A word of caution tho: I used malloc because I needed >1GB pointers,
which resulted in random crashes that took me many months to track
down.  I went back to NewMemoryblock (and used arrays of <=1GB pointer
instead) and the crashes disappeared.  I never figured out why, but it
was a statistical process of elimination that was extremely painful.

My lesson: don't use anything but NewMemoryblock in the RB app (using
malloc and vec_malloc in plugins is fine).

P.


On 4/20/07, Frank Condello <[EMAIL PROTECTED]> wrote:
> On 20-Apr-07, at 8:53 AM, [EMAIL PROTECTED] wrote:
> >
> > El 20/04/2007, a las 6:23, Frank Condello escribió:
> >
> >> I'm pretty sure NewPtr calls down to malloc eventually (in newer
> >> versions of OS X at least) but it might do some other nasty stuff for
> >> backwards compatibility. Regardless, I don't foresee ever needing
> >> allocations of that size - I just need guaranteed alignment.
> >>
> >> Frank.
> >> <http://developer.chaoticbox.com/>
> >> <http://macgameblog.com/>
> >>
> >> On 19-Apr-07, at 10:01 PM, Peter K. Stys wrote:
> >>
> >>> I believe NewMemoryblock uses Apple's NewPtr call so you can check
> >>> there (I had a problem with 1GB block limits and this was the
> >>> reason;
> >>> malloc has no such limit, or at least 2-4GB, ASFAIK)
> >>>
> >>> Peter.
> >>>
> >>> On 4/19/07, Frank Condello <[EMAIL PROTECTED]> wrote:
> >>>> I've started declaring into Accelerate.framework and have read that
> >>>> Mac OS X's malloc always uses 16-byte alignment, but I was
> >>>> wondering
> >>>> if I could rely on this alignment from an RB-supplied
> >>>> MemoryBlock. I
> >>>> know this has been discussed before but I couldn't find a
> >>>> definitive
> >>>> yes or no - I suspect (and hope) the answer is yes, if not, I've
> >>>> got
> >>>> a bit more work to do!
> >>>>
> >>>> TIA,
> >>>> Frank.
> >
> > Not sure if it helps...... but in http://www.nilobject.com/page/3/
> > can I read:
> > "On Mac OS X malloc automatically aligns 16 bytes and pads out
> > allocations to 16-byte offsets.....:"
>
> Apple's docs state that clearly as well, but it doesn't guarantee
> that RB's memoryblock is always aligned. I suspect RB's memoryblock
> *is* aligned on Mac OS X but the lack of a definitive answer drove me
> to write my own light weight memoryblock class last night. Problem
> solved...
>
> Thanks to everyone for their input!
>
> Frank.
> <http://developer.chaoticbox.com/>
> <http://macgameblog.com/>
>
>
> _______________________________________________
> Unsubscribe or switch delivery mode:
> <http://www.realsoftware.com/support/listmanager/>
>
> Search the archives:
> <http://support.realsoftware.com/listarchives/lists.html>
>


-- 
-------------------------------------------------------------------------------
Peter K. Stys, MD
Professor of Medicine(Neurology), Senior Scientist
Ottawa Health Research Institute, Div. of Neuroscience
Ottawa Hospital / University of Ottawa
Ontario, CANADA
tel:    (613)761-5444
fax:    (613)761-5330
http://www.ohri.ca/profiles/stys.asp
-------------------------------------------------------------------------------
_______________________________________________
Unsubscribe or switch delivery mode:
<http://www.realsoftware.com/support/listmanager/>

Search the archives:
<http://support.realsoftware.com/listarchives/lists.html>

Reply via email to