RE: [Zope] zope benchmarks now available
It's a great start. Thanks. I find it very helpful. I'm also very curious about performance on dynamic content. My experience has been, and I think I'm not alone in this, that even with my rarefied understanding of the capabilities of Zope, I find myself exploiting it's advantages. I don't think I have a single static page on my site. I think it's more to do with mindset. I just don't think in terms of "static" pages, I'm more concerned with the logic than the presentation. In fact some of my pages are extremely complex. Perhaps much more than they need to be, but I'm still learning. Another comment I need to make - this is through experience with my manager. On http://www.zope.org/Members/BwanaZulia/zope_benchmarks/benchmarks.html the key statement is "Tomcat is 2x or 3x faster than Zope when execute an SQL query", the key words being "SQL query". Not pages. Even I missed that first. I'm stating the obvious here, but sometimes even the obvious needs to be restated. And a suggestion - I'd be very interested to know if the above was (still) true for DB access. One piece of logic we frequently use here is to generate a form with default values that are dynamically extracted from a database. Explanation - we have some very lengthy forms. Since we don't expect users to complete them in one go, we give them the option of storing whatever they enter in a session in the DB. And each time the user wishes to update the information, the input form is dynamically generated with SQL query variables (I'm still using <dtml-sqlvars ...> which is slower than the alternatives, but like I said earlier, I'm still learning) being used as default values. Thanks again. Regards, Samir. -----Original Message----- From: J Cameron Cooper [mailto:jccooper@jcameroncooper.com] Sent: Wednesday, March 12, 2003 03:02 To: zope@zope.org Subject: [Zope] zope benchmarks now available A little while back I promised to run some benchmarks. Took a little while longer than I thought, and there is a lot to be desired, but I have published the preliminary results. My results (unfortunately) don't agree with the qualitative results of http://www.zope.org/Members/BwanaZulia/zope_benchmarks/benchmarks.html I imagine the more thorough tests I have planned will reveal more. Anyway, if you like numbers (and who doesn't?) you can find my results at http://zopeedge.com/benchmarks It's a little earlier than I had planned to make the site publicly known, so be well warned that the only thing there that really works is the benchmark area. --jcc _______________________________________________ Zope maillist - Zope@zope.org http://mail.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope-dev )
I'm also very curious about performance on dynamic content. My experience has been, and I think I'm not alone in this, that even with my rarefied understanding of the capabilities of Zope, I find myself exploiting it's advantages. I don't think I have a single static page on my site. I think it's more to do with mindset. I just don't think in terms of "static" pages, I'm more concerned with the logic than the presentation. In fact some of my pages are extremely complex. Perhaps much more than they need to be, but I'm still learning.
Indeed. That test page was indeed one of the few times I've ever used static content in Zope. One of my goals for the future is to test the dynamic capabilties better -- I used static and simple include pages to start with because they are (surprise) easier to set up, especially in translation between DTML/ZPT/JSP/SSI. This will force Apache out and allow JBoss in. Time is all it will take to set up a more realistic test suite. In fact, I suspect that one of the reasons for the speed differences in my tests is that Zope treats all requests (even static ones) with the same procedures regarding security, REQUEST parsing, traversal, et al. Apache and the Java servers are doing few if any of these things: when I make them jump these hoops in the future, we'll see some better comparisons.
Another comment I need to make - this is through experience with my manager. On http://www.zope.org/Members/BwanaZulia/zope_benchmarks/benchmarks.html the key statement is "Tomcat is 2x or 3x faster than Zope when execute an SQL query", the key words being "SQL query". Not pages. Even I missed that first. I'm stating the obvious here, but sometimes even the obvious needs to be restated.
I've stayed away from RDBs so far, but
And a suggestion - I'd be very interested to know if the above was (still) true for DB access.
I can definitely put this on the list. In fact, I already have. It may take a bit to implement (a database is a major component.) --jcc
On Wednesday 12 March 2003 12:00 pm, J Cameron Cooper wrote:
I'm also very curious about performance on dynamic content. My experience has been, and I think I'm not alone in this, that even with my rarefied understanding of the capabilities of Zope, I find myself exploiting it's advantages. I don't think I have a single static page on my site. I think it's more to do with mindset. I just don't think in terms of "static" pages, I'm more concerned with the logic than the presentation. In fact some of my pages are extremely complex. Perhaps much more than they need to be, but I'm still learning.
Indeed. That test page was indeed one of the few times I've ever used static content in Zope.
One of my goals for the future is to test the dynamic capabilties better -- I used static and simple include pages to start with because they are (surprise) easier to set up, especially in translation between DTML/ZPT/JSP/SSI. This will force Apache out and allow JBoss in. Time is all it will take to set up a more realistic test suite.
Just for a simple comparison could you compare python scripts doing the same thing to DTML and ZPT also. I have done some of my own tests and found that python scripts are usually 4-10x faster then DTML or ZPT especially when things get complex. I have recently started moving some stuff from DTML to python scripts for speed and it has worked well.
Just for a simple comparison could you compare python scripts doing the same thing to DTML and ZPT also. I have done some of my own tests and found that python scripts are usually 4-10x faster then DTML or ZPT especially when things get complex. I have recently started moving some stuff from DTML to python scripts for speed and it has worked well.
When we get to more advanced computations, I'll naturally use Python Scripts. I'll even make a few tests for presentation using Python Scripts, since you ask. --jcc
participants (3)
-
J Cameron Cooper -
kosh -
Samir Mishra