21) Message boards : Number crunching : Work almost done? (Message 12328)
Posted 23 Jan 2006 by borandi
Post:
HM, I'm participating in nearly all BOINC-Projects.

When LHC sends work, I switch all other projects to "no new work". So, if LHC has work, my boxes concentrate on LHC :-)


Me too =D Mind you, there is now less than 15k WU left :'(
22) Message boards : Number crunching : 362,000 secs per 1,000,000 turns (Message 12321)
Posted 23 Jan 2006 by borandi
Post:
http://lhcathome.cern.ch/workunit.php?wuid=973422

Gotta love those P133's taking almost 4 days and 5hrs just to do 1 large WU.



Good heavens! You actually have Windows XP on that?! Is it truly installed or part of some kind of cluster?


It is truly installed on a 3.2GB HDD.


you might get better throughput with win98 or 95 - a typical P133 has rather small memory for the huge needs of XP, and a lot of that proc time os almost certainly swapping operating system in and out whenever the app makes a system call


My CPU effieciency on an LHC WU of 95.15% begs to differ =D I think in that machine there is 96MB RAM, and it runs fine as long as I only access it through boincView. Note, XP only requires 64MB.
23) Message boards : Number crunching : 362,000 secs per 1,000,000 turns (Message 12320)
Posted 23 Jan 2006 by borandi
Post:
That's pretty funny and it's not even due to the W98 problem that causes the cpu time to increase even when a wu is preempted. It would be better off doing SETI@home and using the optimised SETI application, especially since it gets less granted credit than claimed.

Live long and crunch.


Every time it tries a SETI WU, after 50hrs it always seems to have a computational error. ATM its either on LHC or SZTAKI :D
24) Message boards : Number crunching : 362,000 secs per 1,000,000 turns (Message 12282)
Posted 22 Jan 2006 by borandi
Post:
http://lhcathome.cern.ch/workunit.php?wuid=973422

Gotta love those P133's taking almost 4 days and 5hrs just to do 1 large WU.

Next time LHC has work it may do one again :D

25) Message boards : Cafe LHC : More about light (Message 12268)
Posted 22 Jan 2006 by borandi
Post:
Also, with regards to redshift:



Prediction: light escaping from a large mass should lose energy---the wavelength must increase since the speed of light is constant. Stronger surface gravity produces a greater increase in the wavelength.

This is a consequence of time dilation. Suppose person A on the massive object decides to send light of a specific frequency f to person B all of the time. So every second, f wave crests leave person A. The same wave crests are received by person B in an interval of time interval of (1+z) seconds. He receives the waves at a frequency of f/(1+z). Remember that the speed of light c = (the frequency f) × (the wavelength l). If the frequency is reduced by (1+z) times, the wavelength must INcrease by (1+z) times: lat B = (1+z) × lat A. In the doppler effect, this lengthening of the wavelength is called a redshift. For gravity, the effect is called a gravitational redshift.

Observation: spectral lines from the top layer of white dwarfs are significantly shifted by an amount predicted for compact solar-mass objects. The white dwarf must be in a binary system with a main sequence companion so that the amount the total shift due to the ordinary doppler effect can be determined and subtracted out. Inside a black hole's event horizon, light is redshifted to an infinitely long wavelength.


26) Message boards : Cafe LHC : More about light (Message 12267)
Posted 22 Jan 2006 by borandi
Post:
For clarification:

My point here is trying to find how tightly time and space are connected to each other.

Roughly worded:
That, at the speed of light, does all space get covered in one time(instant) and
at the speed of stopped, does one space see all time?

A source is:
http://www.astronomynotes.com/relativity/s1.htm


From that site:


Since the laws of physics do not depend on your location or motion, Einstein reasoned that the speed of light will be measured to be the same by any two observers regardless of their velocity relative to each other. For example, if one observer is in a rocket moving toward another person at half the speed of light and both observers measure the speed of a beam of light emitted by the rocket, the person at rest will get the same value the person in the rocket ship measures (about 300,000 kilometers/second) instead of 1.5 times the speed of light (=rocket speed + speed of beam of light). This assumption has now been shown to be correct in many experiments. To get the same value of the speed (= distance/time) of light, the two observers moving with respect to each other would not only disagree on the distance the light travelled as Newton said, they would also disagree on the time it took.

