One more for your list; a good chunk of the mathematics the damn things are designed on came out of India a LONG time before Ecole Polytechnique et al got there, and certainly before 1776!
Could they have much worse government in 5 years and in pissing contest drop some nuclear bombs on Pakistan?
Or equally, could it be that in 5 years' time the US will democratically elect some redneck who might drop a few bombs on Saddam for the sake of some idiot pissing contest... oh, wait.
For problems which require serialization, by needing results in hand to go onto the next step in the calculation, these things slow down to the speed of one of the component PC's.
...which in the case of POWER4, is still shit on toast.
Interestingly, there is a version of the POWER4 chip available where one of the processor cores is disabled, allowing the remaining core to leverage the L2 cache and bandwidth for better throughput and cache/pipe/prediction efficiency. This config is aimed directly at the, er, scientific and technical market.
Also interesting to note that these days, most quick processors are superscalar and/or hyperthreading, so even a single processor derives much of the performance from limited parallelisation.
--
Well, not a bad question. The theory is that they give a country a huge military advantage and that's a Bad Thing.
I can think of two counterarguments...
1/ Who gave the US the right to decide?
2/ The Manhattan project operated on pencil, paper, and Feynman's grey matter, and look what they done, Ma!
(OK, that's a slight exaggeration, but most of it was achieved using teams of operators on arithmetic machines. ISTR they did get an IBM something-or-other right near the end of the project, but they could easily have got there without.)
One more for your list; a good chunk of the mathematics the damn things are designed on came out of India a LONG time before Ecole Polytechnique et al got there, and certainly before 1776!
Could they have much worse government in 5 years and in pissing contest drop some nuclear bombs on Pakistan?
Or equally, could it be that in 5 years' time the US will democratically elect some redneck who might drop a few bombs on Saddam for the sake of some idiot pissing contest... oh, wait.
For problems which require serialization, by needing results in hand to go onto the next step in the calculation, these things slow down to the speed of one of the component PC's.
Interestingly, there is a version of the POWER4 chip available where one of the processor cores is disabled, allowing the remaining core to leverage the L2 cache and bandwidth for better throughput and cache/pipe/prediction efficiency. This config is aimed directly at the, er, scientific and technical market.
Also interesting to note that these days, most quick processors are superscalar and/or hyperthreading, so even a single processor derives much of the performance from limited parallelisation. --
Go on Europe...keep underestimating in your arrogance and take the gamble. No...really...do it.
s/Europe/Republican America/
I thank you.
Well, not a bad question. The theory is that they give a country a huge military advantage and that's a Bad Thing.
I can think of two counterarguments...
1/ Who gave the US the right to decide?
2/ The Manhattan project operated on pencil, paper, and Feynman's grey matter, and look what they done, Ma! (OK, that's a slight exaggeration, but most of it was achieved using teams of operators on arithmetic machines. ISTR they did get an IBM something-or-other right near the end of the project, but they could easily have got there without.)