www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - D, Intel and AMD

reply "Chris" <wendlec tcd.ie> writes:
Are there any major differences in how D fares on AMD processors 
as opposed to Intel? The usual story is that Intel is faster and 
more precise when it comes to arithmetic operations (although I 
suspect that Intel is also very good at PR ;) Has anyone come 
across D related issues that are peculiar to one or the other 
architecture or brand?
Oct 29 2013
next sibling parent Iain Buclaw <ibuclaw ubuntu.com> writes:
On 29 October 2013 20:35, Chris <wendlec tcd.ie> wrote:


 Are there any major differences in how D fares on AMD processors as
 opposed to Intel? The usual story is that Intel is faster and more precise
 when it comes to arithmetic operations (although I suspect that Intel is
 also very good at PR ;) Has anyone come across D related issues that are
 peculiar to one or the other architecture or brand?



The only peculiarity I've found is that code ran on Intel Atom is slower
with the -O switch on DMD.  (There's a bug report somewhere for that).

-- 
Iain Buclaw

*(p < e ? p++ : p) = (c & 0x0f) + '0';
Oct 30 2013
prev sibling next sibling parent "John Colvin" <john.loughran.colvin gmail.com> writes:
On Tuesday, 29 October 2013 at 20:35:26 UTC, Chris wrote:

 Are there any major differences in how D fares on AMD 
 processors as opposed to Intel? The usual story is that Intel 
 is faster and more precise when it comes to arithmetic 
 operations (although I suspect that Intel is also very good at 
 PR ;) Has anyone come across D related issues that are peculiar 
 to one or the other architecture or brand?


I can't think of anything intrinsic to D that would result in any 
difference when compared to similar c++ programs. DMD itself may 
sometimes make some poor decisions from an AMD perspective (as 
Ian mentioned), but if you're looking for every last drop of 
performance you will want to be using ldc or gdc anyway.
Oct 30 2013
prev sibling parent reply "Kagamin" <spam here.lot> writes:
On Tuesday, 29 October 2013 at 20:35:26 UTC, Chris wrote:

 Are there any major differences in how D fares on AMD 
 processors as opposed to Intel? The usual story is that Intel 
 is faster and more precise when it comes to arithmetic 
 operations (although I suspect that Intel is also very good at 
 PR ;) Has anyone come across D related issues that are peculiar 
 to one or the other architecture or brand?


I once changed an Athlon-based machine to a Celeron-based. If I 
remember dhrystone benchmarks correctly, the drop in ALU 
performance was significant.
Oct 30 2013
parent "evilrat" <evilrat666 gmail.com> writes:
On Wednesday, 30 October 2013 at 17:16:17 UTC, Kagamin wrote:

 On Tuesday, 29 October 2013 at 20:35:26 UTC, Chris wrote:

 Are there any major differences in how D fares on AMD 
 processors as opposed to Intel? The usual story is that Intel 
 is faster and more precise when it comes to arithmetic 
 operations (although I suspect that Intel is also very good at 
 PR ;) Has anyone come across D related issues that are 
 peculiar to one or the other architecture or brand?


 I once changed an Athlon-based machine to a Celeron-based. If I 
 remember dhrystone benchmarks correctly, the drop in ALU 
 performance was significant.


but shouldn't Celeron compared to Sempron, and not Athlon? O_o
Oct 30 2013