
Re: Re: Famous Mathematica 5.0 not working with Mathematica
Posted:
Jun 1, 2007 2:54 AM


JeanMarc Gulliet wrote: > Anonymous wrote: >> Hello. I am performing an evaluation of Mathematica 6 with the demo >> version on both Windows XP 32 bits and OS X 10.4.9. However both crash >> when evaluating the famous Karl's Mathematica benchmark: >> http://www2.staff.fhvorarlberg.ac.at/~ku/karl/timings50.html >> >> Did anyone experience the same thing? Have you got any suggestion? >> Thank you. > > Hi, > > I have done some more testing, and the culprit seems to be the graphic > engine. Having commented out the following two expressions from Karl's > original notebook, > > Timing[Plot3D[Sin[x y], {x, 0, 2 \[Pi]}, {y, 0, 2 \[Pi]}, > PlotPoints > 4000, DisplayFunction > Identity]][[1]], > Timing[Plot3D[Evaluate[q4[x, t]], {x, 40, 40}, {t, 40, 40}, > PlotPoints > 2000, DisplayFunction > Identity]][[1]] > > Mathematica did not consume too much memory. > > Moreover, if one loads the legacy graphic engine > > In[1]:= << Version5`Graphics` > > then evaluating the following expression > > In[2]:= Plot3D[Sin[x y], {x, 0, 2 \[Pi]}, {y, 0, 2 \[Pi]}, > PlotPoints > 4000, DisplayFunction > Identity] > > Out[2]= SurfaceGraphics > > works fine. However, evaluating the same expression with the new graphic > engine eats up all the memory available. > > In[3]:= << Version6`Graphics` > > In[4]:= Plot3D[Sin[x y], {x, 0, 2 \[Pi]}, {y, 0, 2 \[Pi]}, > PlotPoints > 4000, DisplayFunction > Identity] > > Regards, > JeanMarc > > The option PlotPoints>4000 sets a much, much, much larger number of points in Mathematica 6.0 due to the adaptive sampling feature of Plot3D. (http://www.wolfram.com/products/mathematica/newin6/content/HighImpactAdaptiveVisualization/)
You should try setting the option MaxRecursion>0 in Plot3D to disable that adaptive sampling.
 Igor C. Antonio Wolfram Research, Inc.
To email me personally, remove the dash.

