>> 
>> What is [UInt32] for ObjC? A C-array of uint32_t? 
> 
> Well, the OBj-C sort gets a Swift array and treats it as as a C-array of 
> UInt32. Same as the Swift version.
> 

I think that counts as comparing Swift to C. Perhaps Swift Arrays with a 
generic of uint32_t should in fact optimize down to the same routine written in 
C for sorting uint32_t's or hand-tuned assembler, but that seems a little 
optimistic.  I'd say both this and the Swift Array <--> NSArray test are valid, 
the latter is more a real-world case, unless you in fact use C arrays all over 
your current Cocoa code, most people don't. 


>>> 2. the build-in Swift function sorted(array) crashes with an array of size 
>>> 10 million and values in the range 0 ... 100. Probably due to excessive 
>>> recursion.
>>> 3. the build-in Swift function sorted(array) with an array of size 10 
>>> million and values in the range 0 ... 1000 is about 100 times slower than 
>>> my own quickSort.
>>> 
>> 
>> 2) bad, I think I read something on the dev forums about sorted() running 
>> out of stack space but I thought that was when run on a thread. 
>> 
>> 3) Not brilliant either - what optimization flags were you using for all 
>> this? 
> 
> The default release ones: ObjC: Fastest, Smallest and Swift: Fastest.

Isn't there an 'unchecked' option somewhere, or is that still aliased to 
Fastest? 


_______________________________________________

Cocoa-dev mailing list ([email protected])

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:
https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com

This email sent to [email protected]

Reply via email to