TAS members' thread February-March 2007: DPAD and ex-NanoHive@Home.

Page 13 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Rudy Toody

Rudy Toody

Diamond Member
Sep 30, 2006
4,267
421
126
Originally posted by: BlackMountainCow
Originally posted by: Rudy Toody
How about a math project? Like ABC@Home, SZTAKI, or RCN?
Science doesn't happen without mathematics.
Click to expand...

If you can REALLY show me that the math project is not just for the sake of math (like finding the largest prime) and that the math used actually CAN/WILL have some real life meaning (which a large prime does not) I'd be happy to also do some math projects. But be warned, so far no one was able to convince me to do math.
Click to expand...

This is a formal rebuttal to Christian Diepold's statement about the real life meaning of a math project.

My definition of Mathematics is the discovery and formal proof of a pattern and the presentation of an Arithmetic algorithm to be used by scientists to deal with that pattern in their own work. The scientist will have full confidence that the pattern will remain consistent within the limits of the proof.


I'm going to describe a handful of Distributed Computing projects currently in progress and my take on their pros & cons.

--------------
Seti@Home: Not practical, but good!

The laws of electrodynamics are the same throughout the universe, however the number of methods that can be used to communicate using those laws are infinite. Without knowing the underlying protocol used by the extraterrestrials to transmit the intelligent messages, it's unlikely we will ever identify them.

This project's primary contribution to science is Distributed Computing.

Author's Note: Even though I no longer process their work units, I do have an ongoing $20/month transferred from my bank account to help support them.

--------------
SIMAP: Practical and good!

This is a bookkeeping operation. It enables the classification and storage of biological data to be used by other scientists.

--------------
Einstein@Home: Practical and good!

This is a bookkeeping operation. It maps the locations of pulsars in the universe.

A recent discovery has indicated that all things that behave like pulsars are not pulsars, so this project will provide the coordinates to the objects that require further analysis.

--------------
SZTAKI Desktop Grid: Good potential!

This project searches for patterns in multi-dimensional binary number systems. It's mainly a Math project, however, I can foresee a time when computers outgrow two dimensions. The foundation of computer architecture is binary arithmetic, so some tangible benefits may come of this exercise.

--------------
Rectilinear Crossing Numbers: Good potential!

This project searches for arrangements of three-dimensional grids to determine the fewest number of crossing lines when every point is connected to every other point.

When computers are driven by lasers and are built in three-dimensions, it may be necessary to know how to avoid the interference of crossing laser beams.

--------------
ABC@Home: Unknown potential!

I don't know the significance of the patterns for which the project is searching, but I suspect it's a Math project.

--------------
Folding@Home: Very good in an unforeseen way!

The good coming from this project is the development of the techniques to offload the arithmetic processing onto a GPU. In fact, I think this proof-of-concept was a major factor in the decision by AMD to purchase ATI. I think the concept will define the next generation of personal computers.

--------------
Rosetta@Home, DPAD, and all other projects which rely on random searches to achieve their goals: Not practical!

The combinitorial explosion experienced with even a few hundred parameters means it is extremely unlikely these projects will find anything of value in our lifetime!

--------------
All of the searches for prime numbers: Only practical for stress testing computers!

--------------
All of the calculations of the digits of pi and other constants: Only practical for stress testing computers!

--Fred Kline
 
petrusbroder

petrusbroder

Elite Member
Nov 28, 2004
13,348
1,155
126
Thanks Rudy Toody for your post. It is very interesting indeed and I agree with much of your points.

I am thinkíng though ...
Rosetta and DPAD searches for combination of parameters - yes. But the search is not random - the parameters are constrained in several different ways (and not only mathematically but also biologically resp. physically) and thus the probability of finding a result is not as small as you think.
If it were as you state
... The combinitorial explosion experienced with even a few hundred parameters means it is extremely unlikely these projects will find anything of value in our lifetime!
Click to expand...
then the projects would not be funded at all - and they are funded by large sciece organisations, the funding process has passed through critical review by scientists (a kind of peer review). I have checked the DPAD and Rosetta, and the science seems solid to me ... if I can trust those who have reviewed the projects and my own (limited) knowledge and experience.
The number of parameters is not in the hundreds - that would never work. In DPAD it is 7 in one simulation ... and in Rosetta - IIRC - it is not more than 11. That is a large enough number though ...
 
Wolfsraider

Wolfsraider

Diamond Member
Jan 27, 2002
8,305
0
76
Originally posted by: petrusbroder
Thanks Rudy Toody for your post. It is very interesting indeed and I agree with much of your points.

I am thinkíng though ...
Rosetta and DPAD searches for combination of parameters - yes. But the search is not random - the parameters are constrained in several different ways (and not only mathematically but also biologically resp. physically) and thus the probability of finding a result is not as small as you think.
If it were as you state
... The combinitorial explosion experienced with even a few hundred parameters means it is extremely unlikely these projects will find anything of value in our lifetime!
Click to expand...
then the projects would not be funded at all - and they are funded by large sciece organisations, the funding process has passed through critical review by scientists (a kind of peer review). I have checked the DPAD and Rosetta, and the science seems solid to me ... if I can trust those who have reviewed the projects and my own (limited) knowledge and experience.
The number of parameters is not in the hundreds - that would never work. In DPAD it is 7 in one simulation ... and in Rosetta - IIRC - it is not more than 11. That is a large enough number though ...
Click to expand...

