Re: Mathematica 8: first impressions
- To: mathgroup at smc.vnet.net
- Subject: [mg113998] Re: Mathematica 8: first impressions
- From: blamm64 <blamm64 at charter.net>
- Date: Sat, 20 Nov 2010 06:13:12 -0500 (EST)
- References: <firstname.lastname@example.org>
First Impressions: 1) Regarding NDSolve. I submitted, on the same machine, an identical, not-that-complicated, mutl-rigid-body dynamics set of differential equations (only six unknowns) to Mathematica 8 and to Mathematica 7.0.1: Timing - 7.0.1 = 1.482 seconds; 8.0 = 219.218 seconds (averaged over three submissions of identical model). Even though the system is 'purely' differential (not DAE), I forced Mathematica to use IDA, since *my* experience has been it is superior to other methods for handling stiff systems (maybe it defaults to LSODA where SolveDelayed->True is used and the system is not DAE; I've no idea). Need I even write that's not good? And I mean *identical*; I simply ran the same notebook in 8.0 and then in 7.0.1., after restarting OS. Plotting of characteristics (constraint forces) had time disparities even worse than the disparity in Timing to solve. One such plot ran in 1. seconds in 7.0.1 and 529.562 seconds in 8.0. Perhaps Mathematica 8.0 no longer implements the fantastic suite authored by Hindmarsh, especially IDA??? NDSolve is promulgated by Wolfram as being the 'biggest' "super- function" available in Mathematica and apparently pretty great strides have been taken ... in the wrong direction, at least for this particular case of mine. I might mention the results of the simulations were virtually identical, so NDSolve in Mathematica 8.0 did not uncover anything new. I'll run some even more involved models I've previously created and compare. 2) 64-bit has been around long enough to be pretty well received: In Mathematica 8.0 $MachinePrecision is still 15.9.... on my Intel, Windows XP x64 machine (2x 4-core Intel Xeon X5570, nVidia Quadro FX 4800, 24 GB RAM) . Yeah, sure, Wolfram a good while back implemented Mathematica to use 64 bit addressable memory (which really helps since it's such a RAM hog), but still not to have implemented using 64-bit available 'precision' is very disappointing ... . Yeah, Mathematica uses double precision 32 bit to get 15.9... machine "precision", but by now I think Mathematica, if installed on a 64-bit system, should have $MachinePrecision of 31.8.... , or something close to that on a 64-bit system. Yeah, sure, *Great* strides have been implemented and I'll never tap into even a small fraction of what's available, I'm sure, but to not, apparently, give good attention to *core* functionality is disappointing. I'm sure I'll find some really great features that are very useful and powerful not found in versions previous to 8.0, I still consider myself a beginner with Mathematica, but the above two points in my opinion are valid complaints. -Brian L.