subject:"Re\: \[OT\] End\-user uses for x86\-64 \(was\: Why are still not at 64 bits\)"


On 2/17/07, Ben Scott <[EMAIL PROTECTED]> wrote:

On 2/17/07, Thomas Charron <[EMAIL PROTECTED]> wrote:
>>  People may giggle, but the PS2 Emotion processing chip is 128 bit.
  I believe it has 128-bit floating point/vector data processing
capabilities, but the integer registers are 64-bit, and the address
word is 32-bit.  Right?  If so, in the context of most of the
discussion in this thread, it would be classified as a 32-bit
architecture.


The integer registers are 128 bit on the Emotion chip.  Address words
is a tough one, as it works in a mixed mode.  As maddog pointed out,
the memory addresses are 32 bit.  But everything INSIDE the chip is
128 bit.  To make it even more confusing, the other processors in the
system are 64 bit.  But the use of a 32 bit memory register is simply
a space saving detail.  The processors in the PS2 could very easily
handle 64 bit memory addressing, but why bother and waste the space if
you know it will NEVER need to address memory in those quantities.


  For that matter, doesn't MMX or one of it's follow-on acronyms add
special-purpose 128-bit registers to x86?


I'm not aware of any 128 bit registers, maybee in SSE2/3?  There where
some 64 bit registers, however.


  I'm not trying to imply the pee cee is anything like the PS2; I'm
just pointing out that there are a lot of things about a processor
that get measured in "bits".  Address word size.  Instruction word
size.  Integer registers.  Floating-point registers.  Memory bus(es).
I/O bus(es).  Number of pieces it breaks into when you hit it with a
hammer.  Etc.


 Exactly my point.  There are 32 bit processors that can handle 64
bit.  Most of the focus here seemed to be pretty much purely on the
ability for a 64 bit processor the address additional memory, and
that's a side effect, really.  To sum it up, a 64 bit processor can do
more advanced math sheet faster, which is in turn what makes it able
to do the mom and pop stuff faster.  Generally speaking, almost
anything can run faster, for the most part, on a 64 bit processor, but
they can't do anything that CAN'T be done (granted, with a sometimes
significant performance hit) with a 32 bit processor.

--
-- Thomas
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

On 2/17/07, Jon 'maddog' Hall <[EMAIL PROTECTED]> wrote:

On Sat, 2007-02-17 at 21:08 -0500, Thomas Charron wrote:
> On 2/17/07, Jerry Feldman <[EMAIL PROTECTED]> wrote:
> > On Sat, 17 Feb 2007 17:01:19 -0500
> > "Jon 'maddog' Hall" <[EMAIL PROTECTED]> wrote:
>   People may giggle, but the PS2 Emotion processing chip is 128 bit.
> http://en.wikipedia.org/wiki/Emotion_Engine
Yes, but read it again.  It does not say anything about the address
space being 128 bits.  Just the registers, datapaths, etc.
I can understand having a 128 bit data register, to manipulate those
IPv6 addresses.  I can understand having a 128 bit datapath.  But 128
bit memory address.

Correct.  If we're JUST talking 128 bit memory addresses, then chances
are, it only does so to optimize the datapaths, but not in a raw 128
bit pointer.

But 128 bit / 64 bit is more then just addressable memory.

I guess that's what one of my initial points was meant to be.  If
anything, a pointer being an int and an int being able to be used to
'address more' is a fringe benefit.

--
-- Thomas
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)


On 2/17/07, Ben Scott <[EMAIL PROTECTED]> wrote:

If so, in the context of most of the discussion in this thread ...


s/thread/tangled ball of string/

-- Ben
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-17 Thread Jon 'maddog' Hall

On Sat, 2007-02-17 at 21:08 -0500, Thomas Charron wrote:
> On 2/17/07, Jerry Feldman <[EMAIL PROTECTED]> wrote:
> > On Sat, 17 Feb 2007 17:01:19 -0500
> > "Jon 'maddog' Hall" <[EMAIL PROTECTED]> wrote:
> > > I do not think a "128 bit address space" computer will ever exist, at
> > > least not in the silicon technologies that we are talking about.
> > Probably not for a while, but I'm 100% certain, there will be a 128-bit
> > address space computer somewhere down the line, but at the present
> > time, 64-bit address space is more than sufficient for most of the
> > biggest computers. It will be the governments that will want the
> > 128-bit address space. On the other hand, I seriously doubt desktop
> > computers will ever need anything near 64-bit addresses.
> 
>   People may giggle, but the PS2 Emotion processing chip is 128 bit.
> 
> http://en.wikipedia.org/wiki/Emotion_Engine
> 
>   It all comes down to funky fast math.
> 
Yes, but read it again.  It does not say anything about the address
space being 128 bits.  Just the registers, datapaths, etc.

I can understand having a 128 bit data register, to manipulate those
IPv6 addresses.  I can understand having a 128 bit datapath.  But 128
bit memory address.

md

___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)


On 2/17/07, Thomas Charron <[EMAIL PROTECTED]> wrote:

 People may giggle, but the PS2 Emotion processing chip is 128 bit.


 I believe it has 128-bit floating point/vector data processing
capabilities, but the integer registers are 64-bit, and the address
word is 32-bit.  Right?  If so, in the context of most of the
discussion in this thread, it would be classified as a 32-bit
architecture.

 For that matter, doesn't MMX or one of it's follow-on acronyms add
special-purpose 128-bit registers to x86?

 I'm not trying to imply the pee cee is anything like the PS2; I'm
just pointing out that there are a lot of things about a processor
that get measured in "bits".  Address word size.  Instruction word
size.  Integer registers.  Floating-point registers.  Memory bus(es).
I/O bus(es).  Number of pieces it breaks into when you hit it with a
hammer.  Etc.

-- Ben
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

On 2/17/07, Ben Scott <[EMAIL PROTECTED]> wrote:

On 2/17/07, Thomas Charron <[EMAIL PROTECTED]> wrote:
>  .  The only limitation to a 16 bit processor is being limited to
> 64 KB of data per page at a time.
  Right, just as the beggar's only limitation is that he has no money.

...  Not quite sure how that applies.  I have a feeling a beggar can
get money in other ways then having him page memory.

 Maybee if he pages his wallet?  :-D

>> But it's [16-bit limitation workarounds] so slow, cumbersome,
>> and error-prone as to be a significant obstacle.
>   Which is why different kinds of embedded systems will use multiple
> smaller scale processors.
  Fine if you're building something application specific.  Not so much
if it's a general-purpose computer.  It the ASIC route was magic
business pixie dust, we'd all be using Amiga's.

 Hey now, they where damned purdy machines.

  Moving beyond the UI, there's plenty that benefits from being able
to count to higher than 65535 in a CPU register, or being able to work
on more than 64 KB of stuff at a time.  For example, I suspect
implementing real-time MPEG2 decoding in the Intel 8501 would be
difficult.  Sure, lots of bitty boxes use bitty CPUs, but they have
ASICs with wide pipes instead.

