[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Chicken-users] libcurl?
Re: [Chicken-users] libcurl?
Mon, 3 Oct 2005 16:07:00 -0500
On Mon, Oct 03, 2005 at 10:06:58PM +0200, felix winkelmann wrote:
> On 9/29/05, Ed Watkeys <address@hidden> wrote:
> > This is my biggest problem with many Scheme implementations. As
> > offensive as it is to many Schemers, SMP and multicore CPUs will be
> > the main sources of processor performance improvement for the
> > foreseeable future.
> I think all this hype (yes, hype!) about going SMP and multi-whatever
> is something that should be taken with a bucket or two of salt.
> The question is not whether machines capable of executing multiple
> threads will be mainstream - they just will. The question is how we
> are going to change our programming habits. And one thing I know:
I think it might be hype, but only for part of the market. The average
joe using his computer for downloading mp3s and balancing his checkbook
isn't going to care about multi-core processing. But scientists and
financial institutiona care a lot. And these people have buildings full
of computers on every floor packed as tightly as the fire marshalls
allow. Since many economies are basically number cruched on computers,
performance will always be a goal for people to important to ignore.
> The programming skills of conemporary software developer of are not
> even suffficient to write robust single threaded code. Period.
I totally agree. Though the reasoning behind why is something I could
debate a long time. :)
> Let's face it: software is crap. Feature-laden and bloated, written under
> tremendous time-pressure, often by incapable coders, using dangerous
> languages and inadequate tools, trying to connect to heaps of broken
> or obsolete protocols, implemented equally insufficiently, running
> on unpredictable hardware, we are all more than used to brokenness.
> Adding multithreading to this will not improve the standard of software
> engineering, it will just make it worse.
It will definitely make it worse. :( I follow the gaming technology
industry closely (I did computer graphics as a hobby for a long time)
and they've been talking about multithreading issues for 4 or so years
now. As far as I can tell, most of them are very afraid of it since
language/tool threading for the primary language they use (C/C++) is
atrocious, people really have no idea how to code like that(conditional
variables--whazzat?), and customers are buying SMP machines and demanding
appropriately scaled game technology right now.
> User-level threads have at least better determinism, can
> be debugged properly and alllow for _less_ synchronization overhead.
Yeah, but they only run on one processor...
> - Do we need more performance, seriously?
The average joe, maybe not. Physicists, financial companies, game/world
simulation, and sociologists--you bet.
I think, given human imagination (when it isn't stomped into pieces by
the drudgery of existance), if more and more computing power is presented,
more uses will be found for it until the last true barrier to programming
will be the act of content generation itself.