Re: Performance of RevMedia on matrix diagonalization
Dear Viktoras, thanks a lot for the suggestions. Yes VB is compiled (although in a sort of p-code) and it isexpected to be faster than any scripting language. I did't try to do what you have suggested because I am not mastering other languages but VB6 and FORTRAN. I am also convinced that trying to rewrite matrix algebra using the very fast routines provided natively by Revolution will improve the matters a lot. As usual, because I am not a professional in this field, I have to find some time to do that! Cheers Piero 2009/8/27 viktoras d. vikto...@ekoinf.net Hi Piero, declaring local/global variables in functions and handlers should slightly add to speed. Populating arrays in advance with zeros so that array size won't change inside loops might also be of some value. However, please correct me if I am wrong, VB6 is compiled while Revolution is interpreted, therefore much slower. Did you try implementing same algorithm in other interpreted languages (javascript, php, perl, etc?) it would be interesting to do some benchmarking :-) Using native functions: matrixMultiply(firstArray, secondArray), transpose(array) where possible would also increase speed as they are really fast. Best wishes Viktoras Piero Ugliengo wrote: I am completely new to Revolution. I have downloaded the RevMedia alpha version and played a bit with it. I was impressed by how fast I was able to port a little VB6 code to RevMedia. This code diagonalizes a symmetric matrix using the Jacobi algorithm. I checked against the VB6 code and I got exactly the same results in the same number of iterations so that numerics is the same. However the RevMedia code is at least one order of magnitude slower than the VB6 one. For instance a 50x50 matrix is diagonalized in a couple of seconds in VB6 and it tooks more than a minute on RevMedia. I know that RevMedia is not meant for numerical intensive calculations; however, I would like to use it in a scientific teaching context so some power is needed. I wonder if anybody much expert than me can try the code using the most powerful Revolution Studio and if there is a clever way to deal with matrix algebra. The link to download the rev script is here: http://sites.google.com/site/pierougliengo/download-1/test.rev?attredirects=0 http://sites.google.com/site/pierougliengo/download-1/test.rev?attredirects=0 One can set the size of the matrix in the onmouse routine. Thanks a lot _ Piero Ugliengo ___ use-revolution mailing list use-revolution@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-revolution -- Prof. Piero Ugliengo University of Torino Dip. Chimica IFM, Via P. Giuria, 7 I-10125 Torino ITALY Phone: +39-011-670.4596 FAX: +39-011-236.4596 E-mail: piero.uglie...@unito.it Home page: http://web086.unito.it/cgi-bin/chimifm/persone.pl/Show?_id=ugliengosort=DEFAULTsearch=hits=70 MOLDRAW: http://www.moldraw.unito.it ___ use-revolution mailing list use-revolution@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-revolution
Re: Performance of RevMedia on matrix diagonalization
Hi Bernard, you are certianly right about RB. Indeed I am following their developmentduring the years. I am pretty sure that the same code will be probably faster in RB than in VB6 but I have yet to decide to buy their compiler. As you said, what is attractive in RevMedia is its free character and the fact that one can write a trully portable app running within any browser. It is also rather easy to adopt as a language for teaching programming concepts to people in the scientific field without the effort needed with Java or C++. Python is certainly a beautiful language extremely reach in features and ad hoc libraries and can be adopted as a general purpose language also in a scientific context (Numpy will solve almost all numerical aspects). On the other hand I still have to find an easy way to deploy to a final user a complex Python application which would be straightforward to install on his/her computer. Cheers Piero 2009/8/28 Bernard Devlin bdrun...@gmail.com Viktoras, Maybe the wording in the IDE (and maybe the docs) is misleading. After all, the button in the script editor says 'compile', not just 'save' or 'check syntax'. It might very well be that it is compiled into some form of p-code that doesn't mean that it needs to be re-compiled each time the script inside a control is called. (I believe that Python and other scripting languages also compile down to some form that doesn't require recompilation - I know that REXX used to do this). If it is compiled then one might well have an expectation that it would perform at speeds near to those of VB. After all, Java is compiled to byte code and it performs pretty well these days. (Although it hardly seems reasonable to expect a company the size of RunRev to be able to optimize cross-platform execution the way that a company the size of Sun [with additional help/stimulus from companies like IBM/Oracle]). It might also be interesting/useful for Piero to implement his code in RealBasic. Since it is so close to VB it should be fairly trivial. After all, RB is cross platform and from a small company too. Although if Piero is interested in delivering content through a browser, then RB may not be of much interest to him even if it is faster. Bernard On Thu, Aug 27, 2009 at 8:30 PM, viktoras d.vikto...