Any thoughts about this ? Does it have any chance to bring (hardware) reconfigurable computing to the masses ? In principle, FPGAs can be made into general purpose coprocessors but, in practice, they remain mostly within their own niche, quite well isolated from the mainstream programming. One of the reasons seems to be apparent cultural differences between programming and hardware design worlds. Will acquisition of Altera by Intel somehow result in bringing these two closer ?
>This is largely horse pucky. FPGAs are a trade off of efficiency for generality. FPGA based
uh... the same holds for von Neumann's CPU architecture;o)
>coprocessors only provide a benefit where the algorithm can be implemented more efficiently in >logic than conventional code and it's done at a high enough frequency to warrant the trouble. Few >situations on the desktop meet these criteria.
'situations on the desktop' a quite well handled by the contemporary CPUs. One can only type that fast.... Desktop is not an issue, no matter multi core CPUs, Cell Processor, FPGA based solutions...
>The only situation that comes to mind is video compression/decompression but that is already
The only 'desktop situation'. There's life beyond desktop....
>accelerated quite well by something on the other end of the efficiency/generality spectrum: the GPU.
...like a bunch of engineering/scientific computions that, in places, are embarassingly parallel and, at the same time, embarassingly simple so that dedicating entire Beowulf node to the unit computation is a waste.
Just as example - check TimeLogic's page (http://www.timelogic.com/)- a large class of bioinformatic computations can be accelerated by 2 orders of magnitude. Note, that it translates into substituting Beowulf clusters with a single FPGA-based accelerator board. Hardly horse puckey, I'd say...
>Now, if/when they come out with memory that can be reorganized on-the-fly, perform large-scale >simple massively-parrallel operations, and do some content-addressable tricks, that will be a >signifigant development. I don't know how long it would take that to make it into higher level >programming languages, though. It kinda of turns the job of writing programs on it's head.
Have you ever stumbled on FPGAs ? It's already there. The problem is, as I see it, it does turn writing programs on it's head. Thus, very few people outside of the hardware design crowd know what to do with them. Just think how many people do get exposed to digital design vs programming. How many people do go beyond a vague idea of a processor working on data sitting in memory ? How many CS graduates are utterly unhappy about digital design classes ?
Re:The Superiority of PHP over Perl
on
Perl 6 Essentials
·
· Score: 3, Informative
>limited in what you can do. This is because Perl isn't OO (so you >can't create Node classes, for example, usefull in a linked list) and >because it lacks pointers. Some of you may notice that PHP lacks >pointers, but look deeper! Behind the scenes, hidden from the user
uh.. some might even notice that Perl does NOT 'lack ponters' -
what do you think all those \$a, \@b, \%c are ????
Most of the non-commercial science (including biology) has been done according to this 'open-source' model since eons ego. It has always been the cycle: read publicly available papers on a given topic - do something new - publish results in a publicly available, peer reviewed journal.
I write a lot of scientific papers and presentations with tons of equations. Normally, I use LaTeX for all my stuff, and it works great. However, there are some people (like my supervisor) who are hesitant to use something like LaTeX because of the steep learning curve. They prefer to use
Following up on the scientific papers theme - is anyone working on a bibliography management system for Kword (like EndNote in Win/Mac world) ? It would be another booster for the use by the science/research community.
The idea of making a portal out of it is interesting but BLAST has been around for some time...
True. There's A LOT of sites dealing with various facets of genetic information. Starting from databases containing 'raw' DNA sequences coming off various genomesequencing projects (no, it's not only human; genomes of twenty or so organisms have been or are being sequenced at the moment) all the way to servers trying to interpret the data. Just check Pedro's list or one of the mirrors in Germany or Switzerland. And note that this is not so current (in fact nearly 3 years old) but already pretty long list...
