....when apropos takes a minute to search for matches?
And speaking of minutes, it has been quite a few waiting for loop to
count all the symbols. I wonder why it is so slow, doing maybe
1.5m/minute. And what the hell does apropos know that loop does not
about zooming through all the symbols?
Up to 6.2m now. Is that too many? How would that impact performance? It
is easier asking than Just Timing It.
OK, 6.5m symbols. That's a lot, right?
And where should I put my money in my next computer?
Is AMD any different than Intel, or shouls I save $150?
I know Vista is stupid, but is it slow? I'll be getting a lot of RAM.
How cores? If it is just me doing Web2.0 programming, how many can I
use? Should I spring for the quad core deal?
What about 64-bit vs 32-bit? Any advantage to the former? I do not need
more than 4g ram.
peace,kzo
On 2008-10-30, Kenny <·········@gmail.com> wrote:
> And where should I put my money in my next computer?
The empty bay below the DVD burner.
> Is AMD any different than Intel, or shouls I save $150?
I don't know about ``then'', but different ``from'', certainly.
> I know Vista is stupid, but is it slow? I'll be getting a lot of RAM.
I've been running Vista for like a year and a half. Yes, it is slow, but
the reason for that is all the crap that is enabled by default.
I.e. it's slow in the default setup, due to stupid reasons that
don't appear to have anything to do with the kernel.
If you go through the list of services and scheduled tasks, and kill everything
superfluous and useless, it will be a lot faster.
For instance, I discovered that, out-of-the-box, Vista was running a full hard
drive defragmentation daily! And lots of other stupidities.
It also runs Microsoft Anti-Spyware (which has been renamed Windows Defender).
This seems to slows down the system quite a bit; Defender is among the things
that I killed. Also, I don't run any anti-virus bullshit. Anti-virus software
doesn't work, and infections are a dumb-user problem anyway.
99% of the virus problems out there are because people say yes to ``please run
this .exe to install the codec needed for viewing this porn file''.
So, quite literally, most Windows viruses are in fact sexually transmitted!
Kaz Kylheku wrote:
> On 2008-10-30, Kenny <·········@gmail.com> wrote:
>
>>And where should I put my money in my next computer?
>
>
> The empty bay below the DVD burner.
ba-dump-bump
>
>
>>Is AMD any different than Intel, or shouls I save $150?
>
>
> I don't know about ``then'', but different ``from'', certainly.
Never mind me, help this guy:
"Now that I have some vacation I all over sudden found myself making a
few quick buttons for..."
Reminds me when I asked my older brother how to spell what sounds like
"halfto". Hunh?, he replied. You, know, like in 'I halfto go to the
bathroom'.
>
>
>>I know Vista is stupid, but is it slow? I'll be getting a lot of RAM.
>
>
> I've been running Vista for like a year and a half. Yes, it is slow, but
> the reason for that is all the crap that is enabled by default.
>
> I.e. it's slow in the default setup, due to stupid reasons that
> don't appear to have anything to do with the kernel.
>
> If you go through the list of services and scheduled tasks, and kill everything
> superfluous and useless, it will be a lot faster.
Knowing me I'll take out something crucial, void the warranty, and
somehow launch an ICBM on Canada which is right next to Alaska so its
governor should be OK on international affairs.
>
> For instance, I discovered that, out-of-the-box, Vista was running a full hard
> drive defragmentation daily! And lots of other stupidities.
>
> It also runs Microsoft Anti-Spyware (which has been renamed Windows Defender).
> This seems to slows down the system quite a bit; Defender is among the things
> that I killed. Also, I don't run any anti-virus bullshit. Anti-virus software
> doesn't work, and infections are a dumb-user problem anyway.
>
> 99% of the virus problems out there are because people say yes to ``please run
> this .exe to install the codec needed for viewing this porn file''.
Well how the hell else am I going to see it?!
>
> So, quite literally, most Windows viruses are in fact sexually transmitted!