@ekoinf.net wrote: However, please correct me if I am wrong, VB6 is compiled while Revolution is interpreted, therefore much slower. Did you try implementing same algorithm in other interpreted languages (javascript, php, perl, etc?) it would be interesting to do some benchmarking :-) ___ use-revolution mailing list use-revolution@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-revolution -- Prof. Piero Ugliengo University of Torino Dip. Chimica IFM, Via P. Giuria, 7 I-10125 Torino ITALY Phone: +39-011-670.4596 FAX: +39-011-236.4596 E-mail: piero.uglie...@unito.it Home page: http://web086.unito.it/cgi-bin/chimifm/persone.pl/Show?_id=ugliengosort=DEFAULTsearch=hits=70 MOLDRAW: http://www.moldraw.unito.it ___ use-revolution mailing list use-revolution@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-revolution
Re: Performance of RevMedia on matrix diagonalization
I just tried it in Rev Studio. Changed 'N' to 50 under mouseup, it took a shade over 30 seconds to diagonalise, I'm using a 4 year old AMD 64 3200+. I'll leave the optimising to others for now.. 2009/8/27 Piero Ugliengo piero.uglie...@gmail.com: I am completely new to Revolution. I have downloaded the RevMedia alpha version and played a bit with it. I was impressed by how fast I was able to port a little VB6 code to RevMedia. This code diagonalizes a symmetric matrix using the Jacobi algorithm. I checked against the VB6 code and I got exactly the same results in the same number of iterations so that numerics is the same. However the RevMedia code is at least one order of magnitude slower than the VB6 one. For instance a 50x50 matrix is diagonalized in a couple of seconds in VB6 and it tooks more than a minute on RevMedia. I know that RevMedia is not meant for numerical intensive calculations; however, I would like to use it in a scientific teaching context so some power is needed. I wonder if anybody much expert than me can try the code using the most powerful Revolution Studio and if there is a clever way to deal with matrix algebra. The link to download the rev script is here: http://sites.google.com/site/pierougliengo/download-1/test.rev?attredirects=0 http://sites.google.com/site/pierougliengo/download-1/test.rev?attredirects=0One can set the size of the matrix in the onmouse routine. Thanks a lot _ Piero Ugliengo -- Prof. Piero Ugliengo University of Torino Dip. Chimica IFM, Via P. Giuria, 7 I-10125 Torino ITALY Phone: +39-011-670.4596 FAX: +39-011-236.4596 E-mail: piero.uglie...@unito.it Home page: http://web086.unito.it/cgi-bin/chimifm/persone.pl/Show?_id=ugliengosort=DEFAULTsearch=hits=70 MOLDRAW: http://www.moldraw.unito.it ___ use-revolution mailing list use-revolution@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-revolution ___ use-revolution mailing list use-revolution@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-revolution
Re: Performance of RevMedia on matrix diagonalization
I saw this over on the forum (the diagonalisation takes about twenty seconds on my MacBookPro with rev enterprise), and two things occur to me. 'Arrays' in revolution are actually hash tables, so are very nice for random access to elements, but will not be as efficient for sequential access as more traditional, fixed width arrays in other languages. I suspect that this is the main speed issue with your eigen function. Secondly, I understand that all numbers in revolution (which is an un- typed language) are floats - though there may be conversions to and from integers going on in the engine. I doubt that this will have any effect on the speed of your functions, but it's probably unneccessary to write '1.0' or '1.' - simply '1' will do, I think. I think that using the a[i][j] approach instead of a[i,j] might get you maybe 10% more speed, but then you'd have to convert to a[i,j] anyway in order to use the built in matrixmultiply and transpose functions (this should be updated, I'd have thought). My maths is nowhere near good enough to suggest mathematical optimisations (which you might not want for educational purposes, anyway) and I can't see any obvious optimisations to make in the code you've written. Hopefully, someone here will be able to be more helpful, but this may be one of