I need to get the best throughput I can on single-threaded Matlab tasks. Mostly ODE integration and algebra manipulations, such as matrix inversion, LU factorization, cholesky square roots, etc... Memory requirements are not enormous, but it is possible that memory is still a bottleneck. I'm expecting it's simple CPU power, though.
I'm torn between the Mac Pro and the iMac 17". Unfortunately, our budget year is closing so waiting till the new Mac Pros come out is not possible. I need to buy one on Monday.
I'm looking into insight for comparing the i7 and the Xeon in these two circumstances. I understand that the Xeon can in theory do triple channel memory addressing, but this only if you have 3 RAM modules: no more, no less.
All the extras of a Mac Pro are wasted on this. It's just a question of raw power. The 500E price difference between a 2.66Ghz Mac Pro and a 2.83Ghz i7 iMac is not important. The 1000E+ difference between the Mac Pro with upgraded processors and the iMac IS. In any case, an 8-core would be wasted on this, it's a single-threaded task.
Which do I get come Monday?
I'm torn between the Mac Pro and the iMac 17". Unfortunately, our budget year is closing so waiting till the new Mac Pros come out is not possible. I need to buy one on Monday.
I'm looking into insight for comparing the i7 and the Xeon in these two circumstances. I understand that the Xeon can in theory do triple channel memory addressing, but this only if you have 3 RAM modules: no more, no less.
All the extras of a Mac Pro are wasted on this. It's just a question of raw power. The 500E price difference between a 2.66Ghz Mac Pro and a 2.83Ghz i7 iMac is not important. The 1000E+ difference between the Mac Pro with upgraded processors and the iMac IS. In any case, an 8-core would be wasted on this, it's a single-threaded task.
Which do I get come Monday?