Re: Mathematica 3.01 is slower than Mathematica 3.0 on PowerMacs!
- To: mathgroup at smc.vnet.net
- Subject: [mg9707] Re: [mg9633] Mathematica 3.01 is slower than Mathematica 3.0 on PowerMacs!
- From: David Withoff <withoff>
- Date: Fri, 21 Nov 1997 01:31:55 -0500
- Sender: owner-wri-mathgroup at wolfram.com
> I reported this before, even sent a bug report in to WRI, and more or > less got a big series of yawns. Here is the scoop: There is a rather > extensive Mathematica benchmark series, created by Karl Unterkolfer, > with up to date results kept at: > http://fampm201.tu-graz.ac.at/karl/timings30.html. If you look at > reference 89 on that page, you will see a report that Mathematica 3.01 > is 15% slower than Mathematica 3.0. This is exactly the same result > that I reported to WRI within days of the release of 3.01. > > I remain very displeased about this result. I have heard nothing from > WRI except a grudging acknowledgement "You're right" and some vague > noises to "wait for the next release". When??? > > Does anyone else see 3.01 to be slower than 3.0 on PowerMacs? Or, is > there something very anomalous about Karl's test codes? I don't think > so. I would be interested in reports from Mathematica users and I would > like to hear further about this from WRI. My earlier report is now > verified on a public web site and I would like to see WRI stop ignoring > this issue and provide an explanation. > > -- > Charles Bouldin I am sorry if you got the impression that your report was being ignored by Wolfram Research, or that it had been dismissed with a yawn or a grudging acknowledgement. None of that is true. All of these reports are taken very seriously, and the fact that the problem has not yet been isolated does not mean that it is being ignored. An important technical point here is that while there are some examples that are known to be slower in Version 3.0.1 relative to Version 3.0.0, there are also quite a number of examples that are faster, and none of the slowdowns in question are prohibitive. It would be possible to conclude, depending on the choice of examples, that Version 3.0.1 is either generally slower than Version 3.0.0, or generally faster. Neither conclusion is really justified by the available examples. Still, any examples of unexpected speed degradation are worth investigating. Several people have already looked at your examples, and we all hope that they will eventually discover something that will provide for definitive answers to your questions. Dave Withoff Wolfram Research