Nice catch. Back in my early days selling the Mac Algebra I did let a
college student borrow a computer for a while and then used it to cut my
floppies. Yep. Got a not-too awkward phone call from someone who caught
it with their software before installing. Fortunately was not selling
all that many just then, only had to make a half dozen more awkward
calls, but I do recall that I had in fact infected at least one
customer. The metaphor...hell, maybe it isn't even a metaphor.
kt
On Thu, 30 Oct 2008 21:12:32 +0000 (UTC), Kaz Kylheku <········@gmail.com> wrote:
> 99% of the virus problems out there are because people say yes to ``please run
> this .exe to install the codec needed for viewing this porn file''.
>
> So, quite literally, most Windows viruses are in fact sexually transmitted!
Heh, this belongs to a fortune cookie file somewhere :)
On 2008-12-20, Giorgos Keramidas <········@ceid.upatras.gr> wrote:
> On Thu, 30 Oct 2008 21:12:32 +0000 (UTC), Kaz Kylheku <········@gmail.com> wrote:
>> 99% of the virus problems out there are because people say yes to ``please run
>> this .exe to install the codec needed for viewing this porn file''.
>>
>> So, quite literally, most Windows viruses are in fact sexually transmitted!
>
> Heh, this belongs to a fortune cookie file somewhere :)
If you think so, write to the maintainer of the fortune program.
By the way, looking at my posting again, I have one more piece of
OT] performance-related advice about Vista.
It appears that somehow (upgrades?) Vista can end up disabling most of your
swap space.
The thing started throwing up low memory dialogs at me and its performance
degrading with increasing uptime.
In task manager, there appears to be lots of memory.
But it turned out that I had only 135 megs of swap. I was sure I had configured
a lot more, and that system ran for like two years without popping up these
dialogs.
It could have been this:
``The page file size may become alternately too small or too large when you
start Windows Server 2008 or Windows Vista if there is no available free disk
space, and the page file size is managed by the system.''
[ http://support.microsoft.com/kb/955635 ]
I /was/ nearly out of disk space at some point, but I can't remember if
if the page file size management had been set to automatic.
Kenny <·········@gmail.com> writes:
> OK, 6.5m symbols. That's a lot, right?
That's a lot. But... I just interned 11.6m symbols into package foobar:
(with-open-file (f "/usr/share/dict/words")
(loop for word = (read-line f nil nil)
while word
do (loop for i below 50
do (intern (format nil "~a~a" (string-upcase word) :i) :foobar))))
That took a few minutes, mostly because I forgot to quit the memory
hogs Safari and iPhoto before I told 64 bit LW for mac to do this (my
linux box has 4GB RAM, I tend to forget that 2GB on my MacBook is not
that much for experiments like this, and swapping against a 5400rpm
laptop disk is no fun).
But once it was done, apropos is fast:
FOOBAR 56 > (time (apropos :giraffe :foobar))
Timing the evaluation of (APROPOS :GIRAFFE :FOOBAR)
GIRAFFE8
[stuff deleted]
GIRAFFESQUE44
GIRAFFESQUE42
User time = 0.912
System time = 0.009
Elapsed time = 1.127
Allocation = 78752 bytes
10 Page faults
FOOBAR 57 >
> Is AMD any different than Intel, or shouls I save $150?
It's definitely different, but it depends on operating system and
lisp implementation, I guess.
For 64 bit LispWorks on linux, we get much better results with Core 2
than with Opteron - with 32 bit linux/lisp, Opteron and Core 2 are not
that much different. And depending on your application, 64 bit can give
you advantages even for applications with small memory footprint.
--
(espen)
Espen Vestre wrote:
> Kenny <·········@gmail.com> writes:
>
>
>>OK, 6.5m symbols. That's a lot, right?
