Andy Switala <[EMAIL PROTECTED]> wrote:
> I found this online: http://unixhelp.ed.ac.uk/CGI/man-cgi?posix_memalign.
Thanks to you and Steve for the docs.
I've now a test and platform code for both flavors of memalign. I'll
send it later.
leo
[EMAIL PROTECTED]:~/parrot/languages/perl6% man memalign
POSIX_MEMALIGN(3) Linux Programmer's ManualPOSIX_MEMALIGN(3)
NAME
posix_memalign, memalign, valloc - Allocate aligned memory
SYNOPSIS
#include
int posix_memalign(void **memptr, size_t alignment, size_t size);
Dan Sugalski wrote:
[ zerofilled aligned memory ]
That, as much as anything, argues for an entry in platform.c to get and
return large sections of memory. I know it's reasonably doable on a lot
of platforms, just potentially differently everywhere.
BTW the zero filled memory is not that impor
n't check myself.)
--Andy
-Original Message-
From: Dan Sugalski [mailto:[EMAIL PROTECTED]
Sent: Thursday, May 29, 2003 08:16
To: [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Subject: Re: Add #22337
At 11:00 PM +0200 5/27/03, Leopold Toetsch wrote:
>Dan Sugalski <[EMAIL PROTECTED]&
At 11:00 PM +0200 5/27/03, Leopold Toetsch wrote:
Dan Sugalski <[EMAIL PROTECTED]> wrote:
At 8:03 PM +0200 5/27/03, Leopold Toetsch wrote:
... . We need some tests, from which size memory is
cleard for malloc and memalign.
I tossed the memset for now and saved ~450.000 L2-misses or ~0.2 s.
Whi
At 5:43 PM +0200 5/28/03, Leopold Toetsch wrote:
Leopold Toetsch <[EMAIL PROTECTED]> wrote:
... . We need some tests, from which size memory is
cleard for malloc and memalign.
Here is a small program, which could be put into a test.
Are there systems out there, without memalign, where malloc.c ca
Appended is a refined version of #22337 which it obsoletes.
Key features are:
1) DOD flags (live, on_free_list, ...) are kept in the arenas, 1 nibble
per object
2) arena memory is aquired per memalign() to be able to calculate arena
from object address
3) free_list is per arena now
4) PMC size i
Leopold Toetsch <[EMAIL PROTECTED]> wrote:
>>>... . We need some tests, from which size memory is
>>>cleard for malloc and memalign.
Here is a small program, which could be put into a test.
Are there systems out there, without memalign, where malloc.c can not be
linked with?
/*
* test clean me
Dan Sugalski <[EMAIL PROTECTED]> wrote:
> At 8:03 PM +0200 5/27/03, Leopold Toetsch wrote:
>>... . We need some tests, from which size memory is
>>cleard for malloc and memalign.
>>I tossed the memset for now and saved ~450.000 L2-misses or ~0.2 s.
> While I didn't see any memsets in smallobject
At 3:10 PM -0400 5/27/03, Uri Guttman wrote:
> "DS" == Dan Sugalski <[EMAIL PROTECTED]> writes:
>> obviously gets called for memalign - at least for this arena size)
>> does clear the memory. We need some tests, from which size memory is
>> cleard for malloc and memalign.
>> I tossed
> "DS" == Dan Sugalski <[EMAIL PROTECTED]> writes:
>> obviously gets called for memalign - at least for this arena size)
>> does clear the memory. We need some tests, from which size memory is
>> cleard for malloc and memalign.
>> I tossed the memset for now and saved ~450.000 L2-miss
At 8:03 PM +0200 5/27/03, Leopold Toetsch wrote:
Some additional remarks:
The "memset" in smallobject.c is not necessary on Linux. mmap() (which
obviously gets called for memalign - at least for this arena size)
does clear the memory. We need some tests, from which size memory is
cleard for mallo
Some additional remarks:
The "memset" in smallobject.c is not necessary on Linux. mmap() (which
obviously gets called for memalign - at least for this arena size)
does clear the memory. We need some tests, from which size memory is
cleard for malloc and memalign.
I tossed the memset for now and s
13 matches
Mail list logo