Einstein found that what you measure for length, time, and mass depends on your motion relative to a chosen frame of reference. Everything is in motion. As you sit in your seat, you are actually in motion around the center of the Earth because of the rapid rotation of the Earth on its axis. The Earth is in motion around the Sun, the Sun is in orbit around the center of our Galaxy, the Galaxy is moving toward a large group of galaxies, etc. When you say something has a velocity, you are measuring its change of position relative to some reference point which may itself be in motion. All motion is relative to a chosen frame of reference. That is what the word ``relativity'' means in Einstein's Relativity theories. The only way observers in motion relative to each other can measure a single light ray to travel the same distance in the same amount of time relative to their own reference frames is if their ``meters'' are different and their ``seconds'' are different! [b]Seconds and meters are relative quantities. [b]


The fact that as you approach the speed of light, not only does time change, but distance changes exponentially.

This is assuming you mean 'space' as 'distance' - I define space as being that black stuff around the planets and stuff!

Dont forget there are two types of relativity: 'Special Relativity' and 'General Relativity'. Under General Relativity, Einstein concluded that gravity = acceleration: http://www.astronomynotes.com/relativity/s3.htm . Actually that page has good explanation of space-time :D
27) Message boards : Number crunching : Boinc farms. (Message 12266)
Posted 22 Jan 2006 by borandi
Post:

Being part of this is whats important.
Whether its a Pentium Pro 100 or a quad dual core opteron......
.........fire up that BOINC and get crackin!!!


My P133 racks up 13 cred/day, where my AMD X2 4400+ gets 650. Nevertheless, I'll still spend time on that P133 to get it working :D

I've still got a P166 that wont accept an operating system.... all Linux does is give errors :'(
28) Message boards : Number crunching : Loadsa Work (Message 12249)
Posted 22 Jan 2006 by borandi
Post:
I having 50/50 short/long WU's - with my slowest computer needing 32hrs for 1,000,000 turn calculation.

But this lot of WUs wont take long - lots of people find this a worthwhile project to crunch on - me included!!!
29) Message boards : Number crunching : Pirates ahoy! (Message 12248)
Posted 22 Jan 2006 by borandi
Post:
Aye aye! More Pirate pieces of eight! (thats more WUs)
30) Message boards : Number crunching : Work almost done? (Message 12246)
Posted 22 Jan 2006 by borandi
Post:
The historical pattern is to have "waves" of work, where a couple hundred thousand work units will be created and mostly issued, then another wave, if any, is added. After all the main waves have been issued and mostly returned, the "clean up" runs for a little bit where smaller batches are scavenged of work that is still incomplete (no quorum).


It leaves those who have done work waiting some weeks to pick up those last credits :D
31) Message boards : Cafe LHC : More about light (Message 12245)
Posted 22 Jan 2006 by borandi
Post:
Now here's my interpretation of red-shift, and this was what I was taught in high school physics:

Red shift is visible when a light emitting source (i.e. a star) is moving away from the view of the detector (i.e. a human) as it elongates the wavelength of the electromagnetic radiation.

As a result, a photon parallel to another photon, both in a vacuum, will see each other at the wavelength they are emitted (imagine they have eyes and can see the photons moving).

Given that time is relative to photons, it is different from another reference point. To say time is constant across the universe would indicate that you were in all places at once.

So:

1) Ok, time stops; For the photon, relative to the photon.
2) Yet, it does change. It gets red-shifted.

A1) Yes, I would agree. Only if the photon was purely observing the other photon. If it observed a different reference frame, then time would change.
A2) This doesnt comply with what I've been taught or any scientific journal I have read. Have a link/reference?

