Performance and size issues 2003-07-01 - By Joseph Kesselman
> Does this number mean that the maximum node number is about 65K ?
>Apparently not.
Nope. Large documents will "overflow" into additional DTM IDs. This is an
addressing compromise which should allow us to handle up to 64K documents
simultaneously if each is under 64K in size, or fewer gawdawful-huge
documents.
If you're getting "No more DTM IDs are available", you're either loading a
huge number of documents (including temporary trees/RTFs) or we've got a
bug. This shouldn't be happening under normal circumstances, unless you're
churning a single DTMManager to death or are running a single DTMManager
for a very long time. Can you come up with a simple testcase for us to look
at, or otherwise tell us more about how you're actually using Xalan so we
can analyse this in more detail?
>Could the use of XSLTC improve the performance of my application?
Sometimes yes, sometimes no. Can't hurt to try. Generally, XSLTC _should_
be more heavily performance-optimized than interpretive Xalan and
interpretive should be more flexible (especially from a tooling point of
view) than XSLTC, but there are exceptions to that rule and we're still
trying to improve both modes.
For advice on optimizing stylesheets in general, independent of which XSLT
processor you're using, you might want to check out the XSLT User's mailing
list:
http://www.mulberrytech.com/xsl/xsl-list/index.html
______________________________________
Joe Kesselman, IBM Next-Generation Web Technologies: XML, XSL and more.
"The world changed profoundly and unpredictably the day Tim Berners Lee
got bitten by a radioactive spider." -- Rafe Culpin, in r.m.filk
|
|
