[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: svn commit: r1078357 - in /subversion/trunk: ./ build.conf subversion/libsvn_fs_fs/caching.c subversion/libsvn_fs_fs/fs.h subversion/libsvn_fs_fs/fs_fs.c subversion/svnadmin/main.c

From: Hyrum K Wright <hyrum_at_hyrumwright.org>
Date: Mon, 7 Mar 2011 12:29:53 -0600

On Mon, Mar 7, 2011 at 10:36 AM, Hyrum K Wright <hyrum_at_hyrumwright.org> wrote:
> On Mon, Mar 7, 2011 at 6:11 AM, Philip Martin
> <philip.martin_at_wandisco.com> wrote:
>> stefan2_at_apache.org writes:
>>
>>> Author: stefan2
>>> Date: Sat Mar  5 21:18:33 2011
>>> New Revision: 1078357
>>
>>> Modified: subversion/trunk/subversion/svnadmin/main.c
>>> URL: http://svn.apache.org/viewvc/subversion/trunk/subversion/svnadmin/main.c?rev=1078357&r1=1078356&r2=1078357&view=diff
>>> ==============================================================================
>>> --- subversion/trunk/subversion/svnadmin/main.c (original)
>>> +++ subversion/trunk/subversion/svnadmin/main.c Sat Mar  5 21:18:33 2011
>>> @@ -290,6 +290,11 @@ static const apr_getopt_option_t options
>>>       N_("use format compatible with Subversion versions\n"
>>>          "                             earlier than 1.7")},
>>>
>>> +    {"memory-cache-size",     'M', 1,
>>> +     N_("size of the extra in-memory cache in MB used to\n"
>>> +        "                             minimize redundant operations. Default: 256.\n"
>>> +        "                             [used for FSFS repositories only]")},
>>> +
>>>      {NULL}
>>>    };
>>
>>
>>>
>>> @@ -1650,6 +1656,7 @@ main(int argc, const char *argv[])
>>>    /* Initialize opt_state. */
>>>    opt_state.start_revision.kind = svn_opt_revision_unspecified;
>>>    opt_state.end_revision.kind = svn_opt_revision_unspecified;
>>> +  opt_state.memory_cache_size = 0x10000000; /* 256 MB */
>>
>> 256MB is a magic number.  How was its value determined?
>>
>> It causes the regression tests to fail on machines with 2GB of memory
>> when the tests are run in parallel, as they cannot allocate 10*256MB.
>>
>> There appears to be a significant overhead setting up the cache.
>> Setting the magic number to 16MB makes the tests noticeably faster,
>> setting it to 1MB makes the tests faster still.
>
> FWIW, my test run times increased by ~25% over the weekend.  On IRC,
> Philip speculated that this is a potential cause.

r1078884 reset this value to 1MB, so as to keep people working. We
should probably discuss the appropriate values for cache sizes in
difference scenarios, and get things fixed up prior to release.

-Hyrum
Received on 2011-03-07 19:30:27 CET

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.