World Community Grid Crunch July 2008

Post by **UBT - Mikee** » Sat Jul 05, 2008 6:19 pm

As we've just been overtaken by 'UK' and 'TSBT' are also hot on our heals I suppose we better do something about it! It's also one of our worst performing projects where we're currently 116 in the world!

It's several months since we had a 'Crunchathon' so this is as good a time and project to do one on, especially as we're not participating in anything at the moment.

Someone has registered us for the following World Community Grid challenges...

MUSCULAR DYSTROPHY CRUNCH ends 31/7/2008
ALL WCG PROJECTS AND ALL TEAMS ends 1/8/2008

so I suppose we can use them in conjunction with our own crunch.

For this crunch there will be 2 targets...

1 - Regain UK #1 status
2 - Be the first UK team to get 5 million credits, we currently have 4,752,224 - 235,126 to get!

Pros and cons

Pros

It's a very worthwhile project with several subprojects to select...

AfricanClimate@Home
• Please review the system requirements before opting to participate in this project - this is a climate project so if there is work available be prepared to tie up the PC for weeks!
Discovering Dengue Drugs - Together
FightAIDS@Home
Help Conquer Cancer
Human Proteome Folding 2
Nutritious Rice for the World

You can do one, some or all if your PC can take it and there's work available of course!

Cons

Unfortunately there are a couple!
1 - For the credits hogs among us (that includes me!) the credits are, at best, average.
2 - Most of the projects have a large quorum - Nutritious Rice has a quorum of 10, Human Proteome Folding - Phase 2 has 15. This is balanced by having an initial replication of 19. Help Conquer Cancer is 2 and 2. So it can take a few days for your first results to come through.
3 - The project times can be quite long - most are 8 or more hours for this average PC.

This crunchathon that will take us to the end of the month(!) with a break 18 July 2008 00:00 UTC and end 21 July 2008 00:00 UTC for anyone who wants to do the PrimeGrid Lunar Landing Challenge.

WCG main page for joining

UK position.

World position.

Starting stats...

18:15 5th July 2008

World Position = 116th
UK position = 2nd
UK score = 4,764,874
UKBT score = 4,752,224

Credits required to overtake Team UK = 12,650
Credits required to get to 5,000,000 = 235,126

Start whenever you like!

UBT - Rick Horn · Post by **UBT - Rick Horn** » Sat Jul 05, 2008 7:13 pm

Muscular Dystrophy and African Climate are not available to download at the moment.

Post by **UBT-mark3346** » Sat Jul 05, 2008 8:40 pm

AfricanClimate@Home 74.52
Discovering Dengue Drugs - Together 70.48
FightAIDS@Home 38.10
Help Conquer Cancer 71.89
Human Proteome Folding 2 52.36
Nutritious Rice for the World 81.03

These are the global figures for credits per hour (divide by 7 for Boinc credits)

Rice gives the highest credits by almost 9%

UBT - Nogbo · Post by **UBT - Nogbo** » Sat Jul 05, 2008 10:03 pm

2 - Most of the projects have a large quorum - Nutritious Rice has a quorum of 10, Human Proteome Folding - Phase 2 has 15. This is balanced by having an initial replication of 19. Help Conquer Cancer is 2 and 2. So it can take a few days for your first results to come through.

Please excuse my density on this, but just what does this mean? :?

Post by **UBT-mark3346** » Sat Jul 05, 2008 10:21 pm

Before you get your credits a number of other results must be returned for the same wu - Quorum of one you get them straight away, Q of 10, 10 results must be returned before credits are granted etc.

melter65 · Post by **melter65** » Sat Jul 05, 2008 10:25 pm

For the science, it's a way of cross-checking the results

For the projects, it's a way of making sure we're not cheating

UBT - Nogbo · Post by **UBT - Nogbo** » Sat Jul 05, 2008 10:41 pm

Thanks gents, that is a lot clearer now :D

Post by **UBT-mark3346** » Sat Jul 05, 2008 10:42 pm

Rice is slightly different in that each wu runs 8 hours regardless of machine and although the quorum is 10 (was 14 see below) the wus are not identical to each other

Each result returned by the project contains unique data. We have some additional information in the results that we will use to help us make sure that the very strong majority of results returned contain the correctly computed data.

