> If there is something crucial missing, let me know:)
Missing? How about Avaki, The Legion Project, Sun Grid Engine, and a host of other companies that have working product out there but didn't prescribe to Ian Foster's Globus project as the "One and Only True (Grid) Way".
It's a real pity that (from what I've seen) the OGSA is anything but open, and seems intent that only one implementation shall ever be called "Grid" and it is Globus.
The speed of the electron is not the issue. The speed of the signal propagation in the medium is the issue. That is still a significant fraction of C.
As the previous poster commented, in a vaccum, C is about 30cm per second. While the speed of signal propagation in copper is somewhat slower, 1.0cm/sec (as the earlier poster said) is probably too slow.
Well, I can chime in on with some details on one of these ten.
Greengard and Rokhlin's Fast Multipole Method (FMM) algorithm computes an approximation for the sum total of the interactions between all pairs of elements out of a large group.
For instance, in astrophysics simulations, one quantity that needs to be computed is the total force on a star that results from the gravitational attraction from each of the other stars. If you have to do this computation for each star, then the total ammount of computation required grows as N^2, given N total stars.
This is where the name "The N-body Problem" comes from.
The FMM algorithm essentially models distant groups of particles (stars) as a single mathematical object and by using other fairly complex operations and representations, reduces the overall complexity from N^2 to N.
The importance of this algortihm comes from the fact that in many different types of scientific simulations (astrophysics, molecular modeling, computational fluid dynamics, etc.) the N-body computation was the limiting factor in the performace of these algorithms. Use of FMM and similar algortihms has reduced the overall simulation times by orders of magnitudes for large systems, allowing simulations that once required CPU-decades to be completed in CPU-months.
There are several good sources of FMM material on the web. You can try:
Slashdot Mirror
> If there is something crucial missing, let me know :)
Missing? How about Avaki, The Legion Project, Sun Grid Engine, and a host of other companies that have working product out there but didn't prescribe to Ian Foster's Globus project as the "One and Only True (Grid) Way".
It's a real pity that (from what I've seen) the OGSA is anything but open, and seems intent that only one implementation shall ever be called "Grid" and it is Globus.
Doh!
(open mount insert foot, chew, spit toes)
That would be approx 30cm/ns, *not* second.
What's 9 orders of magnitude between friends, eh?
-b
Still wrong.
The speed of the electron is not the issue. The speed of the signal propagation in the medium is the issue. That is still a significant fraction of C.
As the previous poster commented, in a vaccum, C is about 30cm per second. While the speed of signal propagation in copper is somewhat slower, 1.0cm/sec (as the earlier poster said) is probably too slow.
Total off topic now,
-b
Greengard and Rokhlin's Fast Multipole Method (FMM) algorithm computes an approximation for the sum total of the interactions between all pairs of elements out of a large group.
For instance, in astrophysics simulations, one quantity that needs to be computed is the total force on a star that results from the gravitational attraction from each of the other stars. If you have to do this computation for each star, then the total ammount of computation required grows as N^2, given N total stars.
This is where the name "The N-body Problem" comes from.
The FMM algorithm essentially models distant groups of particles (stars) as a single mathematical object and by using other fairly complex operations and representations, reduces the overall complexity from N^2 to N.
The importance of this algortihm comes from the fact that in many different types of scientific simulations (astrophysics, molecular modeling, computational fluid dynamics, etc.) the N-body computation was the limiting factor in the performace of these algorithms. Use of FMM and similar algortihms has reduced the overall simulation times by orders of magnitudes for large systems, allowing simulations that once required CPU-decades to be completed in CPU-months.
There are several good sources of FMM material on the web. You can try:
Mario's Nbody resources
More Nbody
Leslie Greengard's page
And of course, I'll have to plug our research group page at Duke
Hope this helps
-bill