I wonder peter,

How many vote to back a project not really understanding its function or base values?

How many decide to support a project on the ground floor where little or no info is available?

For Example:

The experiment is called the Neutrino Factory, scheduled for construction some time around 2015. Its primary aim is to fire beams of neutrinos (fundamental particles) through the Earth's interior to detector stations on different continents. They're doing this to measure whether they change type en route (there are 3 types of neutrino) and data from this in turn will allow them to determine the neutrino's mass (and whether it even has mass).
Click to expand...

What did they (dpad) represent in data, was it questions like this that needed answers? Do we have a way yet of knowing the mass (does it exist or not?) (can we trust the information in simulation, since real world results are still 8 years away?

Again:

The reason they want to do this is that the neutrino is just about the most common particle in the universe (billions pass through your body every second) and if it has mass, this could cause the universe to eventually re-collapse on itself. Knowing the mass will also allow scientists to make better models of how the universe began.
Click to expand...

The most common and we know so very little. is the science not just speculation? An unanswered question, we crunch in hopes of finding an answer, Do the scientists not do the same ... choose that which they want to better understand?

As we continue we see:

Actually the machine that's being built (costing at least $1.9bn) has several scientific aims. The neutrinos are used for fundamental physics experiments, but the proton beam that is produced at the start (this hits the target rod at the beginning of the simulation you download) is also going to be used in experiments for neutralizing radioactive waste by transmuting the radioactive elements into stable ones.
Click to expand...

So we are not even crunching the main data but rather the effect on the effect on radioactive waste. This does help but to what extent ? is it that the only true and viable solution to get the grant money came because of a secondary and proveable solution, and that funding the primary goal was too broad or unprovable to be seriously funded in the scientific community?

As we know that funding the uncertain is always hard, I love this next part:

You are simulating the part of the process where the protons hit a target rod and cause pions to be emitted, which decay into muons, which then proceed to a storage ring and decay into electrons and neutrinos. This is a fairly critical part of the apparatus, which catches the pions and confines some of them into a beam while they decay into muons. The efficiency of this dictates that of the entire machine. If the R&D says it isn't efficient enough it may not get funded to be built, but users of this program have already doubled the estimated efficiency, so it probably will.
Click to expand...

It might get built , but even then its a simulation of a thing we know little about, how it works if it has mass and if that mass affects the outcome are we crunching the various possibilities and what happens if the data we crunch is missing key elements?

I am still crunching DPAD :)

I am just saying there are more sides to these projects than meets the eye. If it suits you to run a certain project over another than by all means run that project, even though all my machines are on dpad (whoop de do 3 running it), I respect christian for not running math projects, it shows individualism and creative thought. I agree we should not be forced into a project just because.



Me I am a strange bird lol I run the projects for many reasons ie...

I want to see stats and I want to see them now (I dont get this with dpad lol)(I warned you I was strange)

I want ease of setup an a stable fast client ( I get this with dpad) Even an idiot like me (setup) can set this lil dude up hardest part is the user text (create a notepad file named user, add [ta]name where name is the name you want to show in stats and hit background bat lol

I tried climate prediction (love the idea behind it) but couldnt stop foaming at the mouth waiting for results to show up lasted less than a day.

Lifemapper and sob were great both fed my stats addiction in instantainous stats = profit lol as well as easy setups.

seti was as bearable as dpad but both left me scanning the stats site to see slight changes in stats hours apart. I truly have little patience lol

I do the projects because i like to think I am helping, but the truth is I am more intrested in stats anymore (Rosetta@Home was different in that it was one of the few boinc projects not validated 3 times (again I hate waiting to see work done finally getting listed on my stats).

My favorite projects, I mean the ones I am most intrested are not the projects I end up crunching day to day, Its all about the stability, ease to run, the stats. Its why i grew to hate my favorite project :(

Anyway off my soapbox as its time to check my stats

Mike
 
petrusbroder

petrusbroder

Elite Member
Nov 28, 2004
13,348
1,155
126
I agree Mike,
Most crunchers crunch for the fun of it and with the hope of contributing ...
What I am reacting to is the supposition that these projects
...extremely unlikely these projects will find anything of value in our lifetime!
Click to expand...
because that would make them meaningless. And at least that I check ... and in science, the fun part is trying to find some understanding ...
Exactly the part you quoted
... but users of this program have already doubled the estimated efficiency, so it probably will.
Click to expand...
gives that project some meaningfulness and thus shows, that - in time - the crunching does some good.

I do not think that that you, Rudy Toody and I disagree in general ... and it is interesting to see that Rudy Toody agree in his judgemt of all projetcs except Rosetta and DPAD.

BTW: has not Rosetta had some success?
 
