the slower guys feel its unfair?????

i duno maybe i got it all wrong but i dont think holding back your faster crunchers is good for the project and or the science... the MORE science done in a shorter amout of time is a win win situation isnt it???


if im wrong i apologize was just sitting here drinking a few beers and all this came to me lol,,,,, whether im right i dont know and if im wrong then it will be a learning experience.

Cappy out

night all
ID: 10285 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
jedirock
Avatar

Send message
Joined: 8 Nov 08
Posts: 178
Credit: 6,140,854
RAC: 0
Message 10287 - Posted: 11 Feb 2009, 3:30:17 UTC - in response to Message 10280.  

You know what the project has been barely up for 24hs without a problem, why don't we wait a while before adjusting the credits again.

Between the optimized apps, the GPU app, and the stock app, there is no way of being able to adjust the credit to a fair amount for all (hardware) systems, and Operating Systems just yet.

If you all are so gun ho for the cross project pairity thing, then may I suggest that you go hound the 95% of these projects that are giving out less credit than the seti stock app.I have yet to see even one time that the credit police have gotten a project to raise their credit. They just go around and pound on a project until they are beaten into submission and the list below is the results.

When you can get even 50% of the 90% that give less to give a reasonable amount then I will agree with you that this project is too high. Until then, I use this project to even the field, and gives me the opportunity to crunch other projects and get around the same credit as if there was parity.

Are those on 32- or 64-bit machines? Because I know ABC grants credit closer to Milkyway on my 64-bit crunchers. 32-bit machines crunch twice as long, but ABC doesn't grant them any more credits, because their 64-bit app can do the same calculations in half the time using 64-bit ints.
ID: 10287 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Misfit
Avatar

Send message
Joined: 27 Aug 07
Posts: 915
Credit: 1,503,319
RAC: 0
Message 10289 - Posted: 11 Feb 2009, 3:52:58 UTC - in response to Message 10280.  

I have yet to see even one time that the credit police have gotten a project to raise their credit.

Here's the link: http://boinc.netsoft-online.com/e107_plugins/boinc/get_cpcs.php

Here's the list : Go sick'em

The almighty SETI.........1.000

uFluids...........................0.507

What about the anti-police? Thanks for this useful info. I have word when uFluids is ready to go on its new server the credits will be adjusted accordingly. I've recommended doubling them. :D
me@rescam.org
ID: 10289 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Kevint
Avatar

Send message
Joined: 22 Nov 07
Posts: 285
Credit: 1,076,786,368
RAC: 0
Message 10290 - Posted: 11 Feb 2009, 4:07:55 UTC - in response to Message 10287.  


Are those on 32- or 64-bit machines? Because I know ABC grants credit closer to Milkyway on my 64-bit crunchers. 32-bit machines crunch twice as long, but ABC doesn't grant them any more credits, because their 64-bit app can do the same calculations in half the time using 64-bit ints.


Not always the case - I have a couple boxes that I use on ABC and not here - Why?

Speed limit.

On ABC my 9640's are seeing 13-15K per day, here- 10K.

the newer i7's with the same overclock should do much better than 13-15K per day on ABC - here- 2500 per core per day.

Of course, this is also using an optimized MW app, so it is not quite apples to apples since ABC does not have an optimized app.


uFluids has an application problem, the more cores your unit has, the worse it is. A 8core box thrashes horribly, and gets much less work done than per core than an old Pent D duo. Wanna run uFluids with the current app, set your active cores to no more than 2.
.
ID: 10290 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Stevea

Send message
Joined: 14 Jul 08
Posts: 50
Credit: 8,398,033
RAC: 0
Message 10293 - Posted: 11 Feb 2009, 6:10:02 UTC - in response to Message 10289.  
Last modified: 11 Feb 2009, 6:14:20 UTC

I have yet to see even one time that the credit police have gotten a project to raise their credit.

Here's the link: http://boinc.netsoft-online.com/e107_plugins/boinc/get_cpcs.php

Here's the list : Go sick'em

The almighty SETI.........1.000

uFluids...........................0.507

What about the anti-police? Thanks for this useful info. I have word when uFluids is ready to go on its new server the credits will be adjusted accordingly. I've recommended doubling them. :D


And then the other 42 projects?

These would be a good place to start, don't you think....

Hydrogen@Home...........0.274 ... lol . just pathetic
Docking@Home..............0.360 ... this is a joke right? No good science here
orbit@home...................0.338

Go ahead and get your magic wand out.

LOL... even BOINC's own Beta is pathetic: BOINC alpha test...........0.403
ID: 10293 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
jedirock
Avatar

Send message
Joined: 8 Nov 08
Posts: 178
Credit: 6,140,854
RAC: 0
Message 10294 - Posted: 11 Feb 2009, 6:18:10 UTC - in response to Message 10293.  

LOL... even BOINC's own Beta is pathetic: BOINC alpha test...........0.403

Yeah, well BOINC alpha test is really just for the alpha versions of BOINC, so they can download WUs that can basically be screwed over and have no effect. Just sends out work for the uppercase example app. It's not meant as a full-time project.
ID: 10294 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Stevea

Send message
Joined: 14 Jul 08
Posts: 50
Credit: 8,398,033
RAC: 0
Message 10295 - Posted: 11 Feb 2009, 6:35:56 UTC - in response to Message 10294.  
Last modified: 11 Feb 2009, 6:37:59 UTC

LOL... even BOINC's own Beta is pathetic: BOINC alpha test...........0.403

Yeah, well BOINC alpha test is really just for the alpha versions of BOINC, so they can download WUs that can basically be screwed over and have no effect. Just sends out work for the uppercase example app. It's not meant as a full-time project.