>
>
> That's a lot. But... I just interned 11.6m symbols into package foobar:
>
> (with-open-file (f "/usr/share/dict/words")
> (loop for word = (read-line f nil nil)
> while word
> do (loop for i below 50
> do (intern (format nil "~a~a" (string-upcase word) :i) :foobar))))
>
> That took a few minutes, mostly because I forgot to quit the memory
> hogs Safari and iPhoto before I told 64 bit LW for mac to do this (my
> linux box has 4GB RAM, I tend to forget that 2GB on my MacBook is not
> that much for experiments like this, and swapping against a 5400rpm
> laptop disk is no fun).
>
> But once it was done, apropos is fast:
>
> FOOBAR 56 > (time (apropos :giraffe :foobar))
> Timing the evaluation of (APROPOS :GIRAFFE :FOOBAR)
> GIRAFFE8
> [stuff deleted]
> GIRAFFESQUE44
> GIRAFFESQUE42
> User time = 0.912
> System time = 0.009
> Elapsed time = 1.127
> Allocation = 78752 bytes
> 10 Page faults
Hmmm. I got cpu 6s, real 83s. Lots of giraffes, btw. A second search is
5.9/6.1, ie the first time I am waiting while the huge image underlying
the huge symbol count gets pulled in from disk.
Thanks for the trouble you took!
>
> FOOBAR 57 >
>
>
>>Is AMD any different than Intel, or shouls I save $150?
>
>
> It's definitely different, but it depends on operating system and
> lisp implementation, I guess.
>
> For 64 bit LispWorks on linux, we get much better results with Core 2
> than with Opteron - with 32 bit linux/lisp, Opteron and Core 2 are not
> that much different. And depending on your application, 64 bit can give
> you advantages even for applications with small memory footprint.
Cool. Thx again.
kt
Kenny <·········@gmail.com> writes:
> OK, 6.5m symbols. That's a lot, right?
That's a lot. But... I just interned 11.6m symbols into package foobar:
(with-open-file (f "/usr/share/dict/words")
(loop for word = (read-line f nil nil)
while word
do (loop for i below 50
do (intern (format nil "~a~a" (string-upcase word) i) :foobar))))
That took a few minutes, mostly because I forgot to quit the memory
hogs Safari and iPhoto before I told 64 bit LW for mac to do this (my
linux box has 4GB RAM, I tend to forget that 2GB on my MacBook is not
that much for experiments like this, and swapping against a 5400rpm
laptop disk is no fun).
But once it was done, apropos is fast:
FOOBAR 56 > (time (apropos :giraffe :foobar))
Timing the evaluation of (APROPOS :GIRAFFE :FOOBAR)
GIRAFFE8
[stuff deleted]
GIRAFFESQUE44
GIRAFFESQUE42
User time = 0.912
System time = 0.009
Elapsed time = 1.127
Allocation = 78752 bytes
10 Page faults
FOOBAR 57 >
> Is AMD any different than Intel, or shouls I save $150?
It's definitely different, but it depends on operating system and
lisp implementation, I guess.
For 64 bit LispWorks on linux, we get much better results with Core 2
than with Opteron - with 32 bit linux/lisp, Opteron and Core 2 are not
that much different. And depending on your application, 64 bit can give
you advantages even for applications with small memory footprint.
--
(espen)
Kenny wrote:
> ....when apropos takes a minute to search for matches?
>
> And speaking of minutes, it has been quite a few waiting for loop to
> count all the symbols. I wonder why it is so slow, doing maybe
> 1.5m/minute. And what the hell does apropos know that loop does not
> about zooming through all the symbols?
Did you compile the loop?
> OK, 6.5m symbols. That's a lot, right?
What does your code do? Intern symbols all day?
> And where should I put my money in my next computer?
Take a look at
http://arstechnica.com/guides.ars
In particular the System Guides. I personally think the sweet spot is
around 4GB RAM with a dual-core 64-bit CPU (and OS). Unless you're
willing to learn linux or OS X, Vista's probably the way to go.