They add a little side circuit that can do it cheaply.

http://www.keil.com/dd/chipinfo.asp?did=3323&bhcp=1
http://www.cast-inc.com/cores/mpeg4-e/index.shtml

People do it all the time, cheaply.

  I'm not trying to say small chips are useless or dead or passe, just
that big chips are can enable more than "bigger and faster".

 Yes.  Price.

>>   Sure, the i386 brought a number of other advantages to the table,
>> chief among them a real MMU, but the address word size mattered, too,
>> I think.
> Only in the fact that it was faster ...
  Um.  Perhaps you never worked with "Windows/286".

 It was a moot point by the time the 386 came out.  I also worked
with Minix.  In all fairness, I believe what caused the 386 to trounce
the 286 was the advent of VGA graphics, and people upgrading their
machines for it.

>   And a good reason to completely disregard them because 'Gnome Sux,
> KDE Rulez!
  Do you have a relevant point, here?

 That the Linux community pushing 'You now have choice' breaks down
when it comes to the general public.  And that perhaps we can actually
learn from WHY people prefer Windows in general.

> We have yet to see how that's going to play out, however.  But
> generally, anything that doesn't work like 'Doze does, doesn't fly
> very well.
  They said that for years and years about 1-2-3 and WP and NetWare.
  The "network effect", AKA "installed base", is a very real thing.
It takes something significant to overcome it.  Which was what I was
attempting to look at: Is there anything significant on the horizon?
But we're all looking backwards or down instead of ahead for some
reason.

 And as much as I miss my Netware servers, they're mostly all gone
now.  But this is kinda a moot point I guess.

>>> And there is quite literally NOTHING you cannot do in 32 bit that
>>> you can in 64.
>> Addressing more than 4 GB of RAM without memory
>> windowing/segmentation comes to mind
>   Those are possibilities that address the issue.
  Such as?  Serious question; I'm at most a very casual student of
micro-architectures, so I don't know.  I enjoy learning, though.  So
educate me.  :)

 The easiest to explain is either some form of windowing or segmentation.

 Some decent reading, which I'll probably get flamed for referring to
can be read at

http://technet2.microsoft.com/WindowsServer/en/library/efc41320-713f-4004-bc81-fc8552651033.mspx?mfr=true

 The laymans description ends up with paging, however.

>   Hehehe.  And Windowz is also sometimes credited for the success of
> the Pentium.  Does that define it as a killer app?
  Perhaps.  Indeed, Microsoft bloat has powered quite a bit of
hardware sales.  Have you seen the recommended system configurations
for Vista?  Supercomputers modeling the Earth's climate need less
power.

 Hey now, that gave me an excuse for my new laptop, HUSH UP before my
GF hears you!@  :-D

>   Or where they simply the defacto standard apps people used for
> productivity, and had little to do with IBM-PC?
  That's the whole freaking point!  :-)  A "killer app" isn't
necessarily intrinsic to the thing it boosts; it's just responsible
for the widespread adoption.

 Hrm.  You may have a good point.  Maybe it was the combination of
both applications and the name 'IBM'.  Or maybe it was the cheaper
clones which pushed and pushed the platform further and further.

 I often wondered what would have happened if Apple hadn't locked down the Mac.

> This is fun, but I suspect some list members may be
> telling us to STFU soon enough.
  I note that the silence is deafening.  Perhaps we should take this
off-list?  :-)

 I'm game, but it would be VERY non-typical for me NOT to annoy
people..  Let's go offlist unless someone els

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)


On 2/17/07, Jerry Feldman <[EMAIL PROTECTED]> wrote:

On Sat, 17 Feb 2007 17:01:19 -0500
"Jon 'maddog' Hall" <[EMAIL PROTECTED]> wrote:
> I do not think a "128 bit address space" computer will ever exist, at
> least not in the silicon technologies that we are talking about.
Probably not for a while, but I'm 100% certain, there will be a 128-bit
address space computer somewhere down the line, but at the present
time, 64-bit address space is more than sufficient for most of the
biggest computers. It will be the governments that will want the
128-bit address space. On the other hand, I seriously doubt desktop
computers will ever need anything near 64-bit addresses.


 People may giggle, but the PS2 Emotion processing chip is 128 bit.

http://en.wikipedia.org/wiki/Emotion_Engine

 It all comes down to funky fast math.

--
-- Thomas
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)


On 2/17/07, Jon 'maddog' Hall <[EMAIL PROTECTED]> wrote:

>   And there is quite literally NOTHING you cannot do in 32 bit that
> you can in 64.
Yes there is.  You can mmap a single 5 GB virtual address space.
Now if you had said that there are no problems that you can not solve,
given enough time and processing power, with a 32 bit machine than you
could with a 64-bit machine, I would agree with you, and so would Mr.
Turing.


 *sigh*

 True enough..

 But then one must look at it another way.  If you didn't have to,
why would you?  :-)

--
-- Thomas
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-17 Thread Jerry Feldman

On Sat, 17 Feb 2007 17:01:19 -0500
"Jon 'maddog' Hall" <[EMAIL PROTECTED]> wrote:

> I do not think a "128 bit address space" computer will ever exist, at
> least not in the silicon technologies that we are talking about.
Probably not for a while, but I'm 100% certain, there will be a 128-bit
address space computer somewhere down the line, but at the present
time, 64-bit address space is more than sufficient for most of the
biggest computers. It will be the governments that will want the
128-bit address space. On the other hand, I seriously doubt desktop
computers will ever need anything near 64-bit addresses. 

-- 
Jerry Feldman <[EMAIL PROTECTED]>
Boston Linux and Unix user group
http://www.blu.org PGP key id:C5061EA9
PGP Key fingerprint:053C 73EC 3AC1 5C44 3E14 9245 FB00 3ED5 C506 1EA9


signature.asc
Description: PGP signature
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-17 Thread Jon 'maddog' Hall


>   And there is quite literally NOTHING you cannot do in 32 bit that
> you can in 64.
> 
Yes there is.  You can mmap a single 5 GB virtual address space.

Now if you had said that there are no problems that you can not solve,
given enough time and processing power, with a 32 bit machine than you
could with a 64-bit machine, I would agree with you, and so would Mr.
Turing.

md

___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)


On 2/17/07, Thomas Charron <[EMAIL PROTECTED]> wrote:

 .  The only limitation to a 16 bit processor is being limited to
64 KB of data per page at a time.


 Right, just as the beggar's only limitation is that he has no money.


But it's [16-bit limitation workarounds] so slow, cumbersome,
and error-prone as to be a significant obstacle.


  Which is why different kinds of embedded systems will use multiple
smaller scale processors.


 Fine if you're building something application specific.  Not so much
if it's a general-purpose computer.  It the ASIC route was magic
business pixie dust, we'd all be using Amiga's.


A UI doesn't need more then 16 bits for most applications.


 Perhaps not for the UI in the sense of buttons and widgets.  True
color graphics kind of suck in the 16-bit space, though.  (Sure, it
can be done.  Putting a man on the moon using a really big pile of
dynamite could probably be done.  Doesn't mean it's a good way to do
it.)