Slashdot Mirror
Any thoughts about this ? Does it have any chance to bring (hardware) reconfigurable computing to the masses ? In principle, FPGAs can be made into general purpose coprocessors but, in practice, they remain mostly within their own niche, quite well isolated from the mainstream programming. One of the reasons seems to be apparent cultural differences between programming and hardware design worlds. Will acquisition of Altera by Intel somehow result in bringing these two closer ?
ok... maybe I'm missing something, but what's wrong with using httpd2 for proxying/load balancing ? It seems to provide both ?
lukasz
uh... the same holds for von Neumann's CPU architecture ;o)
>coprocessors only provide a benefit where the algorithm can be implemented more efficiently in
>logic than conventional code and it's done at a high enough frequency to warrant the trouble. Few
>situations on the desktop meet these criteria.
'situations on the desktop' a quite well handled by the contemporary CPUs. One can only type that fast.... Desktop is not an issue, no matter multi core CPUs, Cell Processor, FPGA based solutions...
>The only situation that comes to mind is video compression/decompression but that is already
The only 'desktop situation'. There's life beyond desktop....
>accelerated quite well by something on the other end of the efficiency/generality spectrum: the GPU.
...like a bunch of engineering/scientific computions that, in places, are embarassingly parallel and, at the same time, embarassingly simple so that dedicating entire Beowulf node to the unit computation is a waste.
Just as example - check TimeLogic's page (http://www.timelogic.com/)- a large class of bioinformatic computations can be accelerated by 2 orders of magnitude. Note, that it translates into substituting Beowulf clusters
with a single FPGA-based accelerator board. Hardly horse puckey, I'd say...
>Corolarius: we want 64 or more processors on a die, with tons of busses and 0.01 micron 30 layer dies ;-)
One can have ~300 8-bit RISCs right now. That's how many can fit into a single mordern FPGA...
>Now, if/when they come out with memory that can be reorganized on-the-fly, perform large-scale
>simple massively-parrallel operations, and do some content-addressable tricks, that will be a
>signifigant development. I don't know how long it would take that to make it into higher level
>programming languages, though. It kinda of turns the job of writing programs on it's head.
Have you ever stumbled on FPGAs ? It's already there. The problem is, as I see it, it does turn writing programs on it's head. Thus, very few people outside of the hardware design crowd know what to do with them.
Just think how many people do get exposed to digital design vs programming. How many people do go beyond a vague idea of a processor working on data sitting in memory ? How many CS graduates are utterly unhappy about digital design classes ?
>limited in what you can do. This is because Perl isn't OO (so you
>can't create Node classes, for example, usefull in a linked list) and
>because it lacks pointers. Some of you may notice that PHP lacks
>pointers, but look deeper! Behind the scenes, hidden from the user
uh.. some might even notice that Perl does NOT 'lack ponters' -
what do you think all those \$a, \@b, \%c are ????
sigh...
Most of the non-commercial science (including biology) has been done according to this 'open-source' model since eons ego. It has always been the cycle: read publicly available papers on a given topic - do something new - publish results in a publicly available, peer reviewed journal.
well, the applet is another incarnation of a demo applet provided on Suns's Java pages.
I write a lot of scientific papers and presentations with tons of equations. Normally, I use LaTeX for all my stuff, and it works great. However, there are some people (like my supervisor) who are hesitant to use something like LaTeX because of the steep learning curve. They prefer to use
Following up on the scientific papers theme - is anyone working on a bibliography management system for Kword (like EndNote in Win/Mac world) ? It would be another booster for the use by the science/research community.
The idea of making a portal out of it is interesting but BLAST has been around for some time...
True. There's A LOT of sites dealing with various facets of genetic information. Starting from databases containing 'raw' DNA sequences coming off various genome sequencing projects (no, it's not only human; genomes of twenty or so organisms have been or are being sequenced at the moment) all the way to servers trying to interpret the data. Just check Pedro's list or one of the mirrors in Germany or Switzerland. And note that this is not so current (in fact nearly 3 years old) but already pretty long list...