--borandi
32) Message boards : Cafe LHC : LHC@home needs more publicity (Message 12244)
Posted 22 Jan 2006 by borandi
Post:
Heh the rate that these WU's are being churned through, its not publicity that is needed, its more WU's! My pending credit keeps rising with the 8 machines I have running, but given that 2 weeks ago there were 500,000 WUs left and now only ~100000 it looks like we crunchers will go quicker than the scientists can!
33) Message boards : Number crunching : Boinc farms. (Message 12093)
Posted 16 Jan 2006 by borandi
Post:
I'm keeping an eye on the Playstation 3. It's supposed to be available with a hdd and linux on it, so even if the compilers are not top notch, it could just crunch 8 (7) different tasks and should have one hell of a throughput. That would temp me to get another box again :D


Getting a PS3 would be the cheapest crunch option available. At CES they announced that of the Cell Processors eight cores (they say eight, its prob lots of hyper threading), seven will be available to the PS3 due to architecture.

Assuming they come out at no less than 300 quid (~$550) it'd be great as a cruncher.
34) Message boards : Number crunching : Download failures (Message 12017)
Posted 15 Jan 2006 by borandi
Post:

Did it work?



Worked for me :D
35) Message boards : Cafe LHC : BOINC and Public Universities (Message 12007)
Posted 15 Jan 2006 by borandi
Post:
I've tried here at the University of Hull, UK.

Several problems:

Security - The uni will only have programs that if they access the internet they need appropriate security certificates.
Bandwidth - Lots of computers x lots of crunching = slow network. (funnily enough that didnt stop several students d/l 10gb/day on the uni network)
Cost - Having 3000+ computers on full tilt 12hrs a day is more power consuming than 3000+ idle ones. Universities are strapped for cash as it is.

This is coming from a student - I have to apply to the uni for them to sell me their old computers for my BOINC farm. I'm glad that electricity is included in my rent costs!

BOINC@Hull - the BOINC project/farm I run
36) Message boards : Number crunching : Boinc farms. (Message 12005)
Posted 15 Jan 2006 by borandi
Post:
Just added me new AMD X2 4400+ to me farm :D

Its going through 300 creds/processor/day = 600 creds/day. ATM all on CPDN until I can get that LHC file d/l!

BOINC@Hull
37) Message boards : Number crunching : Download failures (Message 11990)
Posted 14 Jan 2006 by borandi
Post:
I have the same prob on 6 of my machines with that file. :( And I've downloaded ~200 WU between them.
38) Message boards : Number crunching : Boinc farms. (Message 11692)
Posted 19 Dec 2005 by borandi
Post:
Falc - thats a mean setup. I'll put it on my christmas list!
39) Message boards : Number crunching : Boinc farms. (Message 11686)
Posted 16 Dec 2005 by borandi
Post:
Borandi, if your "main" machine is running Windows, you really should have a look at BoincView
it may take a couple of minutes to set up, but it makes the thing a lot easier when you're running multiple "drones" on a network.


Haha I just set it up, then came here and I find this post! BoincView is great - now I dont have to keep switching the monitor between units and keep clicking upload. Fantastic!
40) Message boards : Number crunching : Boinc farms. (Message 11677)
Posted 15 Dec 2005 by borandi
Post:
How bad is your boinc farm???


Lets see..

AMD 2400+ (1.8Ghz) - laptop that overheats just on iTunes
P3 500Mhz
P2 266Mhz
P1 133Mhz (takes 50h per SETI wu)

Thats not to mention I've got my girlfriend using it on her laptop (1.5Ghz).

Ahh... when my AMD X2 4400+ arrives.... then the crunching begins!

The local council here collects old computers from businesses and schools and stuff, and uses them to teach IT Engineers how to fix them and stuff, then sells them on really cheaply to local projects (or donates). I'm hoping to get some of them for BOINC@Hull. I'm not sure what systems they may give - anything from 100Mhz to 1Ghz I think.

Hmm... now I need a networking monitor which is easy to use...


Previous 20 · Next 20


©2024 CERN