We are starting out sending 19 results per workunit and validating when 14 are returned. Each workunit will run for about 8 hours on each computer. A computer running this project will generate an additional structure prediction in about 90 seconds of execution time on a average computer.

This means that if you have a really powerful computer, you may generate 1000 structure predictions in 8 hours while a slower computer may only generate 160. We will be awarding credit by averaging the the claimed credit per structure prediction, and then taking that value and awarding the # of structure predictions in a result * the average claimed credit per structure prediction. So in this case if the slower computer had a claimed credit of 80 and the larger computer had a claimed credit of 450, then the average credit per structure prediction would be 0.475. This means that the slower computer would be granted 76 credits and the faster computer would be awarded 475 credits.

The key point though is that all computers should have a estimated time to completion of around 8 hours.

UBT - The Prof.... · Post by **UBT - The Prof....** » Sat Jul 05, 2008 10:52 pm

Have re-attached to this for the duration, though not one of my favorite projects

melter65 · Post by **melter65** » Sat Jul 05, 2008 10:56 pm

Same here, Prof. I'll run some 'Rice' wu's alongside some of the other, more obscure, projects I'm crunching at the moment

Joshrandom · Post by **Joshrandom** » Sat Jul 05, 2008 11:07 pm

Not one of my favs either, but I'll give it a few cores and see how it goes.

UBT - Rick Horn · Post by **UBT - Rick Horn** » Sat Jul 05, 2008 11:31 pm

Like the projects, but don`t like the credits. :(

UBT - Nogbo · Post by **UBT - Nogbo** » Sun Jul 06, 2008 8:10 am

Put this on two cores last night to see how it goes.

Post by **UBT - Mikee** » Sun Jul 06, 2008 10:04 am

Code: Select all

Start
BUK Credits = 4,764,874
UBT Credits = 4,752,224 
Diff        =    12,650

To get 5,000,000 = 247,776

World Posn  =  116

-------------------------------------------

Latest 0.5 days in

BUK Credits =  4,773,400
UBT Credits =  4,755,953
Diff        =  17,447 (Lost 4,797)

To get 5,000,000 = 244,047 (up 3,729)

World posn  =   116

-------------------------------------------

As expected, it's going to take a while for the first results to come through. Once they do I hope we'll see a big increase in output.

Below is a chart to show what is possible. The large peak at the end of last year for for the first WCG challenge. The top day was 32,768. I'm hoping we're going to top that by a large margin!

BOINCStats daily graph...

UBT - mickyb69 · Post by **UBT - mickyb69** » Sun Jul 06, 2008 6:55 pm

Wow 8 hrs for 76 credits. I don't like the rice

Guess the 3 gigs a slow computer .
I wish i had a fast one that would get me the 475 credits .
Mind you the 3.2's only asking for 116 :evil:

Post by **UBT-mark3346** » Sun Jul 06, 2008 7:13 pm

UBT - mickyb69 wrote:Wow 8 hrs for 76 credits. I don't like the rice
Guess the 3 gigs a slow computer .
I wish i had a fast one that would get me the 475 credits .
Mind you the 3.2's only asking for 116 :evil:

I get around 130 credits for the 8 hours on this:

AuthenticAMD
AMD Athlon(tm) 64 X2 Dual Core Processor 5600+ [x86 Family 15 Model 67 Stepping 3]

melter65 · Post by **melter65** » Sun Jul 06, 2008 10:16 pm

Is WCG known to favour the AMD processors?

Post by **UBT-mark3346** » Sun Jul 06, 2008 10:43 pm

melter65 wrote:Is WCG known to favour the AMD processors?

No more than any other projects as far as I know (My AMDs score better than my Intels on almost every project except for SAP and I think CPDN)

UBT - mickyb69 · Post by **UBT - mickyb69** » Sun Jul 06, 2008 11:19 pm

My amd 2000xp went to heaven :shock:
But even though it ran at 1600 meg it seemed to be better on most projects.
A good example was milkyway ...
3.2 dual (intel) running xp 64 did a WU in about 8 mins 12 seconds
AMD 2000xp (1600 ish ) on win xp did wu's in 9.12 .

On most projects it held its own against a celeron 2.7
So yeh i'd say AMd's are probs better at crunching :lol:

NaRyan · Post by **NaRyan** » Mon Jul 07, 2008 12:50 am

AMD's seems to be better at projects that are FPU based.
Intels seem to be better at projects that are Integer based.

