www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Re: We need to clarify if 'real' is the 'default floating point type' or not.

reply renoX <renosky free.fr> writes:
Ameer Armaly Wrote:

 Maybe we just ought to make int 
 whatever type is fastest, and call the current int something different, 
 maybe dword?


So int would mean 8 bit integers?
I'm joking but for some operation 8 bit int *are* 'fastest' than 32 bit int,
because they use less cache, memory..

Anything else that intX, uintX is necessarily arbitrary (and even intX is not
that simple: some DSP have 24bit integers..), so I'm not sure that it's quite
useful to spend too much time on that topic.

renoX
Mar 04 2008
parent reply "Ameer Armaly" <ameer.armaly furman.edu> writes:
"renoX" <renosky free.fr> wrote in message 
news:fqj8i1$1m01$1 digitalmars.com...

 Ameer Armaly Wrote:

 Maybe we just ought to make int
 whatever type is fastest, and call the current int something different,
 maybe dword?


 So int would mean 8 bit integers?
 I'm joking but for some operation 8 bit int *are* 'fastest' than 32 bit 
 int, because they use less cache, memory..

 Anything else that intX, uintX is necessarily arbitrary (and even intX is 
 not that simple: some DSP have 24bit integers..), so I'm not sure that 
 it's quite useful to spend too much time on that topic.


I don't know why I didn't think of this earlier: auto. If you want optimal 
int types, isn't that part of the  rationale behind type inference?

 renoX
Mar 04 2008
parent reply JMNorris <nospam nospam.com> writes:
"Ameer Armaly" <ameer.armaly furman.edu> wrote in
news:fqjl0e$2mi4$1 digitalmars.com: 


 I don't know why I didn't think of this earlier: auto. If you want
 optimal int types, isn't that part of the  rationale behind type
 inference? 


I assume you're proposing a change in the way auto works.  Currently for:

    auto x = 5 < 7;

x has type bool.
Mar 06 2008
parent "Ameer Armaly" <ameer.armaly furman.edu> writes:
"JMNorris" <nospam nospam.com> wrote in message 
news:fqonmv$25gh$1 digitalmars.com...

 "Ameer Armaly" <ameer.armaly furman.edu> wrote in
 news:fqjl0e$2mi4$1 digitalmars.com:


 I don't know why I didn't think of this earlier: auto. If you want
 optimal int types, isn't that part of the  rationale behind type
 inference?


 I assume you're proposing a change in the way auto works.  Currently for:

    auto x = 5 < 7;

 x has type bool.


Not really. Suppose you're on a 64-bit machine.
auto x = 5;
should yield a long, where as on a 32-bit machine it should yield an int
It's not perfect by any stretch, but it's easier given the current 
circumstances.
Mar 06 2008