|

»

Execution Speed

View: New views

6 Messages — Rating Filter: Alert me

	Execution Speed by roger_p :: Rate this Message: Reply to Author \| View Threaded \| Show Only this Message I have been experiencing a sudden slowdown in SableCC. My language has been growing in complexity, so I am not surprised that it is taking longer to process the syntax and generate a parser. However, it has suddenly jumped with the last few changes from a 45minute time on a windows pc to over 6 hours! I noticed that it is consuming about 3gig of memory ... so it could be that it is simply windows running out of memory and the page swapping killing me, or is there some limit/design issue in Sable that would lead to a sudden, dramatic increase in the running time as I add more alternatives, etc. Thanks! -- Be Yourself @ mail.com! Choose From 200+ Email Addresses Get a Free Account at www.mail.com! _______________________________________________ SableCC-Discussion mailing list SableCC-Discussion@... http://lists.sablecc.org/listinfo/sablecc-discussion
	Re: Execution Speed by Etienne M. Gagnon :: Rate this Message: Reply to Author \| View Threaded \| Show Only this Message Hi Roger, The performance of any Java program drops dramatically as soon as the Java virtual machine (jvm) causes memory swapping. There's no artificial barrier where SableCC's performance drop. You might want to look into your jvm's garbage collection options and see if you can tune it to behave better when swapping memory. This is a sticky issue for all garbage collected languages. Etienne PS: The unhelpful answer would be "add more ram to your pc"... :) Roger Pomeroy a écrit : > I have been experiencing a sudden slowdown in SableCC. My language has been growing in complexity, so I am not surprised that it is taking longer to process the syntax and generate a parser. However, it has suddenly jumped with the last few changes from a 45minute time on a windows pc to over 6 hours! I noticed that it is consuming about 3gig of memory ... so it could be that it is simply windows running out of memory and the page swapping killing me, or is there some limit/design issue in Sable that would lead to a sudden, dramatic increase in the running time as I add more alternatives, etc. > > Thanks! > > -- Etienne M. Gagnon, Ph.D. SableCC: http://sablecc.org _______________________________________________ SableCC-Discussion mailing list SableCC-Discussion@... http://lists.sablecc.org/listinfo/sablecc-discussion
	RE: Execution Speed by Christopher Van Kirk :: Rate this Message: Reply to Author \| View Threaded \| Show Only this Message If you're running a 32 bit version of Windows then 3gb is probably going to cause you to disk swap. My recollection is that 3gb is the maximum addressable ram on Windows, and some portion of that limit is reserved by your graphics adapter (e.g. if your video card has 256mb ram, then your non-graphics memory limit is 2.75gb). Why is so much ram being taken up? Can you not break your AST's down into smaller chunks and work on them individually? -----Original Message----- From: sablecc-discussion-bounces+chris.vankirk=fdcjapan.com@... [mailto:sablecc-discussion-bounces+chris.vankirk=fdcjapan.com@.... org] On Behalf Of Roger Pomeroy Sent: Sunday, August 09, 2009 1:41 PM To: sablecc-discussion@... Subject: Execution Speed I have been experiencing a sudden slowdown in SableCC. My language has been growing in complexity, so I am not surprised that it is taking longer to process the syntax and generate a parser. However, it has suddenly jumped with the last few changes from a 45minute time on a windows pc to over 6 hours! I noticed that it is consuming about 3gig of memory ... so it could be that it is simply windows running out of memory and the page swapping killing me, or is there some limit/design issue in Sable that would lead to a sudden, dramatic increase in the running time as I add more alternatives, etc. Thanks! -- Be Yourself @ mail.com! Choose From 200+ Email Addresses Get a Free Account at www.mail.com! _______________________________________________ SableCC-Discussion mailing list SableCC-Discussion@... http://lists.sablecc.org/listinfo/sablecc-discussion _______________________________________________ SableCC-Discussion mailing list SableCC-Discussion@... http://lists.sablecc.org/listinfo/sablecc-discussion