I know on Milkyway the AMD's (when running linux x64) made the Intel Quad look silly, yet when running Riesel's LLR the Intel was a lot faster.

I think on ABC the Intel was faster than the AMD since ABC is an Integer based project.

fellie · Post by **fellie** » Mon Jul 07, 2008 8:27 am

On it all chips since mid-day on 5th July :D

Doing dddt, faah, and rice, as I have the cancer / folding badges.

Will switch back to all projects as and when I get the badges

Post by **UBT - Mikee** » Mon Jul 07, 2008 8:27 pm

Okay guys - I'm stating to panic!

TSBT are only a couple of thousand behind us and they scored 53,000 against our 17,000!

We need to get a rush on or we're dropping to 3rd!

Code: Select all

Start
UK Credits  = 4,764,874
UBT Credits = 4,752,224 
Diff        =    12,650

To get 5,000,000 = 247,776

World Posn  =  116

-------------------------------------------

Latest 1 day in

UK Credits  =  4,808,352
UBT Credits =  4,783,478 
Diff        =  24,874 (Lost 12,224)

To get 5,000,000 = 216,552 (up 3,729)

World posn  =   116

Post by **UBT - Mikee** » Tue Jul 08, 2008 6:41 pm

Well there's good news and bad news - unfortunately there's more bad news than good!

Good news - we gained a World place!

Bad news - we lost it again as TSBT overtook us with claymores waving, lifting their kilts and showing us their backsides as they stormed past!

Good news - we're not losing by as much as before as output is slowly rising.

Bad news - we're now 3rd in the UK listing!

This isn't going as well...

Code: Select all

UK Credits  = 4,764,874
UBT Credits = 4,752,224 
Diff        =    12,650

To get 5,000,000 = 247,776

World Posn  =  116

-------------------------------------------

Latest 2 day in

UK Credits  =  4,813,633
UBT Credits =  4,787,969 
Diff        =     28,664 (Lost 16,014)

To get 5,000,000 = 212,031 (up 35,745)

World posn  =   116

Post by **UBT - Mikee** » Wed Jul 09, 2008 9:19 pm

I'm sorta hoping that this is like an oil tanker and we'll eventually swing around and start making some headway as, at the moment, we're going backwards!

Code: Select all

UK Credits  = 4,764,874
UBT Credits = 4,752,224 
Diff        =    12,650

To get 5,000,000 = 247,776

World Posn  =  116

-------------------------------------------

Latest 3 days in

UK Credits  =  4,851,896
UBT Credits =  4,820,484 
Diff        =     31,412 (Lost 18,792)

To get 5,000,000 = 179,516 (up 68,260)

World posn  =   116

UBT - Nogbo · Post by **UBT - Nogbo** » Wed Jul 09, 2008 9:30 pm

Well I have had this running 24 hours a day on 2 cores since Saturday and so far I have got 1,103 Boinc credits :(

Not so much a crunch as a nibble :lol:

melter65 · Post by **melter65** » Wed Jul 09, 2008 10:20 pm

Ah! The wonders of WCG credits!

Post by **UBT - Mikee** » Thu Jul 10, 2008 8:21 pm

Well...

We're nearly matching 'UK'. We were just 900 behind them yesterday.

Code: Select all

UK Credits  = 4,764,874
UBT Credits = 4,752,224 
Diff        =    12,650

To get 5,000,000 = 247,776

World Posn  =  116

-------------------------------------------

Latest 4 days in

UK Credits  =  4,871,974
UBT Credits =   4,839,669 
Diff        =     33,350 (Lost 19,665)

To get 5,000,000 = 160,331 (up 87,445)

World posn  =   116

This is pants!

Not because there seems to be a lack of interest in this crunch but the credits are woefully poor. I'm down to a third of what I usually get :(
Reckon I'll stick with it until I get the badge or we get to 5 million - whichever comes sooner and get back to a project with better rewards!

UBT - The Prof.... · Post by **UBT - The Prof....** » Fri Jul 11, 2008 1:07 pm

Have two wu's, both on about 40%. Will keep going for now, but not really interested in this. Time would be better spent on other projects, and will be after these two finish :(