On Thu, Oct 30, 2008 at 02:49:16PM -, Stephen Pelc wrote:
> There's nothing that makes a program in one language execute any
> faster than one in another language other than the quality of
> the implementation of the compiler. That in turn assumes that
> performance is the primary objective.
>> That's one of the problems with performance optimisation. You cannot
>> optimise everything (for a large software project you never have time)
>> so
>> you have to work on the areas you think are important / used the most.
>> Of
>> course if you get the wrong areas, then it doesn't help ;).
>
>
Rob Malpass said:
> Fred said that one reason he liked Forth was that it was
> fast - which got me thinking
> Nowadays would we refer to one language being any faster than another? I
> guess we can exclude from this debate the speed of compilation. I'd
> guess for most (non-climate-model
On Thu, Oct 30, 2008 at 08:30:02AM -, Stephen Rowles wrote:
> > Things like the Hot Spot compiler in Java, which can re-optimise
> > byte code as the program is running depending on how its actually
> > being used can complicate things (ending up with Java performing
> > better than C in *some*
> If the app requires that the garbage be collected from time to time
(i.e.
> it uses that heap), then you have an effective priority inversion: the
> high-priority app will be prevented from operating until the
low-priority
> GC has completed.
>
> So all this does is to shift the problem, not fix
> It's not uncommon to find the GC running as a lower priority task to
> claim real-time improvements. Still isn't all that predictable though.
If the app requires that the garbage be collected from time to time (i.e.
it uses that heap), then you have an effective priority inversion: the
high-pri
> It's amazing how many people jump into "optimisation" without having
> collected any metrics on their code first...
Similarly, it is astonishing to see how some people don't *look* at the
metrics. A long while back, I instrumented a chunk of UI code (using
Quantify, IIRC) which gave a "clear" i
> GC in real time applications can be a bad idea, though I believe
> there are real-time versions of garbage collectors out there.
It's not uncommon to find the GC running as a lower priority task to
claim real-time improvements. Still isn't all that predictable though.
Peter
***
> That's one of the problems with performance optimisation. You cannot
> optimise everything (for a large software project you never have time) so
> you have to work on the areas you think are important / used the most. Of
> course if you get the wrong areas, then it doesn't help ;).
It's amazing
> Things like the Hot Spot compiler in Java, which can re-optimise
> byte code as the program is running depending on how its actually
> being used can complicate things (ending up with Java performing
> better than C in *some* circumstances).
This is something that is particularly relevant for la
On Wednesday 29 October 2008 20:53:06 Vic wrote:
> > So once a program is compiled - is any language faster than any other
> > nowadays? Two programs: one in C, one in Java doing the identical job -
> > is there anything anywhere that says one will always be faster for
> > non-trivial application
Hi,
On Wed, 29 Oct 2008, Vic wrote:
(quoting Rob Malpass)
> > So once a program is compiled - is any language faster than any other
> > nowadays? Two programs: one in C, one in Java doing the identical job -
> > is there anything anywhere that says one will always be faster for
> > non-trivial
> I'd
> guess for most (non-climate-modelling) applications - compilation speed on
> modern hardware isn't much of an overhead.
Go rebuild something like OpenOffice or gcc4 and try to make the same
claim...
> So once a program is compiled - is any language faster than any other
> nowadays? Two
Hi all
I watched an interesting programme on youtube the other day - it was an episode
of Micro Live from 1986 [1]. Fred Harris (remember him?) was talking about
Forth and doing some compiling on the Acorn Electron. The team were comparing
things like Basic to C to Forth. Fred said that o
14 matches
Mail list logo