	Re: Execution Speed by Rowan Worth :: Rate this Message: Reply to Author \| View Threaded \| Show Only this Message If I'm reading correctly there are no ASTs involved yet - this is still in parser generation. I must say the Roger has incredible patience - the only time SableCC has taken more than about 3 minutes to generate the parser for me was when I tickled something and put it in an infinite loop... -Rowan On Mon, Aug 10, 2009 at 7:41 AM, Christopher Van Kirk<chris.vankirk@...> wrote: > Why is so much ram being taken up? Can you not break your AST's down into > smaller chunks and work on them individually? > > -----Original Message----- > From: > sablecc-discussion-bounces+chris.vankirk=fdcjapan.com@... > [mailto:sablecc-discussion-bounces+chris.vankirk=fdcjapan.com@.... > org] On Behalf Of Roger Pomeroy > Sent: Sunday, August 09, 2009 1:41 PM > To: sablecc-discussion@... > Subject: Execution Speed > > I have been experiencing a sudden slowdown in SableCC. My language has been > growing in complexity, so I am not surprised that it is taking longer to > process the syntax and generate a parser. However, it has suddenly jumped > with the last few changes from a 45minute time on a windows pc to over 6 > hours! _______________________________________________ SableCC-Discussion mailing list SableCC-Discussion@... http://lists.sablecc.org/listinfo/sablecc-discussion
	Re: Execution Speed by Etienne M. Gagnon :: Rate this Message: Reply to Author \| View Threaded \| Show Only this Message Hi Rowan, The LALR(1) tables computation algorithm has an exponential complexity worst case. I have seen over 60 minutes computation for big grammars. (I've got a colleague that keeps asking me for SableCC 4 because he has to deal with such a grammar...) I guess that Christopher meant "... break your grammar into smaller chunks...". Depending on the type of grammar, that could work. E.g. use a simpler grammar to do a first "high level parsing", then use additional parsers and/or manual tree transformations, to further develop the AST. Have fun! Etienne Rowan Worth a écrit : If I'm reading correctly there are no ASTs involved yet - this is still in parser generation. I must say the Roger has incredible patience - the only time SableCC has taken more than about 3 minutes to generate the parser for me was when I tickled something and put it in an infinite loop... -Rowan On Mon, Aug 10, 2009 at 7:41 AM, Christopher Van Kirkchris.vankirk@... wrote: Why is so much ram being taken up? Can you not break your AST's down into smaller chunks and work on them individually? -- Etienne M. Gagnon, Ph.D. SableCC: http://sablecc.org _______________________________________________ SableCC-Discussion mailing list SableCC-Discussion@... http://lists.sablecc.org/listinfo/sablecc-discussion
	RE: Execution Speed by Christopher Van Kirk :: Rate this Message: Reply to Author \| View Threaded \| Show Only this Message Some parts of this message have been removed. Learn more about Nabble's security policy. Yes, that’s what I meant. Sorry, I’m not so good with languages ;) -----Original Message----- From: sablecc-discussion-bounces+chris.vankirk=fdcjapan.com@... [mailto:sablecc-discussion-bounces+chris.vankirk=fdcjapan.com@...] On Behalf Of Etienne M. Gagnon Sent: Monday, August 10, 2009 11:43 PM To: Discussion mailing list for the SableCC project Subject: Re: Execution Speed Hi Rowan, The LALR(1) tables computation algorithm has an exponential complexity worst case. I have seen over 60 minutes computation for big grammars. (I've got a colleague that keeps asking me for SableCC 4 because he has to deal with such a grammar...) I guess that Christopher meant "... break your grammar into smaller chunks...". Depending on the type of grammar, that could work. E.g. use a simpler grammar to do a first "high level parsing", then use additional parsers and/or manual tree transformations, to further develop the AST. Have fun! Etienne Rowan Worth a écrit : If I'm reading correctly there are no ASTs involved yet - this is still in parser generation. I must say the Roger has incredible patience - the only time SableCC has taken more than about 3 minutes to generate the parser for me was when I tickled something and put it in an infinite loop... -Rowan On Mon, Aug 10, 2009 at 7:41 AM, Christopher Van Kirkchris.vankirk@... wrote: Why is so much ram being taken up? Can you not break your AST's down into smaller chunks and work on them individually? -- Etienne M. Gagnon, Ph.D. SableCC: http://sablecc.org _______________________________________________ SableCC-Discussion mailing list SableCC-Discussion@... http://lists.sablecc.org/listinfo/sablecc-discussion