LOL: Well they could give more credit than that, since they are the parity nutjobs.

They can't even keep their own project in line with parity, but expect everybody else to...... nice...
ID: 10295 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
David @ TPS

Send message
Joined: 20 Nov 08
Posts: 24
Credit: 2,561,361
RAC: 0
Message 10298 - Posted: 11 Feb 2009, 13:55:25 UTC
Last modified: 11 Feb 2009, 14:04:17 UTC

Parity - Schmarity...... THE LAST TIME I LOOKED THIS PROJECT WAS NOT AT BERZERKELY

Lets just call a spade a spade.

Some people crunch just for the science and that is great!

Some people crunch for the competition and that is fine.

I am a little of both, but more for the competition.

What that boils down to is this. I have several computers of various ages and horsepower. I set goals for myself and strive for them. I am going to allocate my horsepower where it will help my goals, pure and simple. Right now, SETI pays well, so most is there. A couple on Einstein to reach 3 million. 3 on Rosetta which pays NOTHING 13K seconds pays 12 credits. THEN I found MW! This project pays well, even on the P-4's, it helps my Combined Boincstats, and I believe that the project is worthwhile. WOW! I started here with a P-4 and a XEON, left when the credits got slashed, then, when the opt-ap was reinstated, I came back with the P4, a P-D and 2 of my quads, which can all do max work per day. I get NO penalty for taking time away from SETI, and if I do not get MW, I have plenty of other work. I compete for world ranking on boincstats, but mainly with myself. Can I find just the right combination of projects and tweak that extra few flops out of a box somewhere? As long as the credits on MW remain where they are, or even rise, the rest of my acreage (small farm) will be coming to MW. Who benefits? ME and THE PROJECT... a win/win.

MW has had problems of late, so has SETI. So what? Any project trying to improve is gonna have more bugs than a bait shop! If your project is worth crunching (aka pays well and is science worthy) your crunchers will STILL come back after an outage, no matter how long. I have no problem with the 12/core limit. My boxes just switch back to whatever else they are crunching and wait for the next d/l.

*edit* -- Travis, you (and everyone else) are doing well! Just remember, it is YOUR project, not D.A.'s. If a company wants workers, they have to pay to get them, and if they want HARD workers, they have to pay more than minimum wage. You have a good project starting, you need crunchers, so chase the moths out of your "credit wallet" and hire some good help!

*slips climbing off soap box, twisting ankle*

Now, I am going to hobble out and attach another box or 2 to MW.
ID: 10298 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Dan T. Morris
Avatar

Send message
Joined: 17 Mar 08
Posts: 165
Credit: 410,228,216
RAC: 0
Message 10306 - Posted: 11 Feb 2009, 18:23:58 UTC

Here is a very great idea, leave the credits where they are now........

And the world will be happier for it.



ID: 10306 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile GalaxyIce
Avatar

Send message
Joined: 6 Apr 08
Posts: 2018
Credit: 100,142,856
RAC: 0
Message 10378 - Posted: 12 Feb 2009, 7:52:09 UTC - in response to Message 10285.  

i feel if your machines can do the work

faster then soemone elses then you DAMN right you should be rewarded for it and not punished because some of the others are still running slow a$$ p3's or p4 single cores.... why we bigger better machines have to be downgraded to their level is beyond my comprehension. i worked Heavy construction all my life. now if a job was bidded for 1 month to complete and i could get it done in 2 weeks i got a BONUS for saving the company time and money.....

but now and here we have a SPEED LIMIT so us faster guys cant go as fast as we can why????? becasue the slower guys feel its unfair?????

It's a good question. What I would say is that I have my share of old P4s contributing to MW ;)


ID: 10378 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Misfit
Avatar

Send message
Joined: 27 Aug 07
Posts: 915
Credit: 1,503,319
RAC: 0
Message 10379 - Posted: 12 Feb 2009, 8:06:24 UTC - in response to Message 10378.  

It's a good question. What I would say is that I have my share of old P4s contributing to MW ;)

My old P4 is at the local landfill. :/ Who knew.
me@rescam.org
ID: 10379 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile GalaxyIce
Avatar

Send message
Joined: 6 Apr 08
Posts: 2018
Credit: 100,142,856
RAC: 0
Message 10425 - Posted: 12 Feb 2009, 18:34:39 UTC - in response to Message 10379.  

It's a good question. What I would say is that I have my share of old P4s contributing to MW ;)

My old P4 is at the local landfill. :/ Who knew.

I already had the HD, RAM and power supplies stripped out of my old P4s for possible reuse and the rest in a garage corner ready for the tip. When I saw how many credits there were to be had with the opp apps I managed to rebuild four of them with the bits. (Don't you just hate it when you turn off an old PC, and it never works again, not matter what you do with it?)

I never thought I would catch up or exceed my SETI credits so quickly. At the time of writing MilkyWay is about to become my highest scoring project.



ID: 10425 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Previous · 1 . . . 12 · 13 · 14 · 15

Message boards : Number crunching : credit comparison to other projects

©2024 Astroinformatics Group