those things that revolution simply isn't best suited for :( Best, Mark Smith On 27 Aug 2009, at 08:16, Piero Ugliengo wrote: I am completely new to Revolution. I have downloaded the RevMedia alpha version and played a bit with it. I was impressed by how fast I was able to port a little VB6 code to RevMedia. This code diagonalizes a symmetric matrix using the Jacobi algorithm. I checked against the VB6 code and I got exactly the same results in the same number of iterations so that numerics is the same. However the RevMedia code is at least one order of magnitude slower than the VB6 one. For instance a 50x50 matrix is diagonalized in a couple of seconds in VB6 and it tooks more than a minute on RevMedia. I know that RevMedia is not meant for numerical intensive calculations; however, I would like to use it in a scientific teaching context so some power is needed. I wonder if anybody much expert than me can try the code using the most powerful Revolution Studio and if there is a clever way to deal with matrix algebra. The link to download the rev script is here: http://sites.google.com/site/pierougliengo/download-1/test.rev? attredirects=0 http://sites.google.com/site/pierougliengo/download-1/test.rev? attredirects=0One can set the size of the matrix in the onmouse routine. Thanks a lot _ Piero Ugliengo -- Prof. Piero Ugliengo University of Torino Dip. Chimica IFM, Via P. Giuria, 7 I-10125 Torino ITALY Phone: +39-011-670.4596 FAX: +39-011-236.4596 E-mail: piero.uglie...@unito.it Home page: http://web086.unito.it/cgi-bin/chimifm/persone.pl/Show? _id=ugliengosort=DEFAULTsearch=hits=70 MOLDRAW: http://www.moldraw.unito.it ___ use-revolution mailing list use-revolution@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-revolution ___ use-revolution mailing list use-revolution@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-revolution
Re: Performance of RevMedia on matrix diagonalization
Dear Mark, thanks a lot for the advise. Things like 1.0 istead of 1 were just my ignorance of the language.On the other hand I thought that the syntax a[i][j] instead of the more readable a[i,j] was imposed by the language. Cheers Piero 2009/8/27 Mark Smith li...@futilism.com I saw this over on the forum (the diagonalisation takes about twenty seconds on my MacBookPro with rev enterprise), and two things occur to me. 'Arrays' in revolution are actually hash tables, so are very nice for random access to elements, but will not be as efficient for sequential access as more traditional, fixed width arrays in other languages. I suspect that this is the main speed issue with your eigen function. Secondly, I understand that all numbers in revolution (which is an un-typed language) are floats - though there may be conversions to and from integers going on in the engine. I doubt that this will have any effect on the speed of your functions, but it's probably unneccessary to write '1.0' or '1.' - simply '1' will do, I think. I think that using the a[i][j] approach instead of a[i,j] might get you maybe 10% more speed, but then you'd have to convert to a[i,j] anyway in order to use the built in matrixmultiply and transpose functions (this should be updated, I'd have thought). My maths is nowhere near good enough to suggest mathematical optimisations (which you might not want for educational purposes, anyway) and I can't see any obvious optimisations to make in the code you've written. Hopefully, someone here will be able to be more helpful, but this may be one of those things that revolution simply isn't best suited for :( Best, Mark Smith On 27 Aug 2009, at 08:16, Piero Ugliengo wrote: I am completely new to Revolution. I have downloaded the RevMedia alpha version and played a bit with it. I was impressed by how fast I was able to port a little VB6 code to RevMedia. This code diagonalizes a symmetric matrix using the Jacobi algorithm. I checked against the VB6 code and I got exactly the same results in the same number of iterations so that numerics is the same. However the RevMedia code is at least one order of magnitude slower than the VB6 one. For instance a 50x50 matrix is diagonalized in a couple of seconds in VB6 and it tooks more than a minute on RevMedia. I know that RevMedia is not meant for numerical intensive calculations; however, I would like to use it in a scientific teaching context so some power is needed. I wonder if anybody much expert than me can try the code using the most powerful Revolution Studio and if there is a clever way to deal with matrix algebra. The link to download the rev script is here: http://sites.google.com/site/pierougliengo/download-1/test.rev ?attredirects=0 http://sites.google.com/site/pierougliengo/download-1/test.rev ?attredirects=0One can set the size of the matrix in the onmouse routine. Thanks a lot _ Piero Ugliengo -- Prof. Piero Ugliengo University of Torino Dip. Chimica IFM, Via P. Giuria, 7 I-10125 Torino ITALY Phone: +39-011-670.4596 FAX: +39-011-236.4596 E-mail: piero.uglie...