 Moving beyond the UI, there's plenty that benefits from being able
to count to higher than 65535 in a CPU register, or being able to work
on more than 64 KB of stuff at a time.  For example, I suspect
implementing real-time MPEG2 decoding in the Intel 8501 would be
difficult.  Sure, lots of bitty boxes use bitty CPUs, but they have
ASICs with wide pipes instead.

 I'm not trying to say small chips are useless or dead or passe, just
that big chips are can enable more than "bigger and faster".


  Sure, the i386 brought a number of other advantages to the table,
chief among them a real MMU, but the address word size mattered, too,
I think.


Only in the fact that it was faster ...


 Um.  Perhaps you never worked with "Windows/286".


  And a good reason to completely disregard them because 'Gnome Sux,
KDE Rulez!


 Do you have a relevant point, here?


We have yet to see how that's going to play out, however.  But
generally, anything that doesn't work like 'Doze does, doesn't fly
very well.


 They said that for years and years about 1-2-3 and WP and NetWare.

 The "network effect", AKA "installed base", is a very real thing.
It takes something significant to overcome it.  Which was what I was
attempting to look at: Is there anything significant on the horizon?
But we're all looking backwards or down instead of ahead for some
reason.


And there is quite literally NOTHING you cannot do in 32 bit that
you can in 64.


Addressing more than 4 GB of RAM without memory
windowing/segmentation comes to mind


  Those are possibilities that address the issue.


 Such as?  Serious question; I'm at most a very casual student of
micro-architectures, so I don't know.  I enjoy learning, though.  So
educate me.  :)


  Hehehe.  And Windowz is also sometimes credited for the success of
the Pentium.  Does that define it as a killer app?


 Perhaps.  Indeed, Microsoft bloat has powered quite a bit of
hardware sales.  Have you seen the recommended system configurations
for Vista?  Supercomputers modeling the Earth's climate need less
power.


 With the IBM-PC, people point to Lotus 1-2-3 and Word Perfect.
There was a time when, if you saw an IBM-PC-compatible in an office,
you were nearly certain to see one of those two programs up on the
screen.


  Or where they simply the defacto standard apps people used for
productivity, and had little to do with IBM-PC?


 That's the whole freaking point!  :-)  A "killer app" isn't
necessarily intrinsic to the thing it boosts; it's just responsible
for the widespread adoption.


If they where available for the Apple ][, would we all be using Mac's?


 Quite likely, yes.  And bitching about how Apple's monopoly sucks.


This is fun, but I suspect some list members may be
telling us to STFU soon enough.


 I note that the silence is deafening.  Perhaps we should take this
off-list?  :-)

-- Ben
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-17 Thread Jon 'maddog' Hall


> Yes, there are uses for 64 bit address space, just as a 128 bit address 
> space would enable use to tackle unthinkable problems.

I do not think a "128 bit address space" computer will ever exist, at
least not in the silicon technologies that we are talking about.

Just to take advantage of a 64-bit address space you would probably have
to have all of the disks ever made put together into one system.

For a processor to simply touch each byte of a 64-bit address space,
running at 4 GHz would take about 4 billion seconds (or about 126
years).

I think I read someplace that 128 bits would allow you to address every
Proton and neutron in the known universe, but I doubt that (a little).

While 8, 16 and even 32 bits caused most programmers some pain from time
to time, I think 64 bits has enough address space to keep even Dr. Knuth
happy for a while.

md

___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

On 2/17/07, Ben Scott <[EMAIL PROTECTED]> wrote:

On 2/17/07, Thomas Charron <[EMAIL PROTECTED]> wrote:
>> machine.  They don't understand why, but they know they can play
>> digital music while writing a term paper on their new Dell, while
>> their old Apple ][ or IBM-PC Model 5150 couldn't handle that.
>   That has nothing to do with sized bits I'm afraid.
  I'm inclined to disagree for the general case.  As long as it's a
general purpose computer (and not a bunch of small
application-specific computers), the limitations imposed by only being
able to deal with 64 KB of data at a time are pretty real.

.  The only limitation to a 16 bit processor is being limited to
64 KB of data per page at a time.

It may in
theory be possible to do things in that small a space, by doing enough
task switching or memory windowing or breaking the data down into tiny
chunks or  multi-register arithmetic.  But it's so slow, cumbersome,
and error-prone as to be a significant obstacle.

 Which is why different kinds of embedded systems will use multiple
smaller scale processors.  A UI doesn't need more then 16 bits for
most applications.

  Sure, the i386 brought a number of other advantages to the table,
chief among them a real MMU, but the address word size mattered, too,
I think.

 Only in the fact that it was faster, and 'the next best thing' in an
environment where replacing a system tended to happen in a 2-3 year
cycle.

>   And until there's someone forcing their choices upon them, the
> general public will be utterly confused by the choices.
  The general public tends to be utterly confused, period.

 True 'nuff.  :-)

> They don't give a crap about Bonobo vs KParts, they just want to know
> why the app gstreamer doesn't seem to be able to sit in the KDE
> taskbar like Joan down the street does on HER desktop.
  Well, in theory, standards can address the issue of multiple
choices.  There actually is a standard for that task bar tray icon
stuff, which demonstrates that standards can work in practice, too.
Fortunately, the FLOSS community has a lot of good reasons to like
standards.

 And a good reason to completely disregard them because 'Gnome Sux,
KDE Rulez!  Nonono!  KDE can actually make a decent print icon, fug
Gnome!  Nono, wait!  Let's put the files in /etc/init.d!  NONONONO!
Someplace else!  Ok, fine, at least we can have apache running as
apache..  NONONO!  RUN it as httpd!

  People in the 'doze world get pissed when X doesn't work with Y,
too.  Sometimes, they get even more pissed when they find out that it
doesn't work simply because Microsoft or Apple wanted to improve their
bottom line.

 We have yet to see how that's going to play out, however.  But
generally, anything that doesn't work like 'Doze does, doesn't fly
very well.  There ARE some exceptions, but they are few.

> And there is quite literally NOTHING you cannot do in 32 bit that
> you can in 64.
  Addressing more than 4 GB of RAM without memory
windowing/segmentation comes to mind

 Those are possibilities that address the issue.  64 bit is another.

>>   That's a so-called "killer app".  Entire industries have risen and
>> fallen on such things in the past.
> Name one killer app that killed the 386, or 486.
  It obviously hasn't happened, yet, and may never.  Duh.
  VisiCalc, the first mass-market spreadsheet program, is widely
credited for the commercial success of the Apple ][.

 Hehehe.  And Windowz is also sometimes credited for the success of
the Pentium.  Does that define it as a killer app?

  With the IBM-PC, people point to Lotus 1-2-3 and Word Perfect.
There was a time when, if you saw an IBM-PC-compatible in an office,
you were nearly certain to see one of those two programs up on the
screen.

 Or where they simply the defacto standard apps people used for
productivity, and had little to do with IBM-PC?  If they where
available for the Apple ][, would we all be using Mac's?

  It looks like networked digital music may totally reshape the music
"industry" (cartel).

