Pre-YTEP: remove caching of datasets
Hi all, Before I get too far into writing up a YTEP, I wanted to know if anyone would object to disabling "caching" of datasets in-memory. Right now, if you do: ds1 = yt.load("something") ds2 = yt.load("something") "ds1 is ds2" will return True; this was done so that you could load a dataset, then unpickle an object with a reference to that dataset, and get the dataset back rather than a newly-loaded one. But I think a) folks don't pickle much b) it causes a lot of problems. So I'm wondering if it would be okay to just turn this off. -Matt
On Fri, Sep 25, 2015 at 12:55 PM, Matthew Turk <matthewturk@gmail.com> wrote:
Hi all,
Before I get too far into writing up a YTEP, I wanted to know if anyone would object to disabling "caching" of datasets in-memory.
Right now, if you do:
ds1 = yt.load("something") ds2 = yt.load("something")
"ds1 is ds2" will return True; this was done so that you could load a dataset, then unpickle an object with a reference to that dataset, and get the dataset back rather than a newly-loaded one. But I think a) folks don't pickle much b) it causes a lot of problems. So I'm wondering if it would be okay to just turn this off.
Big +1 to this. It's been the cause of a number of head-scratchers for me. It also motivated the parts of this change that touch domain_left_edge and domain_right_edge, which is IMO a bit silly but necessary given the way the caching works: https://bitbucket.org/ngoldbaum/yt/commits/dd8d2317489904eb14b24cb63d7e8cfd0...
-Matt _______________________________________________ yt-dev mailing list yt-dev@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
+1 On Fri, Sep 25, 2015 at 6:59 PM, Nathan Goldbaum <nathan12343@gmail.com> wrote:
On Fri, Sep 25, 2015 at 12:55 PM, Matthew Turk <matthewturk@gmail.com> wrote:
Hi all,
Before I get too far into writing up a YTEP, I wanted to know if anyone would object to disabling "caching" of datasets in-memory.
Right now, if you do:
ds1 = yt.load("something") ds2 = yt.load("something")
"ds1 is ds2" will return True; this was done so that you could load a dataset, then unpickle an object with a reference to that dataset, and get the dataset back rather than a newly-loaded one. But I think a) folks don't pickle much b) it causes a lot of problems. So I'm wondering if it would be okay to just turn this off.
Big +1 to this. It's been the cause of a number of head-scratchers for me. It also motivated the parts of this change that touch domain_left_edge and domain_right_edge, which is IMO a bit silly but necessary given the way the caching works:
https://bitbucket.org/ngoldbaum/yt/commits/dd8d2317489904eb14b24cb63d7e8cfd0...
-Matt _______________________________________________ yt-dev mailing list yt-dev@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
_______________________________________________ yt-dev mailing list yt-dev@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
+1 On Fri, Sep 25, 2015 at 11:16 AM, Britton Smith <brittonsmith@gmail.com> wrote:
+1
On Fri, Sep 25, 2015 at 6:59 PM, Nathan Goldbaum <nathan12343@gmail.com> wrote:
On Fri, Sep 25, 2015 at 12:55 PM, Matthew Turk <matthewturk@gmail.com> wrote:
Hi all,
Before I get too far into writing up a YTEP, I wanted to know if anyone would object to disabling "caching" of datasets in-memory.
Right now, if you do:
ds1 = yt.load("something") ds2 = yt.load("something")
"ds1 is ds2" will return True; this was done so that you could load a dataset, then unpickle an object with a reference to that dataset, and get the dataset back rather than a newly-loaded one. But I think a) folks don't pickle much b) it causes a lot of problems. So I'm wondering if it would be okay to just turn this off.
Big +1 to this. It's been the cause of a number of head-scratchers for me. It also motivated the parts of this change that touch domain_left_edge and domain_right_edge, which is IMO a bit silly but necessary given the way the caching works:
https://bitbucket.org/ngoldbaum/yt/commits/dd8d2317489904eb14b24cb63d7e8cfd0...
-Matt _______________________________________________ yt-dev mailing list yt-dev@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
_______________________________________________ yt-dev mailing list yt-dev@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
_______________________________________________ yt-dev mailing list yt-dev@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
-- Cameron Hummels NSF Postdoctoral Fellow Department of Astronomy California Institute of Technology http://chummels.org
+1! John ZuHone Kavli Center for Astrophysics and Space Research Massachusetts Institute of Technology 77 Massachusetts Ave., 37-582G Cambridge, MA 02139 (w) 617-253-2354 (m) 781-708-5004 jzuhone@space.mit.edu jzuhone@gmail.com http://www.jzuhone.com
On Sep 25, 2015, at 3:04 PM, Cameron Hummels <chummels@gmail.com> wrote:
+1
On Fri, Sep 25, 2015 at 11:16 AM, Britton Smith <brittonsmith@gmail.com> wrote: +1
On Fri, Sep 25, 2015 at 6:59 PM, Nathan Goldbaum <nathan12343@gmail.com> wrote:
On Fri, Sep 25, 2015 at 12:55 PM, Matthew Turk <matthewturk@gmail.com> wrote: Hi all,
Before I get too far into writing up a YTEP, I wanted to know if anyone would object to disabling "caching" of datasets in-memory.
Right now, if you do:
ds1 = yt.load("something") ds2 = yt.load("something")
"ds1 is ds2" will return True; this was done so that you could load a dataset, then unpickle an object with a reference to that dataset, and get the dataset back rather than a newly-loaded one. But I think a) folks don't pickle much b) it causes a lot of problems. So I'm wondering if it would be okay to just turn this off.
Big +1 to this. It's been the cause of a number of head-scratchers for me. It also motivated the parts of this change that touch domain_left_edge and domain_right_edge, which is IMO a bit silly but necessary given the way the caching works:
https://bitbucket.org/ngoldbaum/yt/commits/dd8d2317489904eb14b24cb63d7e8cfd0...
-Matt _______________________________________________ yt-dev mailing list yt-dev@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
_______________________________________________ yt-dev mailing list yt-dev@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
_______________________________________________ yt-dev mailing list yt-dev@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
-- Cameron Hummels NSF Postdoctoral Fellow Department of Astronomy California Institute of Technology http://chummels.org _______________________________________________ yt-dev mailing list yt-dev@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
+1 John On 09/25/2015 01:55 PM, Matthew Turk wrote:
Hi all,
Before I get too far into writing up a YTEP, I wanted to know if anyone would object to disabling "caching" of datasets in-memory.
Right now, if you do:
ds1 = yt.load("something") ds2 = yt.load("something")
"ds1 is ds2" will return True; this was done so that you could load a dataset, then unpickle an object with a reference to that dataset, and get the dataset back rather than a newly-loaded one. But I think a) folks don't pickle much b) it causes a lot of problems. So I'm wondering if it would be okay to just turn this off.
-Matt _______________________________________________ yt-dev mailing list yt-dev@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
-- John Wise Assistant Professor of Physics Center for Relativistic Astrophysics, Georgia Tech http://cosmo.gatech.edu
participants (6)
-
Britton Smith -
Cameron Hummels -
John Wise -
John ZuHone -
Matthew Turk -
Nathan Goldbaum