I am trying to work around this in SAGE by controlling the number of applets open at once. There are two problems:
1) Safari (and probably FF) on MacOS don't appear to properly free up memory when an applet is destroyed. Chrome on MacOS does, as I have no problems as long as I limit the number of applets open at once. 2) We don't have a good way of estimating how much memory the applets are using in javascript. Could we get javascript access to the information in the "system menu" or have I just missed how to do that? Now what I am doing is just counting applets. This usually works, but is a problem if one of the applets is displaying a lot of surfaces. Then it uses more than the average amount of memory. Jonathan On Mar 29, 2011, at 8:07 AM, jmol-users-requ...@lists.sourceforge.net wrote: > From: Robert Hanson <hans...@stolaf.edu> > Subject: Re: [Jmol-users] Jmol, MacOS, FF & Safari... > To: jmol-users@lists.sourceforge.net > Message-ID: > <AANLkTi=ZKcdEEXbgKAO2imTCtGd49etNCEWkV4=_h...@mail.gmail.com> > Content-Type: text/plain; charset="iso-8859-1" > > That's right. It's not a plug-in. You can't launch a huge number of applets > on a page. Can SAGE work around that? > > On Mon, Mar 28, 2011 at 8:08 AM, Jonathan Gutow <gu...@uwosh.edu> wrote: > >> Dear all, >> >> I've been working this last week on Jmol in SAGE math (www.sagemath.org). >> I thought people might be interested in what I've learned. In the process >> I've unearthed a number of issues on MacOS. >> >> a) When launching a lot of applets on a page in FF sometimes not all the >> applets finish initializing. There appears to be a memory/process collision >> problem. It is so intermittent that I have not been able to pin it down. >> Things seem to be worst on 64 bit installations. >> >> b) Safari also has problems with launching a lot of applets. This one I >> could trace to memory issues. I think Safari has a memory leak as the >> problem occurs even after most of the applets are forced to close. >> Interestingly, this problem shows up as a security violation when calls >> attempt to get an image from the applet as a string. >> >> c) Jmol works very well in Chrome on MacOS. I was unable to produce any of >> the oddities I saw in FF or Safari. >> >> Notes: >> 1) The tests involve launching pages that have lots of Jmols on them. It >> is very easy to run up against the memory limits of the JVM in the browser, >> so the launches are carefully controlled to be sequential and no more than 4 >> of the applets are allowed to stay open simultaneously. A repeating >> javascript process replaces an already open Jmol with an image before >> opening the next one. >> 2) To see examples of what the pages in Sagemath look like go to >> http://<http://141.233.197.45:8888/home/pub/0> >> 141.233.197.45:8888/home/pub/0 >> 3) FF and Chromium work very well on Ubuntu Linux >> 4) I haven't performed these demanding tests on a Windows machine yet. >> >> Jonathan >> Dr. Jonathan H. Gutow >> Chemistry Department gu...@uwosh.edu >> UW-Oshkosh Office:920-424-1326 >> 800 Algoma Boulevard FAX:920-424-2042 >> Oshkosh, WI 54901 >> http://www.uwosh.edu/facstaff/gutow >> > Dr. Jonathan H. Gutow Chemistry Department gu...@uwosh.edu UW-Oshkosh Office:920-424-1326 800 Algoma Boulevard FAX:920-424-2042 Oshkosh, WI 54901 http://www.uwosh.edu/facstaff/gutow ------------------------------------------------------------------------------ Enable your software for Intel(R) Active Management Technology to meet the growing manageability and security demands of your customers. Businesses are taking advantage of Intel(R) vPro (TM) technology - will your software be a part of the solution? Download the Intel(R) Manageability Checker today! http://p.sf.net/sfu/intel-dev2devmar _______________________________________________ Jmol-users mailing list Jmol-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jmol-users