Rudy Toody

Rudy Toody

Diamond Member
Sep 30, 2006
4,267
421
126
... but users of this program have already doubled the estimated efficiency, so it probably will.
Click to expand...

I caution you about generalities like doubled.

The last time I saw my dentist for my annual exam, I told him that in the last year I had doubled the amount of flossing I do. He said that was great! I said that now I floss twice a year.

This type of hyperbole is commonly used to justify decisions without being precise.

In the mid eighties, I was heavily involved in AI and found that all approaches to using the algorithms has serious weaknesses. For some to get funding they would have to resort to generalities or misleading percentages.

For example, a drug company's testing might indicate that less than 1/100th of 1 percent might destroy their liver by using the drug. Doesn't sound bad until you realize that the drug is intended for the entire world population. That ends up being a large number.
 
Wolfsraider

Wolfsraider

Diamond Member
Jan 27, 2002
8,305
0
76
Oh I do agree just wanted to post a bit, lol, its been a while and I think everyone is entitled to share opinions :)

I never realized how bad my stats watching was untill dpad lol
 
petrusbroder

petrusbroder

Elite Member
Nov 28, 2004
13,348
1,155
126
Originally posted by: Rudy Toody
... but users of this program have already doubled the estimated efficiency, so it probably will.
Click to expand...

I caution you about generalities like doubled.

The last time I saw my dentist for my annual exam, I told him that in the last year I had doubled the amount of flossing I do. He said that was great! I said that now I floss twice a year.

This type of hyperbole is commonly used to justify decisions without being precise.

In the mid eighties, I was heavily involved in AI and found that all approaches to using the algorithms has serious weaknesses. For some to get funding they would have to resort to generalities or misleading percentages.

For example, a drug company's testing might indicate that less than 1/100th of 1 percent might destroy their liver by using the drug. Doesn't sound bad until you realize that the drug is intended for the entire world population. That ends up being a large number.
Click to expand...

Hmmm, what drug would that be? Alcohol? The producers want to sell it to all the world and it damages the liver ... ;)

No drug is intended for the whole world population. It is intended for the sick only - and those are a minority. Usually, without medication, most of the sick either get worse or die.

Lets look at tuberculosis. 400 millions people may have it. Without mediciation but with good food and clear, uninfected air, 150 millions will survive 10 years. 250 millions will die. The drug treatment kills 1 in 10 000, helps 8 000 in 10 000 and has limited effect in 1 999 of 10 000. Instead of 250 million dead there will be 25 000 - plus those who have limited effect: 49.975 Millions. 200 Millions will survive. And it is a bad drug which kills 1 in 10 000.

In surgery it is a good hospital that has an infection rate of 2%. 0.1% of those infected die. Lets say 100 000 get operated for medical reasons. 2 000 get infected. 2 die. Lousy stats? No - because without the operations many many more would die.

You can turn the stats almost any way. In the above examples: an increased efficiency of the drug (from 80% to 90%, which is a 10% increase) would mean that some additional 25 millions patients with tuberculosis would survive for 10 years. In surgery doubled efficiency in infection prevention would mean that 1 patient (instead of 2) in 100 000 operations would die due to infection. Considering that the number of opertions i counted in millions per year that would mean a saving of at least 50 people, probably more.

It all depends on the perspective: if we calculate in human lives a doubled efficiency is a lot. If we count in neutrinos, the impact on life is - at least in the short run - negligable.

The efficiency being doubled in DPAD may be a generalisation. If you check the results page you'll see that in some models the efficiency har increased much more.

I am not argueing too much against you. It is a virtue to be cautious. But the "I said to the dentist"-example is not quite valid when dealing with large numbers - whether it is tens of thousands of neurtinos or millions of humans.
 
petrusbroder

petrusbroder

Elite Member
Nov 28, 2004
13,348
1,155
126
Oh - and it all depends on the dentist. The conversation with my dentist would have been like this:

Petrus: "I have doubled my flossing in the last year."
Dentist: "Hmmmm, yeah, looking at your teeth you may have flossed your teeth about ... 4 times ... and that was probably during the week just after the previous consultation!"

I still like her a lot: she is really good with her hands, sharp in her mind and has a great sense of black humor.
 
petrusbroder

petrusbroder

Elite Member
Nov 28, 2004
13,348
1,155
126
The voting as of now (see timestamp above!)

_1 vote: _Leiden Classical
_1 vote: _Rectilinear Numbres crossing (RCN)
_3 votes: World Community Grid
_1 vote: _Eon____________________]
_6 votes

Please vote - TAS would appreciate more votes - 14 or more is a good number ;)
 
petrusbroder

petrusbroder

Elite Member
Nov 28, 2004
13,348
1,155
126
Gentlewomen and Gentlemen,

You are very much encouraged to do your civic duty as TAS-members and to vote! Not "really soon" but today.

There are still only 6 votes cast, so please do vote - here!
 
You must log in or register to reply here.

TRENDING THREADS

COMPANY
RESOURCES
FOLLOW
Top Bottom