- Daniel
D Herring wrote:
> Kenny wrote:
>
>> ....when apropos takes a minute to search for matches?
>>
>> And speaking of minutes, it has been quite a few waiting for loop to
>> count all the symbols. I wonder why it is so slow, doing maybe
>> 1.5m/minute. And what the hell does apropos know that loop does not
>> about zooming through all the symbols?
>
>
> Did you compile the loop?
Down to 1.4s cpu and real. Bouncing is a drag, but when I do I'll mebbe
test from scratch, see if the 80s deal aso goes away.
>
>
>> OK, 6.5m symbols. That's a lot, right?
>
>
> What does your code do? Intern symbols all day?
>
Not my code. Image-dumped hash tables are serving as a poor man's db.
>
>> And where should I put my money in my next computer?
>
>
> Take a look at
> http://arstechnica.com/guides.ars
>
> In particular the System Guides. I personally think the sweet spot is
> around 4GB RAM with a dual-core 64-bit CPU (and OS). Unless you're
> willing to learn linux or OS X, Vista's probably the way to go.
Awesome. Thx for such a precise recommendations. I love not having to
think, and this NG is great for that.......that might have come out
wrong....
kxo
Kenny <·········@gmail.com> wrote:
+---------------
| ....when apropos takes a minute to search for matches?
|
| And speaking of minutes, it has been quite a few waiting for loop to
| count all the symbols. I wonder why it is so slow, doing maybe
| 1.5m/minute.
+---------------
Hmmm... I only have a ~20K symbols lying around,
but it can count them pretty fast, ~29 Msymbol/s
[I did it 1000 times to get a reliable result]:
cmu> (time
(let ((count 0))
(dotimes (i 1000)
(do-all-symbols (ignore)
(declare (ignorable ignore))
(incf count)))
count))
; Compiling LAMBDA NIL:
; Compiling Top-Level Form:
; Evaluation took:
; 0.76f0 seconds of real time
; 0.747812f0 seconds of user run time
; 0.0f0 seconds of system run time
; 1,420,375,501 CPU cycles
; 0 page faults and
; 368,008 bytes consed.
;
22453000
cmu> (/ 22453000 0.76f0)
2.9543422e7
cmu>
How are you counting them? I know about DO-ALL-SYMBOLS (above),
but I don't know any way to use LOOP to count symbols in more
than one package at a time...
+---------------
| And what the hell does apropos know that loop does not
| about zooming through all the symbols?
+---------------
Hmmm... Let's see here... Well, in CMUCL, APROPOS loops over
(LIST-ALL-PACKAGES) and does a DO-SYMBOLS to iterate over each
package. Should be slower than the above, I'm guessing. Also,
we're gonna get a *bunch* of duplications, since several packages
(USE :CL) [though that really shouldn't affect the *rate*].
cmu> (time
(let ((count 0))
(dotimes (i 1000)
(dolist (p (list-all-packages))
(do-symbols (ignore p)
(declare (ignorable ignore))
(incf count))))
count))
; Compiling LAMBDA NIL:
; Compiling Top-Level Form:
; Evaluation took:
; 23.66f0 seconds of real time
; 23.278011f0 seconds of user run time
; 1.6f-5 seconds of system run time
; 43,886,881,567 CPU cycles
; 0 page faults and
; 368,008 bytes consed.
;
87208000
cmu> (/ 87208000 23.66f0)
3685883.3
cmu>
Whoa! That's a *lot* slower! And if done with LOOP:
cmu> (time
(let ((count 0))
(dotimes (i 10)
(dolist (p (list-all-packages))
(loop for sym being the symbols in p do (incf count))))
count))
; Compiling LAMBDA NIL:
; Compiling Top-Level Form:
; Evaluation took:
; 2.55 seconds of real time
; 2.474624 seconds of user run time
; 0.007834 seconds of system run time
; 4,723,603,806 CPU cycles
; [Run times include 0.14 seconds GC run time]
; 0 page faults and
; 15,292,712 bytes consed.
;
859200
cmu> (/ 859200 2.55)
336941.2
cmu>
(*Ouch!*) More than 10 *times* slower!
+---------------
| And where should I put my money in my next computer?
| Is AMD any different than Intel, or shouls I save $150?
+---------------
I tend to like AMD. The above was done with a 1.855 GHz Athlon.
+---------------
| I know Vista is stupid, but is it slow?
+---------------
Dunno. The above was done on FreeBSD. ;-} ;-}
-Rob
-----
Rob Warnock <····@rpw3.org>
627 26th Avenue <URL:http://rpw3.org/>
San Mateo, CA 94403 (650)572-2607
Kenny <·········@gmail.com> writes:
> And where should I put my money in my next computer?
http://www.apple.com
Helpfully,
Patrick
------------------------------------------------------------------------
S P Engineering, Inc. | Large scale, mission-critical, distributed OO
| systems design and implementation.
···@spe.com | (C++, Java, Common Lisp, Jini, middleware, SOA)
Patrick May wrote:
> Kenny <·········@gmail.com> writes:
>
>>And where should I put my money in my next computer?
>
>
> http://www.apple.com
Got one, thx. Well, five, but only two are OS X and only one of those
did not die because Apple makes crappy hardware these days. And no Mac
runs the AllegroCL Windows IDE... or is Vine that good?
kt
Kenny <·········@gmail.com> writes:
> Patrick May wrote:
>> Kenny <·········@gmail.com> writes:
>>
>>>And where should I put my money in my next computer?
>>
>>
>> http://www.apple.com
>
> Got one, thx. Well, five, but only two are OS X and only one of those
> did not die because Apple makes crappy hardware these days. And no Mac
> runs the AllegroCL Windows IDE... or is Vine that good?
Well, don't know about Vine. But there's Parallels and this definitely
does work very well. Used daily for professional work.
Another option: Bootcamp. This gives you native Windows on your Apple
Mac. Works really well, too. Used for high performance DSP stuff and
Software Defined Radio which really requires highest performance.
Result: One piece of damn pretty hardware and OS X plus two options to
run Windows. Plus running Linux using Parallels (also done over here for
porting stuff to Debian and Suse).
Need for a pure Windows machine: None left. Deal.
Frank
--
Frank Goenninger
"Don't ask me! I haven't been reading comp.lang.lisp long enough to
really know ..."
Frank GOENNINGER wrote:
> Kenny <·········@gmail.com> writes:
>
>
>>Patrick May wrote:
>>
>>>Kenny <·········@gmail.com> writes:
>>>
>>>
>>>>And where should I put my money in my next computer?
>>>
>>>
>>> http://www.apple.com
>>
>>Got one, thx. Well, five, but only two are OS X and only one of those
>>did not die because Apple makes crappy hardware these days. And no Mac
>>runs the AllegroCL Windows IDE... or is Vine that good?
>
>
> Well, don't know about Vine. But there's Parallels and this definitely
> does work very well. Used daily for professional work.
>
> Another option: Bootcamp. This gives you native Windows on your Apple
> Mac. Works really well, too. Used for high performance DSP stuff and
> Software Defined Radio which really requires highest performance.
>
> Result: One piece of damn pretty hardware and OS X plus two options to
> run Windows. Plus running Linux using Parallels (also done over here for
> porting stuff to Debian and Suse).
>
> Need for a pure Windows machine: None left. Deal.
Wait. The Mac I have now sounds like a 747 turbine after spinning up,
can't even hear The Voices let alone program a computer.
With Windows I can go to SuperQuietPC.com and get something I am not
even sure is on if it wasn't for the stupid OS I have to ignore.
Hmmm....
kt
From: Raffael Cavallaro
Subject: Re: Is it a bad sign...
Date:
Message-ID: <gego4d$271$1@aioe.org>
On 2008-10-31 20:19:18 -0400, Kenny <·········@gmail.com> said:
> Wait. The Mac I have now sounds like a 747 turbine after spinning up,
> can't even hear The Voices let alone program a computer.
>
> With Windows I can go to SuperQuietPC.com and get something I am not
> even sure is on if it wasn't for the stupid OS I have to ignore.
>
> Hmmm....
You were talking about spending money for a new machine right? So
you're comparing some wheezing relic of a mac to the new windows pc you
could buy? double-you tee f*ck?
Get a Mac laptop - they just updated them. As Patrick pointed out you
can run Windows natively with BootCamp, or virtualized with Parallels
or VMWare Fusion. Cross platform testing is made much simpler (want to
make sure the web version of your algebra software runs right on 5
different browsers on 3 different platforms on the same machine without
rebooting?)
My son is a junior in college doing a CS degree. Naturally there are
courses where he must use some windows software or other; others where
any *nix platform -like Mac OS X or linux - is fine. For gaming he
mostly plays WarCraft on MacOS X. His summer job was at a firm that
sells Windows only C# developer tools. He does all of this on a 15"
MacBook pro.
He grew up on Windows - all the school computers ran it - but after a
semester of Mac OS X he didn't want to go back. As he put it, "of all
the mainstream OSes it gets in your way the least."
Raffael Cavallaro wrote:
> On 2008-10-31 20:19:18 -0400, Kenny <·········@gmail.com> said:
>
>> Wait. The Mac I have now sounds like a 747 turbine after spinning up,
>> can't even hear The Voices let alone program a computer.
>>
>> With Windows I can go to SuperQuietPC.com and get something I am not
>> even sure is on if it wasn't for the stupid OS I have to ignore.
>>
>> Hmmm....
>
>
> You were talking about spending money for a new machine right? So you're
> comparing some wheezing relic of a mac to the new windows pc you could
> buy? double-you tee f*ck?
It is not clear to me that age is a factor. I would not necessarily
expect a PC to be quiet, and as I did last time would seek out one
advertised to be quiet.
The Macs do not even discuss noise, and a desktop starts at $2800.
Wtf!!! Oh, you used that. An iMac puts the whole noisy animal 24" away
with no place to hide. And I do not need a flat panel, thank you, I
have four.
>
> Get a Mac laptop - they just updated them.
checking prices...working...working...oh my.
Uh, what was the point again? How did this tail start wagging that? I
live in a lisp image and now Al Gore's Internet, a PC will be fine.
Macs are great, tho. When I retire from programming I will grab one in
two seconds.
kt
Kenny <·········@gmail.com> writes:
> [...]
> Wait. The Mac I have now sounds like a 747 turbine after spinning up,
> can't even hear The Voices let alone program a computer.
>
> With Windows I can go to SuperQuietPC.com and get something I am not
> even sure is on if it wasn't for the stupid OS I have to ignore.
My Macbook Pro laptop is by far the quietest computer I've ever used
or heard. I put it on a Griffin Elevator stand so that the screen is
at a similar height to my external monitor - it makes a great
development machine and it's portable to boot. Time Machine backups
and SuperDuper! cloning are awesome.
I keep my Ubuntu server in another room so my office is *super* quiet.
Of course I was switching from Linux to FreeBSD with a pretty face. I
really can't recommend buying a Mac just to run Windows in Parallels
or VMWare Fusion or boot camp. It would work fine, but you'd be paying
a premium for OS X and not using it. On the other hand, a MBP is an
*incredibly* nice laptop. Best quality screen I've seen, relatively
light, 5 hour battery life, etc.
> Hmmm....
>
> kt
Frank GOENNINGER <·············@nomail.org> wrote:
>
> Result: One piece of damn pretty hardware and OS X plus two options to
> run Windows. Plus running Linux using Parallels (also done over here for
> porting stuff to Debian and Suse).
>
a) What Linux distro? I have trouble getting Fedora to run (I don't get
much of a choice in this matter).
b) Does it share filesystems like it can with Windows?
c) Does cut and paste work?
