[plt-dev] flvector

On Dec 21, 2009, at 10:44 AM, Noel Welsh <noelwelsh at gmail.com> wrote:

> On Sat, Dec 19, 2009 at 2:54 PM, Matthew Flatt <mflatt at cs.utah.edu>  
> wrote:
>>
>> I can see the use for many flavors of vectors when dealing with  
>> foreign
>> libraries. But is there another reason to bother with "signed 16-bit
>> integer" instead of just using exact integers?
>>
>
> I think arrays of doubles are by far the most common requirement,
> followed by arrays of ints, then distantly followed by floats, 16-bit
> ints and bytes in that order.

FWIW, I agree with Noel's assessment with the exception that I think  
bytes are  more important: I use byte arrays all the time for working  
with images. To simplify life, one often works with int arrays  
(especially when using a high level language), but things like masking  
work out just fine on 8-bit values.

I'd personally like to see support for all of these array types as the  
initial unsafe ops support has already been very beneficial to me.  
I've been able to get 2-4x speedups on things like path planning code  
with very little coding investment. The fact that bugs cause DrScheme  
to bomb out is a downside, and I'd definitely appreciate some  
seperation between the editor and the evaluation process, but the  
performance improvements are more than welcome.

Thanks, Matthew!

Anthony