@unito.it Home page: http://web086.unito.it/cgi-bin/chimifm/persone.pl/Show ?_id=ugliengosort=DEFAULTsearch=hits=70 MOLDRAW: http://www.moldraw.unito.it ___ use-revolution mailing list use-revolution@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-revolution ___ use-revolution mailing list use-revolution@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-revolution -- Prof. Piero Ugliengo University of Torino Dip. Chimica IFM, Via P. Giuria, 7 I-10125 Torino ITALY Phone: +39-011-670.4596 FAX: +39-011-236.4596 E-mail: piero.uglie...@unito.it Home page: http://web086.unito.it/cgi-bin/chimifm/persone.pl/Show?_id=ugliengosort=DEFAULTsearch=hits=70 MOLDRAW: http://www.moldraw.unito.it ___ use-revolution mailing list use-revolution@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-revolution
Re: Performance of RevMedia on matrix diagonalization
Hi Piero, declaring local/global variables in functions and handlers should slightly add to speed. Populating arrays in advance with zeros so that array size won't change inside loops might also be of some value. However, please correct me if I am wrong, VB6 is compiled while Revolution is interpreted, therefore much slower. Did you try implementing same algorithm in other interpreted languages (javascript, php, perl, etc?) it would be interesting to do some benchmarking :-) Using native functions: matrixMultiply(firstArray, secondArray), transpose(array) where possible would also increase speed as they are really fast. Best wishes Viktoras Piero Ugliengo wrote: I am completely new to Revolution. I have downloaded the RevMedia alpha version and played a bit with it. I was impressed by how fast I was able to port a little VB6 code to RevMedia. This code diagonalizes a symmetric matrix using the Jacobi algorithm. I checked against the VB6 code and I got exactly the same results in the same number of iterations so that numerics is the same. However the RevMedia code is at least one order of magnitude slower than the VB6 one. For instance a 50x50 matrix is diagonalized in a couple of seconds in VB6 and it tooks more than a minute on RevMedia. I know that RevMedia is not meant for numerical intensive calculations; however, I would like to use it in a scientific teaching context so some power is needed. I wonder if anybody much expert than me can try the code using the most powerful Revolution Studio and if there is a clever way to deal with matrix algebra. The link to download the rev script is here: http://sites.google.com/site/pierougliengo/download-1/test.rev?attredirects=0 http://sites.google.com/site/pierougliengo/download-1/test.rev?attredirects=0One can set the size of the matrix in the onmouse routine. Thanks a lot _ Piero Ugliengo ___ use-revolution mailing list use-revolution@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-revolution
Re: Performance of RevMedia on matrix diagonalization
Viktoras, Maybe the wording in the IDE (and maybe the docs) is misleading. After all, the button in the script editor says 'compile', not just 'save' or 'check syntax'. It might very well be that it is compiled into some form of p-code that doesn't mean that it needs to be re-compiled each time the script inside a control is called. (I believe that Python and other scripting languages also compile down to some form that doesn't require recompilation - I know that REXX used to do this). If it is compiled then one might well have an expectation that it would perform at speeds near to those of VB. After all, Java is compiled to byte code and it performs pretty well these days. (Although it hardly seems reasonable to expect a company the size of RunRev to be able to optimize cross-platform execution the way that a company the size of Sun [with additional help/stimulus from companies like IBM/Oracle]). It might also be interesting/useful for Piero to implement his code in RealBasic. Since it is so close to VB it should be fairly trivial. After all, RB is cross platform and from a small company too. Although if Piero is interested in delivering content through a browser, then RB may not be of much interest to him even if it is faster. Bernard On Thu, Aug 27, 2009 at 8:30 PM, viktoras d.vikto...@ekoinf.net wrote: However, please correct me if I am wrong, VB6 is compiled while Revolution is interpreted, therefore much slower. Did you try implementing same algorithm in other interpreted languages (javascript, php, perl, etc?) it would be interesting to do some benchmarking :-) ___ use-revolution mailing list use-revolution@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-revolution
