Jim Fulton wrote:
My intuition is still that sharing objects between threads will
introduce a host of subtle bugs.
Yes, I'll +lots to this.
I'm much more interested in seeing a memory-limited cache of some
description an sad to see this thread derail into sharing data between
threads, which
[EMAIL PROTECTED], zodb-dev@zope.org
Objet : Re: [ZODB-Dev] [Enhancement Proposal] Memory size limited Cache
Jim Fulton wrote:
My intuition is still that sharing objects between threads will
introduce a host of subtle bugs.
Yes, I'll +lots to this.
I'm much more interested in seeing
Jim Fulton wrote at 2006-10-6 17:04 -0400:
...
Thread-safety is not an issue for read-only data.
Yes it is. An object that is read-only from a persistence
point of view isn't read-only at the Python level.
Okay: thread-safety is not an issue for (truely) read-only data.
It is potentially an
Dieter Maurer wrote:
Jim Fulton wrote at 2006-10-6 17:04 -0400:
...
Or consider object activation and deactivation. If a ghost is
shared among multiple threads, then __setstate__ could
be called from separate threads.
But, why should this be a problem? They would install the same
state.
...
|Jim Fulton]
Or consider object activation and deactivation. If a ghost is
shared among multiple threads, then __setstate__ could
be called from separate threads.
[Dieter Maurer]
But, why should this be a problem? They would install the same
state.
__setstate__ implementations are
Dieter Maurer wrote:
The pickle size is a *VERY* rough estimation (probably wrong
by a factor of 5 to 15)
But, as you point out, much better than a hard coded 1 ;-)
We probably would get a much better estimation using
PySizer but probably at a significantly higher cost.
Right, I guess
Dieter Maurer wrote:
Jim Fulton wrote at 2006-10-6 12:21 -0400:
...
We would have a use case for this, too.
We've started moving toward just using a single application thread per process
(with many processes). There isn't really much advantage in running multiple
threads if you have multiple
Jim Fulton wrote at 2006-10-6 15:13 -0400:
Dieter Maurer wrote:
Jim Fulton wrote at 2006-10-6 12:21 -0400:
...
We would have a use case for this, too.
We've started moving toward just using a single application thread per
process
(with many processes). There isn't really much advantage in
Dieter Maurer wrote:
...
The only benefit of multiple threads is that it is somewhat
less likely that expensive requests will block inexpensive ones.
Multiple threads can share resources such as main memory and
a common cache (proposed by the original poster).
If there is only one thread,
Jim Fulton wrote at 2006-10-6 16:18 -0400:
Dieter Maurer wrote:
...
The only benefit of multiple threads is that it is somewhat
less likely that expensive requests will block inexpensive ones.
Multiple threads can share resources such as main memory and
a common cache (proposed by the
Dieter Maurer wrote:
Jim Fulton wrote at 2006-10-6 16:18 -0400:
Dieter Maurer wrote:
...
...
We abused Zope a bit and have build a desktop application with it.
One of the main critiques of our customers is too high memory
consumption. Many have computers with 256 MB memory and
they
Dieter Maurer wrote:
A new proposal:
http://www.zope.org/Wikis/ZODB/MemorySizeLimitedCache
It outlines how to implement a ZODB cache limited not be the
number of containing objects but by their estimated memory size.
Feedback welcome -- either here or in the Wiki.
I'm going to reply here
: Re: [ZODB-Dev] [Enhancement Proposal] Memory size limited Cache
Dieter Maurer wrote:
A new proposal:
http://www.zope.org/Wikis/ZODB/MemorySizeLimitedCache
It outlines how to implement a ZODB cache limited not be the
number of containing objects but by their estimated memory size
Jim Fulton wrote at 2006-10-5 08:46 -0400:
...
I'll note that, as a guiding principle, any refactoring we do
should allow pure-python implementation. This means that APIs
need to be Python APIs, although we should consider efficient
C implementations when designing these APIs.
I
14 matches
Mail list logo