 Naw.  People not wanting to pay them if they don't have to reshapes
them.  :-D  Too bad it's not working quite that way at an OS level.

  Shall I go on?  :)

 By all means!  This is fun, but I suspect some list members may be
telling us to STFU soon enough.

--
-- Thomas
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)


On 2/17/07, Thomas Charron <[EMAIL PROTECTED]> wrote:

machine.  They don't understand why, but they know they can play
digital music while writing a term paper on their new Dell, while
their old Apple ][ or IBM-PC Model 5150 couldn't handle that.


  That has nothing to do with sized bits I'm afraid.


 I'm inclined to disagree for the general case.  As long as it's a
general purpose computer (and not a bunch of small
application-specific computers), the limitations imposed by only being
able to deal with 64 KB of data at a time are pretty real.  It may in
theory be possible to do things in that small a space, by doing enough
task switching or memory windowing or breaking the data down into tiny
chunks or  multi-register arithmetic.  But it's so slow, cumbersome,
and error-prone as to be a significant obstacle.

 Sure, the i386 brought a number of other advantages to the table,
chief among them a real MMU, but the address word size mattered, too,
I think.


  And until there's someone forcing their choices upon them, the
general public will be utterly confused by the choices.


 The general public tends to be utterly confused, period.


They don't give a crap about Bonobo vs KParts, they just want to know
why the app gstreamer doesn't seem to be able to sit in the KDE
taskbar like Joan down the street does on HER desktop.


 Well, in theory, standards can address the issue of multiple
choices.  There actually is a standard for that task bar tray icon
stuff, which demonstrates that standards can work in practice, too.
Fortunately, the FLOSS community has a lot of good reasons to like
standards.

 People in the 'doze world get pissed when X doesn't work with Y,
too.  Sometimes, they get even more pissed when they find out that it
doesn't work simply because Microsoft or Apple wanted to improve their
bottom line.


And there is quite literally NOTHING you cannot do in 32 bit that
you can in 64.


 Addressing more than 4 GB of RAM without memory
windowing/segmentation comes to mind


  That's a so-called "killer app".  Entire industries have risen and
fallen on such things in the past.


Name one killer app that killed the 386, or 486.


 It obviously hasn't happened, yet, and may never.  Duh.

 VisiCalc, the first mass-market spreadsheet program, is widely
credited for the commercial success of the Apple ][.

 With the IBM-PC, people point to Lotus 1-2-3 and Word Perfect.
There was a time when, if you saw an IBM-PC-compatible in an office,
you were nearly certain to see one of those two programs up on the
screen.

 It looks like networked digital music may totally reshape the music
"industry" (cartel).

 Desktop publishing and graphics put the Mac on the map, and is
largely responsible for keeping it there.

 It looks like VoIP might well reshape the voice telecom industry.

 Amesbury, MA, where I work, used to be a bustling city, because they
manufactured carriages (the horse-drawn kind).

 The Internet put the metered-service, "walled garden" online service
industry into the history books (Compuserve, MCI Mail, GEnie, Prodigy,
etc.).

 Shall I go on?  :)

-- Ben
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

On 2/17/07, Ben Scott <[EMAIL PROTECTED]> wrote:

machine. They don't understand why, but they know they can play
digital music while writing a term paper on their new Dell, while
their old Apple ][ or IBM-PC Model 5150 couldn't handle that.

That has nothing to do with sized bits I'm afraid.

> I'm not going to argue that 64 bits won't make a difference, just that
> typical home and office users won't notice until it enables something
> that no one has done before, or not done well.

Right. Exactly. I'm wondering what those things might be. What
will x86-64 let

people-doing-ordinary-tasks-like-reading-email-browsing-the-web-writing-letters-looking-at-pictures-and-calculating-their-taxes

do that the limitations of their x86-32 computer prevented them from
doing?

Nothing. But processors can be made to do things faster when they
can deal with more data at a time (gross exageration).

However, most of them are still running a 32-bit OS. Be it Windows
or Linux or Mac OS, most of the installations are still 32-bit, even
if the OS has a 64-bit variant available. In the 'doze world, this is
largely because of the support and compatibility nightmares described
previously. 64-bit Windows causes lots and lots of problems, and
there's currently very little benefit to be had by
people-doing-ordinary-tasks-like-reading-email-browsing-the-web-writing-letters-looking-at-pictures-and-calculating-their-taxes.

So, from that standpoint, the 64-bit potential for the
people-doing-ordinary-tasks-like-reading-email-browsing-the-web-writing-letters-looking-at-pictures-and-calculating-their-taxes
market is still untapped.

It was tapped, but no one cared. I was half hoping they'd simply
release Vista and say it's 64 bit, period. Otherwise, no one really
cares.

Now, let's just say, hypothetically, that something materializes in
the Linux world which needs a 64-bit system to work, and is also
compelling to
people-doing-ordinary-tasks-like-reading-email-browsing-the-web-writing-letters-looking-at-pictures-and-calculating-their-taxes.
Maybe it's a really cool fully-immersive VR world (which will, of
course, immediately be used for sexual purposes). If it's only
available for Linux, then suddenly,
people-doing-ordinary-tasks-like-reading-email-browsing-the-web-writing-letters-looking-at-pictures-and-calculating-their-taxes
will have a *compelling reason* to check Linux out. And maybe, just
maybe, they'll discover all the other benefits of FLOSS while they're
at it.

And until there's someone forcing their choices upon them, the
general public will be utterly confused by the choices. They won't
understand why they can't run A with B when Joe down the street does.
They don't give a crap about Bonobo vs KParts, they just want to know
why the app gstreamer doesn't seem to be able to sit in the KDE
taskbar like Joan down the street does on HER desktop.

And there is quite literally NOTHING you cannot do in 32 bit that
you can in 64.

That's a so-called "killer app". Entire industries have risen and
fallen on such things in the past. As long as Windows sucks at x86-64
and Linux doesn't suck at x86-64, this question will remain very
interesting to me.
Am I the only one (aside from, perhaps, ESR) who thinks this way?

Name one killer app that killed the 386, or 486.

--
-- Thomas
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

On 2/17/07, Jim Kuzdrall <[EMAIL PROTECTED]> wrote:

>   That's not really true.  16-bit machines are *very* limited.  There
> is not a whole lot you can do in 64 kilobytes of RAM (all you can
> directly address with a 16-bit address word).
Not quite so.  As a programmer of embedded systems, I would point
out that sales of microprocessors with address spaces of 16-bits (or
less) exceed those of the larger machines by orders of magnitude.

 Definatly true.  There is a whole lot you can do in a limited
bitspace.  And there are ways to get around the 16-bit address words
fairly easily.

The reference to the automobile engine cylinder count in another
post is a good one.  Four, six, and eight cylinder engines each
continue to be commonly used.  The Lotus has a 12 or 16 cylinder
engine, I know of no other.
By analogy, does the 64-bit machine run on 8-cylinders or
12-cylinders?

 Hehe, I'd say 16-cylinder.  But it doesn't really compare to
processors at all.  Generally, engines with more cylinders produce
less power per cylinder, but it balances out.  :-P

--
-- Thomas
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

On 2/17/07, Jason Stephenson <[EMAIL PROTECTED]> wrote:

Typical end users as defined before don't really care about the
differences. As long as they can do more or less what they want to do
with the computer, they won't really notice the difference.

I think you and I actually agree. I'm not saying
people-doing-ordinary-tasks-like-reading-email-browsing-the-web-writing-letters-looking-at-pictures-and-calculating-their-taxes
actually care about memory models or address word size or any of that
crap. They generally don't. Ignoramuses might get into penis-length
contests because 64 is bigger than 32, but that's about it.

However,
people-doing-ordinary-tasks-like-reading-email-browsing-the-web-writing-letters-looking-at-pictures-and-calculating-their-taxes
do care about what they can do with their pee cee. There are things
they can do on a 32-bit machine which they cannot do on a 16-bit
machine. They don't understand why, but they know they can play
digital music while writing a term paper on their new Dell, while
their old Apple ][ or IBM-PC Model 5150 couldn't handle that.

I'm not going to argue that 64 bits won't make a difference, just that
typical home and office users won't notice until it enables something
that no one has done before, or not done well.

Right. Exactly. I'm wondering what those things might be. What
will x86-64 let
people-doing-ordinary-tasks-like-reading-email-browsing-the-web-writing-letters-looking-at-pictures-and-calculating-their-taxes
do that the limitations of their x86-32 computer prevented them from
doing?

Eventually, you won't be able to buy mainstream hardware with anything
less than 64 bit CPUs, and from looking at the latest Dell PC Catalogs
that I get, that time appears to be now or very soon.

Indeed, that is very much happening. It's practically impossible to
buy an IBM-PC-compatible these days that is not x86-64 capable. Some
laptops are about the only exceptions. So the hardware is there.

Am I the only one (aside from, perhaps, ESR) who thinks this way?

-- Ben
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-17 Thread Jim Kuzdrall

On Saturday 17 February 2007 12:10, Ben Scott wrote:
> On 2/17/07, Jason Stephenson <[EMAIL PROTECTED]> wrote:
> > If end users are defined as home users and office users, then 64
> > bits will never matter to them, just like 32 bits doesn't matter to
> > them today.
>
>   That's not really true.  16-bit machines are *very* limited.  There
> is not a whole lot you can do in 64 kilobytes of RAM (all you can
> directly address with a 16-bit address word).

Not quite so.  As a programmer of embedded systems, I would point 
out that sales of microprocessors with address spaces of 16-bits (or 
less) exceed those of the larger machines by orders of magnitude.

There are many things done by 16-bit address space computers that 
can't be done by the larger ones.  Like make cheap, low powered, small 
consumer products.

Everybody has a place in the chorus.

The reference to the automobile engine cylinder count in another 
post is a good one.  Four, six, and eight cylinder engines each 
continue to be commonly used.  The Lotus has a 12 or 16 cylinder 
engine, I know of no other.

By analogy, does the 64-bit machine run on 8-cylinders or 
12-cylinders?

Jim Kuzdrall
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-17 Thread Michael ODonnell



Dang!  teledildonics.com is already registered:

  Registrant:
  Riggs, Roy
   1508 BOONE CT
   MURFREESBORO, TN 37130-5032
   US

   Domain Name: TELEDILDONICS.COM

   Administrative Contact:
  Riggs, Roy[EMAIL PROTECTED]
 
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-17 Thread Jason Stephenson

Ben's point about the advantages of more memory and the comparison to 
the 16-bit to 32-bit transition is well taken, but I don't think that 
changes my main point:


Typical end users as defined before don't really care about the 
differences. As long as they can do more or less what they want to do 
with the computer, they won't really notice the difference.


Sure, their 64-bit PC will be faster than their old 32-bit PC, but isn't 
that what they expect? It's the same thing going from a 4 cylinder to an 
8 cylinder automobile engine. You generally expect better performance.


They don't really care how that improved performance comes about.

I'm not going to argue that 64 bits won't make a difference, just that 
typical home and office users won't notice until it enables something 
that no one has done before, or not done well. If you can only run the 
ultrarealistic, VR porn with teledildonics module on the latest 64 bit 
tech, then people will run out and buy 64 bit specifically to get 64 
bit. However, before there is a compelling reason for them to get 64 
bit, they won't care as long as what they get this year is better than 
what they had last year.


Eventually, you won't be able to buy mainstream hardware with anything 
less than 64 bit CPUs, and from looking at the latest Dell PC Catalogs 
that I get, that time appears to be now or very soon.

___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-17 Thread Jerry Feldman

On Sat, 17 Feb 2007 12:10:59 -0500
"Ben Scott" <[EMAIL PROTECTED]> wrote:

>   That's not really true.  16-bit machines are *very* limited.  There
> is not a whole lot you can do in 64 kilobytes of RAM (all you can
> directly address with a 16-bit address word).  Anything running on an
> 8086 (i.e., MS-DOS and all its software) has to play all sorts of
> games just to address the full megabyte the IBM-PC architecture allows
> for in Real Mode.  
There is always the next plateau. The early pcs (generic), Apple II,
Commodore 64, Trash 80, and the first IBM PCs were all 8-bit (although
the PC had an 8-bit version of a 16-bit chip). Then we got the 16-bit
chips (such as the Atari ST, Commodore Amiga) and the PC clones that
were truly 16 bit, and the Mac. There was a relatively big transition
from 16-bits to 32-bits. And, still there remained a lot of 16-bit
Windows code. But, back then, home computers were not so ubiquitous as
they are today, and the transition can be somewhat painful. I remember
that we had a number of partners on the Alpha that refused to spend the
effort to port their applications to 64-bit, and we came up with a way
they could build a 32-bit application without having to supply a full
set of 32-bit libraries. In the x86-64 bit world, you have a chip that
fully supports the x86 32-bit environment and instruction set. You can
run a 32-bit OS and it does not any better. You can have a 64-bit OS,
and with some support, 32-bit applications built on 32-bit systems will
run natively. But, you need 32-bit and 64-bit libraries. 
-- 
Jerry Feldman <[EMAIL PROTECTED]>
Boston Linux and Unix user group
http://www.blu.org PGP key id:C5061EA9
PGP Key fingerprint:053C 73EC 3AC1 5C44 3E14 9245 FB00 3ED5 C506 1EA9


signature.asc
Description: PGP signature
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-17 Thread Jerry Feldman

On Sat, 17 Feb 2007 11:00:29 -0500
Jason Stephenson <[EMAIL PROTECTED]> wrote:

> Sixty-four bitness will never matter to end users, but it will become 
> ubiquitous over time, and sooner or later nearly everyone will have 
> computers and devices with 64 bit CPUs and operating systems and the 
> vast majority of people won't really notice.
It certainly will become ubiquitous because the last time I checked
both AMD and Intel's roadmap, they will only be making 64-bit chips.
(Not sure about the embedded market). 

For the routine home user, 64-bits should make their computers perform
better because of the linear address space and graphics support. But,
for the average home user, the mount of memory is more important than
the CPU speed or bitedness. Last year I bought a 64-bit laptop because
I do a lot of 64-bit stuff, but other than that, for my classes at
Northeaster, a 32-bit system would have been just fine. 
-- 
Jerry Feldman <[EMAIL PROTECTED]>
Boston Linux and Unix user group
http://www.blu.org PGP key id:C5061EA9
PGP Key fingerprint:053C 73EC 3AC1 5C44 3E14 9245 FB00 3ED5 C506 1EA9


signature.asc
Description: PGP signature
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

On 2/17/07, Jason Stephenson <[EMAIL PROTECTED]> wrote:

If end users are defined as home users and office users, then 64 bits
will never matter to them, just like 32 bits doesn't matter to them
today.

That's not really true. 16-bit machines are *very* limited. There
is not a whole lot you can do in 64 kilobytes of RAM (all you can
directly address with a 16-bit address word). Anything running on an
8086 (i.e., MS-DOS and all its software) has to play all sorts of
games just to address the full megabyte the IBM-PC architecture allows
for in Real Mode. And that's really just a way to enable a
more-than-16-bit-address-space without actually having to spend more
money on hardware; but it shifts a disproportionate burden to
software. If you've ever had to program in a windowed/segmented
memory model, you'll know what I mean. It really, really sucks.

The 32-bit flat memory model (as enabled on the IBM-PC platform by
the i386) yields a lot of real benefits to
people-doing-ordinary-tasks-like-reading-email-browsing-the-web-writing-letters-looking-at-pictures-and-calculating-their-taxes.
They can do things like browse a multimedia web page, or manage their
collection of family photos, or do real-time WYSIWYG page layout of
the church newsletter, all while listening to their collection of
pirated digital music. It's extremely difficult to do that in a
single 64 KB memory space.

On the other hand, I'm pressed to conceive of what practical benefit
people-doing-ordinary-tasks-like-reading-email-browsing-the-web-writing-letters-looking-at-pictures-and-calculating-their-taxes
will get from x86-64. Two gigabytes[1] is a *lot* of address space,
even for Microsoft's bloated code. There's not much I can think of
that benefits from more than that. Digital video editing is a good
one. Virtual reality (which includes most games) is likely to be
another. There doesn't seem to be much else of interest to
people-doing-ordinary-tasks-like-reading-email-browsing-the-web-writing-letters-looking-at-pictures-and-calculating-their-taxes.

Footnotes
-
[1] Two gigabytes is the correct figure in the context of Microsoft's
bloated code. The NT kernel splits the 4 GB address space between 2
GB for userland programs and 2 GB for kernel purposes, so userland
programs are effectively limited to 2 GB of address space.[2][3] This
has been a real problem in some cases.[4]
[2] There is a way to change it to a 3GB/1GB split, but that
introduces a lot of other issues, and is generally only of use to
people running a single memory-hungry program, not several smaller
ones.
[3] The Linux kernel uses a different design, and does not suffer from
this problem.
[4] Microsoft's server products, especially Exchange[5], are
frequently hampered by this limit, when 2 GB is not enough, but
4294967296 GB is too much.
[5] Exchange before 2007 was limited to 32-bit and did not support
memory windowing.Exchange 2007 is 64-bit only.
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-17 Thread Jason Stephenson

If end users are defined as home users and office users, then 64 bits 
will never matter to them, just like 32 bits doesn't matter to them 
today. For the majority of people, its just a yard stick, like 4 
cylinder vs. 6 cylinder vs. 8. Most have some notion of what it means, 
that more is generally "better," but they don't often base their buying 
decision on that one metric.


They'll buy the computer that's available today that the salesman and 
the demo software or the review online or in the magazine says that they 
can do the work that they want to do on it. They don't particularly need 
to know or care that the CPU has a 64 bit address space. As long as it 
mostly works and does what they want to do in a "reasonable" amount of 
time, they are happy.


End users are buying 64-bit machines today, they're getting them with 
64-bit or 32-bit operating systems, and for the most part, it doesn't 
really matter to them. Very few of them would notice the difference in 
performance in the work that they are doing.


Sixty-four bitness will never matter to end users, but it will become 
ubiquitous over time, and sooner or later nearly everyone will have 
computers and devices with 64 bit CPUs and operating systems and the 
vast majority of people won't really notice.


Yes, there are uses for 64 bit address space, just as a 128 bit address 
space would enable use to tackle unthinkable problems. Yes, the 
performance issues are there, but if it really matter to John and Jane 
Computeruser, 64 bit computers would have become common place in the 
market before now and DEC would have bought Compaq.


As a friend of mine said in 1992, "I wish they'd stop wasting time with 
semiconductors and get on to doing computations with light."


Cheers,
Jason
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

On Fri, 2007-02-16 at 18:30 -0500, Bill McGonigle wrote:
> On Feb 16, 2007, at 14:31, Jon 'maddog' Hall wrote:
> 
> > I will note, however, that you can not mmap in an 8GB flash into a
> > single address space with a 32-bit processor.
> 
> maddog, is this another of your profound observations?  That 64-bit  
> addressing might be more interesting in the low-end/embedded space  
> than the high end?  Wouldn't that be a kick in the pants for  
> everybody expecting the 64-bit train to be headed southbound.
> 
> -Bill

(sigh)

Once upon a time in a place far, far away I talked a young man into
porting his 32-bit kernel to a 64-bit platform.

My main goal for this was to create a platform where researchers and
educators could investigate and expand programming algorithms in a place
where you never ran out of address spacea place where boundless
programming existed.

I recognized that whenever CPU address space had a step function, Don
Knuth would bring out another version of Volume 3 (Sorting and Searching
Techniques) in his never-finished but often re-wrote "Art of Computer
Programming".  I realized that every time address space went from 8-bit
to 16-bit to 32-bit he would bring out yet a few more algorithms that
now made sense, where they did not make sense before because the address
space was too small.

I actually wrote a talk about the concept of boundless programming.  I
utilized (believe it or not) the routines defined in the Posix Real Time
specification (I think it was Posix 1003.11_something-or-other):

mmap
lock
threads
semaphores
async I/O
(there are more.  I am tired)

and noted that with enough address space you could spread your data out
that you might not have collisions very often, and when you did you
might be able to take a little while to clean them up since they were so
seldom.

(By the way, these were techniques that I used back in 1973 on IBM
mainframes, with "huge" address spaces of 16 Mbytes.  A lot has to do
with the speed of the processor versus the size of memoryTemporal
Locality and Spatial Locality (or lack of it).)

A professor up at Dartmouth heard about this, and did a study about how
long it would take for a 64 bit address space to fill up if you never
re-used the addressesnever did garbage collection.  I tried to get
him an Alpha for this, but they were still in too much demand.  All I
could get him were some DecStation 3100s (mips based).

At DecStation 3100 speeds and "normal" programming loads, he estimated
that it would be seven hundred years (or something like that) before all
the space was used up and you had to do garbage collection.  We both
agreed that the processor would probably crash and have to be rebooted
before then (windows it was a given) so it was effectively "never" that
you would have to do garbage collection.  And if you were ever forced to
do it, it would be highly effective, probably recovering terabytes of
data area at a time with little overhead.

I was disappointed that no one have ever seemed to do much in the area
of this work.  Don still has not answered my challenge to re-write
Volume 3 again.  But he is retired and will not even read email.
Something I am considering lately.

Good night.

md

___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-16 Thread Bill McGonigle


On Feb 16, 2007, at 13:44, Ben Scott wrote:


 One is my understanding: Even if you're working with a 64-bit
architecture, isn't most software still dealing with 32-bit values?
Does throughput double without re-writing all the code to take
advantage of that?


I recall reading somewhere [I'll never find the reference, I'm sure]  
that there were some compiler optimizations to take two 32-bit ops  
(say in a loop) and optimize them into a single 64-bit op.  These  
depend on conditions that have been predicted a priori, though, like  
loop optimization.   And you'd have to assume that loading four ops  
into, say, an SSE3  unit, wouldn't be a better way to spend time.


I'm not sure what kind of operations would benefit from doing 2 at a  
time rather than 4 at a time.  My first reaction would be IEEE  
floating point or something with hardware assist that can't be fed  
into an integer vector processor.  But I'm consciously avoiding  
thinking about the binary representation of this because that kind of  
thing usually just results in neuronal injury on a Friday evening.


-Bill
-
Bill McGonigle, Owner   Work: 603.448.4440
BFC Computing, LLC  Home: 603.448.1668
[EMAIL PROTECTED]   Cell: 603.252.2606
http://www.bfccomputing.com/Page: 603.442.1833
Blog: http://blog.bfccomputing.com/
VCard: http://bfccomputing.com/vcard/bill.vcf

___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-16 Thread Bill McGonigle


On Feb 16, 2007, at 14:31, Jon 'maddog' Hall wrote:


I will note, however, that you can not mmap in an 8GB flash into a
single address space with a 32-bit processor.


maddog, is this another of your profound observations?  That 64-bit  
addressing might be more interesting in the low-end/embedded space  
than the high end?  Wouldn't that be a kick in the pants for  
everybody expecting the 64-bit train to be headed southbound.


-Bill

-
Bill McGonigle, Owner   Work: 603.448.4440
BFC Computing, LLC  Home: 603.448.1668
[EMAIL PROTECTED]   Cell: 603.252.2606
http://www.bfccomputing.com/Page: 603.442.1833
Blog: http://blog.bfccomputing.com/
VCard: http://bfccomputing.com/vcard/bill.vcf

___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-16 Thread Dan Jenkins


Bayard Coolidge wrote:


Well, I can think of at least one amateur radio application that comes to
mind, antenna modelling. And, of course, that's not limited to _amateur_
radio, but I'm sure the professionals might already be using "professional
workstations" from Sun, HP, IBM, or whomever, running a licensed UNIX
variant, etc., etc. 
 

Actually, the products I've encountered were all running under 64-bit 
Windows XP. They were in the $25,000 - $40,000 per seat range, which I 
gather may be the low  to mid end for these sorts of simulators.


Note: I know nothing of antenna modelling or microwave simulators, I 
just helped play one on TV... er, I meant helped setup the stations to 
evaluate them for a client. The Unix versions may have already been 
culled before they made their selections (mayhap for cost or data 
compatibility reasons). The simulations needed 8 GB of RAM and ran for days.


--
Dan Jenkins ([EMAIL PROTECTED])
Rastech Inc., Bedford, NH, USA --- 1-603-206-9951
*** Technical Support Excellence for over a Quarter Century

___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-16 Thread Jim Kuzdrall

On Friday 16 February 2007 14:27, Michael ODonnell wrote:
> >> End users of portable processing may benefit from clock rate
> >> reduction.  The 64-bit internal and main memory paths double the
> >> processor's instruction throughput.
>
> If all data paths doubled in width you'd certainly see increased
> througput/efficiency.  But many internal data paths in current
> designs are already wider than 32bits, and cache line sizes (which
> dominate the nature of exchanges w/main memory, and which are
> also already quite large) wouldn't necessarily double just because
> your programming model now allows you to utter 64bit addresses.

I did say "Almost".  Certainly, all of the FPU data lines are 64-bit 
- they even were in the 486, weren't they?

What most users will observe as higher performance won't come from 
the CPU.  The CPU already spends an awful lot of idle cycles waiting 
for I/O.

Faster mass memory (a replacement for mechanical disks) and faster 
Internet transfers will sell a lot more computers than doubling 
processing speed.

If the economies of fabrication make a 64-bit designs as cheap as 
32-bit designs, then the lower power might give a minor advantage.

Actually, what will happen is quite simple.  Microsoft buys large 
stake in Intel.  Vista-2010 will only run on 64-bit processors.  Quick.  
Buy Intel stock.

Jim Kuzdrall 

___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-16 Thread Ric Werme

Jim Kuzdrall wrote:

>End users of portable processing may benefit from clock rate 
>reduction.  The 64-bit internal and main memory paths double the 
>processor's instruction throughput.  A given 32-bit performance can be 
>had at half the clock rate in a 64-bit processor.  (Almost.  There are 
>other factors.)

Well, there are aa lot of factors.  Early Alpha code was 30% bigger than
32-bit MIPS code, some due to with 64 bit pointers, some due to with
instruction scheduling (well placed NOPs can make Alpha faster), and some due
to with the young compilers.  Also, just because you can write an instruction
that accesses 64 bits of data doesn't mean that's what memory sees.  Alpha and
other chips read a cache line worth at a time, 64 _bytes_ on I think all
implementations of Alpha.  That helps keep the cache control logic small and
fast.

If people are doing 32 bit math, a 64 bit ALU won't provide 2X the
performance.  You can add more ALUs, barrel shifters, etc, but those are
power hungry beasts.

>The internally consumed power is proportional to the clock rate for 
>CMOS technology.  Laptops and handhelds which are limited by battery 
>life or heat removal may find the additional PC board area and cost 
>justified.

Don't forget backlight, disk, and other substantial power consumers!

  -Ric Werme
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-16 Thread Jerry Feldman

On Fri, 16 Feb 2007 14:14:59 -0500
"Jon 'maddog' Hall" <[EMAIL PROTECTED]> wrote:

> I know that Intel did not invent complex instructions, but they have
> done more to break compiler-writer's hearts than any other company.
Naw, they are keeping compiler writers in business. All our old friends
from the GEM group in Spitbrook Rd. are still up there wearing Intel
shirts. Don Harbert is Clem Cole's boss, but Steve Hobbs just retired.
Itanic is a compiler writer's dream because they always have a
challenge to make their code work.

