casperyc

957 Reputation

10 Badges

14 years, 120 days
University of Kent
Dr

Social Networks and Content at Maplesoft.com

MaplePrimes Activity


These are replies submitted by casperyc

I think it depends on memory ??

I don't used to have this problem before 2014 during my PhD study....so I got absolutely no idea what's going on here.......

@Markiyan Hirnyk Thanks! Rather unual way, but works perfectly fine!

@acer On my i7 4790K system, I just did a quick test with n=8000.
 I ran in order these commands from a fresh start kernel:

 

n = 8000
m = RandomReal[{}, {n, n}];

SingularValueDecomposition[m]; // AbsoluteTiming

     {207.795, Null}

MatrixRank[m]; // AbsoluteTiming

    {25.5141, Null}

 

Note that since Version 5.0 (released in 2003), SingularValues has been superseded by SingularValueList and SingularValueDecomposition.

 

If you dont mind creat a Mathematica test file and instructions, for example, I shall run SingularValueDecomposition and MatrixRank seperately from fresh kernels? I can test them on my i7 940X system.

 

 

@acer

Digged up my "old" laptop:

i7 940X

4G RAM

Windows 7 SP1 X64 Ultimate -- English

 

Code used for Mathematica:

n=7000;

m = RandomReal[{}, {n, n}];

MatrixRank[m] // AbsoluteTiming

 

So here are some results:

Mathematica 9.0 - 148 sec

Mathematica 10.0.2 - 142 sec

Mathematica 10.1 - 51 sec

 

with n = 8000;

Mathematica 9.0 - 234 sec

Mathematica 10.0.2 - 231 sec

Mathematica 10.1 - 79 sec

 

@tomleslie Something like this?

@Alejandro Jakubi Thanks. This was taken from a Chinese site. I didn't do much thinking. I will forward your comments.

@acer I only used the code as given by the author from http://12000.org/my_notes/rankTest/test.htm .

I just thought it would be comparable since I have used the same code. Yes, I am aware the versions of MMA are different. But the results were just superising since there was a huge difference. I do have a copy of MMA 10.0.1, which I can test later.

 

I will post more information when I have time. Just for information, my system is i7 4970K, GTX980M, 32G RAM, 250 SSD.

 

@Christopher2222 Do you have the Maple file and the Mathematica file for Kamke's ODEs? I would be interested to run them on my system.Thanks!

@casperyc With n=8000, Mathematica took 25 seconds while Maple used 169sec.

@Christopher2222 hmmmm. I will try some of the code on my system.

 

Could there be something wrong with the results?

 

Maple took nearly 2min (seems normal), but MMA only took 15 seconds?

 

 

@Alejandro Jakubi Thanks! But I would not find ODE solving interesting at all as we know that Mathematica is not good at this at all (at least for my research). I have only been using Mathematica for global optimization and doing some parallel computing, which I find it easier to setup than in Maple.

I hope that Maple would have something that benchmarks the system. I am not asking for some comparison between the two programs if that was not clear. So that I know for example, whether my desktop or my laptop is 'faster' in running Maple.

I wonder if this have been developed?

I know that Mathematica can do a benchmarking testing, but does Maple now have a similar test?

 

Thanks!

@Christopher2222 I have been trying to get a hold of it for ages. But I can't find it anywhere though.

1 2 3 4 5 6 7 Last Page 1 of 22