Re: Fast hash of NSData?
Re: Fast hash of NSData?
- Subject: Re: Fast hash of NSData?
- From: Graham Cox <email@hidden>
- Date: Sat, 30 Nov 2013 11:37:57 +0100
On 30 Nov 2013, at 12:41 am, Kyle Sluder <email@hidden> wrote:
> That’s not a good idea. If the file on disk changes out from under you, the cache no longer reflects the data.
That can’t happen - the image data is embedded in the file stream at the moment, not held as a separate file. While I intend to change to that at some point, I don’t do it now. When I do, I’ll also (have to) change the way that the image references in the main stream are written. So the hash approach is only really a stop-gap - it gets me better performance without changing the basic architecture.
@Scott Ribe:
> I recently (last week, no kidding) investigated this question, and went with murmur hash.
I went with a 64-bit FNV-1a, which looks on a par with murmur, according to the page linked by Eric: http://programmers.stackexchange.com/questions/49550/which-hashing-algorithm-is-best-for-uniqueness-and-speed. Which admittedly I am taking at face value, having only profiled my own code for speed, not randomness, etc. The main reason being that an implementation of FNV-1a was right there on the page :)
—Graham
_______________________________________________
Cocoa-dev mailing list (email@hidden)
Please do not post admin requests or moderator comments to the list.
Contact the moderators at cocoa-dev-admins(at)lists.apple.com
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden