Most programs do not use more than one processor. Photoshop does on some filters I think. Office runs the spellchecker in another thread, but usualy you do not gain much with 2 cpu:s.
It works (if I remember correct) that when you try to *acces* a byte that you do not own, you will get a pagefault. The program is linked to a special library that catch the signal and from the address calculates witch machine to get the data from. Data is then transfeard in chunks. I.e. you can not send data, only recive; data will be send when another node wants your data. You need barriers and other syncronisation primitives of course.
Yes, of course they use message passing, but implicit.
I.e. when you use virtual memory on a remote node the pagefault will triger a routine that do a RPC call, you still need 64 bit address space in order to access the memory (who owns a cluster with a total memory of less than 4gig???.
Often, when doing cluster work you map *all* the memory of the cluster into virtual memory. You can then read from other computers using normal pointers. However 32 bit address space is not enough.
You then have to use explicit message passing, like MPI.
Me too, but the case is that you can only use the G5 as a 32-bit processor in MacOS X.
sizeof(long) = 4 sizeof(void*) = 4
Both of which should be 64 bit.
long long is a gnu extension to c (and c++) that is always 64 bit _even_ on x86.
It works by using _two_ 32 bit register.
The G5 (like the power4) can be run in two modes, 32-bit or 64-bit. When running in 32-bit mode the register are 32-bit. So in what he says is that programs will be faster because the G5 is a faster processor (with built in sqrt funcktion), something that has _nothing_ to do with 64 bit computing.
I thought that MacOS was responsive, if that is a lie, you make a point in that some programs that are slowed down by its GUI will be faster. But then I do *only* hear about how good and *fast* the GUI is.
I thought Quicktime was *not* part of the OS, and would be updated anyway. But are you telling me that old OS X will not be able to see new film formats???
I would be suprised if vecLib is faster, such a lib should be handcrafted in assembler. Do you mean that the old assembler was made in a rush?
Do you expect that any of the programs below will run faster?
If I turn it around, do you think photoshop will be faster?
If the system is not responsive I see it as a bug and not as a new feature. But *why* do I allways hear this arguing when a new version is comming. Why do I *never* hear it before.
Yeah! And people tell me it should be called Linux instead for GNU/Linux cause only the kernel is the OS! But of course the GUI will be more snappy if the graphic sub-system is optimised.
I realy do not think responsivnes and speed is the same thing though, most people will think that you now can do your render jobs on a G3.
And I find it amusing that there exist no flaws in the OS except at the next release; no one has told me that MacOS is so unresponsive;-)
Programs generally do not get *that* much faster. It is the same for Apple as for Microsoft; do not expect a revolutionary faster OS. Remember that the time taken by the OS, in most cases is below 1% of the total program execution time.
Slashdot Mirror
You can get UltraSparcs for less than $3000
You could get Alphas for less than $3000
You could get Opteron for less than $3000
*before* you could get a 64-bit system from Apple.
I have a 10+ years old AlphaStation 255
Most programs do not use more than one processor.
Photoshop does on some filters I think. Office
runs the spellchecker in another thread, but
usualy you do not gain much with 2 cpu:s.
I think you are wrong.
What *could* give preformance increase is
to run the applications in 64 bit mode.
Then the Athlon64 would get 16 refisters, and that would be a real preformance increase.
I will try to find it...
It works (if I remember correct) that when you
try to *acces* a byte that you do not own,
you will get a pagefault. The program is linked
to a special library that catch the signal and
from the address calculates witch machine to
get the data from. Data is then transfeard in
chunks. I.e. you can not send data, only recive;
data will be send when another node wants your
data. You need barriers and other syncronisation
primitives of course.
I post again if I find a link...
Its not NUMA, it is fast interconnect network.
You pagefault and an RPC message will be sent.
Yes, of course they use message passing, but
implicit.
I.e. when you use virtual memory on a remote node
the pagefault will triger a routine that do a
RPC call, you still need 64 bit address space
in order to access the memory (who owns a cluster with a total memory of less than 4gig???.
To be fair it is no problem with the hardware,
it is the OS, that needs to be 64 bit.
That is, the cluster would probebly be better
using BSD, GNU/Linux or AIX.
Yes, MacOS X is 32 bit.
Often, when doing cluster work you map *all*
the memory of the cluster into virtual memory.
You can then read from other computers using normal
pointers. However 32 bit address space is not enough.
You then have to use explicit message passing,
like MPI.
Me too, but the case is that you can only use
the G5 as a 32-bit processor in MacOS X.
sizeof(long) = 4
sizeof(void*) = 4
Both of which should be 64 bit.
long long is a gnu extension to c (and c++)
that is always 64 bit _even_ on x86.
It works by using _two_ 32 bit register.
The G5 (like the power4) can be run in two modes,
32-bit or 64-bit. When running in 32-bit mode
the register are 32-bit. So in what he says
is that programs will be faster because the G5
is a faster processor (with built in sqrt
funcktion), something that has _nothing_ to do
with 64 bit computing.
Mozilla for web/mail
Open Office for office replacement
The Gimp for bitmap graphics
Sodipodi for vector graphics
Gaim for instante messaging
*ALL* of which exist for both win and unix.
I.e you do not have to learn different applications.
X11amp/mplayer for media on unix.
>> How could you write a program which accesses
>> files, with a GUI, without calling the OS?
You can make programs that do not call the
OS at all. I have said *nothing* else.
I have not said that gcc has not improved, I say
it will not run faster because of an OS upgrade.
Come on, rendering text has been real time
for 30 years.
If you actually do wait for text to appear, then
why do you pay more than $3000 for a computer.
You wait for computational work, if not, the
GUI is badly designed. And that is a fact.
I do *not* consider HTML engine from KDE to
be part of the OS, I do not wait when I send
mail (and neither does mac people).
I thought that MacOS was responsive, if that is
a lie, you make a point in that some programs that
are slowed down by its GUI will be faster.
But then I do *only* hear about how good and
*fast* the GUI is.
I thought Quicktime was *not* part of the OS, and
would be updated anyway. But are you telling me
that old OS X will not be able to see new film
formats???
I would be suprised if vecLib is faster, such a
lib should be handcrafted in assembler. Do you
mean that the old assembler was made in a rush?
Do you expect that any of the programs below will
run faster?
SPEC
GCC
Photoshop
3Dstudio
So, you expect dramatic inprovment in the following programs:
SPEC
GCC
PHOTOSHOP
3DSTUDIO
I do not.
>>*all* apps spend a not insignificant amount of
>>their time calling code that comes with the OS.
Thats *not* true.
In fact you do not have to call the OS at *all*.
>> I don't expect a revolutionary faster OS, I do
>> expect a much more responsive system.
Fair enough!
I do not consider the GUI to be part of the OS.
If I turn it around, do you think photoshop
will be faster?
If the system is not responsive I see it as a bug
and not as a new feature. But *why* do I allways
hear this arguing when a new version is comming.
Why do I *never* hear it before.
And exactly *how* does this make photoshop faster???
Yeah! And people tell me it should be called
;-)
Linux instead for GNU/Linux cause only the kernel
is the OS! But of course the GUI will be more
snappy if the graphic sub-system is optimised.
I realy do not think responsivnes and speed is the
same thing though, most people will think that
you now can do your render jobs on a G3.
And I find it amusing that there exist no flaws
in the OS except at the next release; no one
has told me that MacOS is so unresponsive
How is that possible? I thought MacOS X was rock
stable!!! How the hell can some one with an uptime
of years find that the new version is more stable?
I do not belive it.
Programs generally do not get *that* much faster.
It is the same for Apple as for Microsoft; do not
expect a revolutionary faster OS. Remember that
the time taken by the OS, in most cases is below
1% of the total program execution time.
It was previewed on an x86 computer...
Well that was kind of what I tried to say ;-)
Well I do not think RMS has that a big impact on
technology, more on society.
>More one reson (besides price) to not buy a dvd
>recorder now.
Besides price?
I think they begin to be _affordable_.