At one time, one or all of these didn't work for me and I gave up trying
to run Linux w/ Parallels.
····@stablecross.com (Bob Felts) writes:
> a) What Linux distro? I have trouble getting Fedora to run (I don't get
> much of a choice in this matter).
I installed debian on Parallels. I didn't bother to try to make the
X server work, though (used the mac X server instead).
> b) Does it share filesystems like it can with Windows?
I just set up rsync between the file systems.
> c) Does cut and paste work?
See a)....
--
(espen)
Kenny <·········@gmail.com> writes:
> Patrick May wrote:
>> Kenny <·········@gmail.com> writes:
>>
>>>And where should I put my money in my next computer?
>>
>>
>> http://www.apple.com
>
> Got one, thx. Well, five, but only two are OS X and only one of those
> did not die because Apple makes crappy hardware these days. And no Mac
> runs the AllegroCL Windows IDE... or is Vine that good?
VMWare Fusion is that good.
>
> kt
On Thu, 30 Oct 2008 15:29:58 -0400, Kenny <·········@gmail.com> wrote:
>Is AMD any different than Intel, or shouls I save $150?
Yes. Intel currently has the edge in overall performance. AMD has
IMO a better multicore architecture, but they've not exploited it
fully yet. Historically, AMD has had better integer performance and
Intel has had better floating point performance, if that makes a
difference.
>I know Vista is stupid, but is it slow? I'll be getting a lot of RAM.
Vista is ok if you disable all the junk that runs by default.
>How cores? If it is just me doing Web2.0 programming, how many can I
>use? Should I spring for the quad core deal?
You're better off with 2 real machines and a network ... client/server
(even with VMs) on the same machine won't allow you to exercise all
the failure modes.
WRT quad vs dual, that's up to you. More cores means more chances for
buggy multithread code to screw up so a quad might help a bit with
development. It probably won't help much with compile times ... most
compilers are single threaded.
>What about 64-bit vs 32-bit? Any advantage to the former? I do not need
>more than 4g ram.
You can't access 4GB with 32-bit chips unless you play games. See
http://en.wikipedia.org/wiki/Physical_Address_Extension
Other than that it's a trade off. Code expands about 30% using 64-bit
pointers and execution tends to be a slower overall due to memory
bandwidth. OTOH, 64-bit memory management is much better and fixnums
are larger so there is less need for bignums. With floating point you
may not see much difference depending on what 32-bit chip you're
coming from.
George
P� Fri, 31 Oct 2008 22:22:33 +0100, skrev George Neuner
<········@comcast.net>:
>
> Other than that it's a trade off. Code expands about 30% using 64-bit
> pointers and execution tends to be a slower overall due to memory
> bandwidth. OTOH, 64-bit memory management is much better and fixnums
> are larger so there is less need for bignums. With floating point you
> may not see much difference depending on what 32-bit chip you're
> coming from.
>
This 32 vs. 64 is a bit of a fuzzy area. Most processors today have 64 bit
extensions (800 series and up), but run on a 32 bit operating system. But
the processor can adapt to to some of the operations in 64 bit internally.
This is probably why the difference is not as staggering as one might
expect. Also there is the SSEx/XMD and floating point issues. And of
course you have to consider internal cache size and supportive hardware
like the GPU and IDE/ATA/SATA drive controller etc. Thus metering
processor and program performance is becoming increasingly hard and
application specific.
--------------
John Thingstad
George Neuner wrote:
> WRT quad vs dual, that's up to you. More cores means more chances for
> buggy multithread code to screw up so a quad might help a bit with
> development. It probably won't help much with compile times ... most
> compilers are single threaded.
But programs are usually composed from several source files that can be
compiled in parallel. cf. make -j4
For the program at work, a quad-code divided the compilation times
almost in four.
--
__Pascal Bourguignon__
http://www.informatimago.com