Mailing Lists: Apple Mailing Lists
Image of Mac OS face in stamp
 
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: optimised memcpy


On May 23, 2007, at 11:16 AM, Brett George wrote: 
On May 23, 2007, at 11:07 AM, Brett George wrote:

I am using memcpy on an x86 machine and finding it to be quite slow
compared to other optimised routines.


I am moving 16384 words of 32-bit floating point data and it's taking
around 49us.


Your mention of floating point leads me to verify whether you're talking
about the memcpy in the kernel or the one in userland?

This is a floating point library in the kernel.

Those two terms ("floating point" and "kernel") just don't belong together.

However, bcopy in both the kernel and userland are quite heavily optimised. My math says:

% expr ((16384 * 4) / 0.000049) / 1024 / 1024 / 1024
1.24561543297

i.e. a bit over 1.2GiB/sec. Which hardware are you running on, and how is it configured?

 = Mike

_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-drivers mailing list      (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/darwin-drivers/email@hidden

This email sent to email@hidden
References: 
 >optimised memcpy (From: "Brett George" <email@hidden>)
 >Re: optimised memcpy (From: Garth Cummings <email@hidden>)
 >Re: optimised memcpy (From: Brett George <email@hidden>)



Visit the Apple Store online or at retail locations.
1-800-MY-APPLE

Contact Apple | Terms of Use | Privacy Policy

Copyright © 2007 Apple Inc. All rights reserved.