-- 
Jerry Feldman <[EMAIL PROTECTED]>
Boston Linux and Unix user group
http://www.blu.org PGP key id:C5061EA9
PGP Key fingerprint:053C 73EC 3AC1 5C44 3E14 9245 FB00 3ED5 C506 1EA9


signature.asc
Description: PGP signature
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

> It's not *end-user computing*.  End-user computing is stuff my mom does, or 
> your Aunt
> Marge or the high school teenager.  :-)

Sorry, I thought we were talking English, and where I come from, and
"end user" is *anyone* that *uses* the software instead of making it or
administering it, whether they do it as a job or not.  Even developers
are "end users" of compilers and other tools, since they use them to do
their work.

I gave you the term the industry uses for what you seem to be talking
about is "commercial computing" or "home computing" and for the most
part the answer is that a 64-bit address space is not something they
need, although they can benefit from the rest of the 64-bit
architecture.

I will note, however, that you can not mmap in an 8GB flash into a
single address space with a 32-bit processor.

md

___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-16 Thread Michael ODonnell



>> End users of portable processing may benefit from clock rate
>> reduction.  The 64-bit internal and main memory paths double the
>> processor's instruction throughput.

If all data paths doubled in width you'd certainly see increased
througput/efficiency.  But many internal data paths in current
designs are already wider than 32bits, and cache line sizes (which
dominate the nature of exchanges w/main memory, and which are
also already quite large) wouldn't necessarily double just because
your programming model now allows you to utter 64bit addresses.
 
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)


> Does throughput double without re-writing all the code to take
> advantage of that?
Yes.

And other issues such as double-word arithmetic (especially in floating
point) are done as a single clock-tick.  But this has little to do with
larger address space.

> 
>   The second is semi-historical: As I recall, your argument is the
> same argument Intel put forth in favor of Itanic -- I mean, Idiotanium
> -- I mean, Itanium -- and we all know that is going nowhere at
> velocities approaching C.
> 

/*HOLD FOR BIGGEST RANT OF MONTH*/

The Itanic did not take off because it simply broke all binary
compatibility on the theory that they could make it up with emulation
and everyone would be happy.  Wrong on both counts.

The Itanic's instruction set was (and I use the past tense purposely) an
ultra-wide instruction set that was studied and rejected in computer
science years ago, but it still keeps coming up like bad coffee, and
some jerk in Intel/HP probably thought that they could get a couple of
patents on the instruction set to keep AMD from copying it.

Complex instructions sets in general, and ultra-wide instruction sets in
particular, are a compiler-writer's nightmare.  They are "tuned" to
particular job flows that we have no idea what they are, where they come
from, or if they ever will exist.

Compare this to the Alpha, with its very clean RISC instruction set that
allows the language choice, compiler choice, and pragmas to tell the CPU
what the job flow will be and how to optimize.  And THEN, even after
creating the object code, the CPU would re-optimize instruction and
register execution on the fly.

I know that Intel did not invent complex instructions, but they have
done more to break compiler-writer's hearts than any other company.

Thank heaven AMD made them understand that compatibility was important,
and that you did not have to throw the baby out with the bathwater.

/* END OF RANT*/

And if I never see another ultra-wide instruction again, my life will be
much happier.

maddog

___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-16 Thread Bayard Coolidge

Well, I can think of at least one amateur radio application that comes to
mind, antenna modelling. And, of course, that's not limited to _amateur_
radio, but I'm sure the professionals might already be using "professional
workstations" from Sun, HP, IBM, or whomever, running a licensed UNIX
variant, etc., etc. 

But I have noted a surge in interest in the amateur radio community in
antenna modelling using the NEC code. For a long time, "lite" versions
have been available for use under Windows, but some full-strength versions for 
Linux are now available. With the availability of 64 bit versions and fast CPUs 
and large(r) physical memory available to the
ordinary home user at a reasonable price, antenna modelling that was
once the domain of scientists can be readily done in the ham shack.

OK, granted this is specific to a special hobby/interest, but I'm sure
there are other applications, formerly ensconced in research labs, that
could now be readily used at home now.

73,

Bayard, N1HO

"Brake for moose - it can save your life -- NHF&GD"

Ben Scott said:

  I've been asking the question, "How would an end-user significantly
benefit from x86-64?"


 
-
We won't tell. Get more on shows you hate to love
(and love to hate): Yahoo! TV's Guilty Pleasures list.___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-16 Thread Ben Scott


On 2/16/07, Jim Kuzdrall <[EMAIL PROTECTED]> wrote:

  I've been asking the question, "How would an end-user significantly
benefit from x86-64?" ... Anyone got any other ideas?


End users of portable processing may benefit from clock rate
reduction.  The 64-bit internal and main memory paths double the
processor's instruction throughput.


 Hmmm.  Interesting idea.

 Two issues I see:

 One is my understanding: Even if you're working with a 64-bit
architecture, isn't most software still dealing with 32-bit values?
Does throughput double without re-writing all the code to take
advantage of that?

 The second is semi-historical: As I recall, your argument is the
same argument Intel put forth in favor of Itanic -- I mean, Idiotanium
-- I mean, Itanium -- and we all know that is going nowhere at
velocities approaching C.

-- Ben
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-16 Thread Ben Scott


On 2/16/07, Jon 'maddog' Hall <[EMAIL PROTECTED]> wrote:

You seem to be defining every "end user" as "mom-and-pop-home", or "bank
teller".  In the scientific and engineering world ...


 NOTE WELL: The following exasperated rant is written with a smile on
my face and laughter in my throat.  :-)

 Oh, for crying out loud, *LIKE I WROTE IN MY MESSAGE*, "heavy duty"
computing is outside of the scope of this discussion.  Yes, I am
explicitly defining the term "end-user" that way, for the purposes of
this discussion, because that's what I'm interested in.  It's already
*well established* that large memories or bigger integers benefit
heavy duty computing.  Much of that world has *already* moved to
64-bit computing for that very reason.  Which is why we *don't need to
have a discussion about it*.  :-)

 People doing that are not doing end-user computing.  Nobody (well,
almost nobody) models complex weather systems as a casual hobby.  They
do it because it's they're freaking *job*.  It's not *end-user
computing*.  End-user computing is stuff my mom does, or your Aunt
Marge or the high school teenager.  :-)

 Meanwhile, I find the question of "Will the end-user -- the "mom and
pop", the "bank teller", the "home user", the MySpace people, the big
list of things they do I wrote in my original message -- will *they*
reap any benefit from x86-64?", that question is more interesting.  If
there's a compelling reason for the *that* population to move to
x86-64, then things get interesting.

 If there is *not* a compelling case for that, well, I expect 64-bit
computing will remain in the domain of the "heavy duty" -- servers,
multimedia production, engineering, scientific, all the back-room
number-crunching and bit-moving operations where Linux already has a
significant foothold.  People in that space might use Linux (or
Solaris or HP-UX or some other non-doze OS) precisely *because*
Microsoft's support of 64-bit computing sucks so much.

 But meanwhile, in the "end-user" space, if some compelling reasons
to use x86-64 exists, then that's one more selling point for Linux in
the end-user space, and might well contribute to critical mass.  Which
is why I find this question interesting, and how I think it ties into
FLOSS.  It could be part of that "killer app" the pundits are always
talking about.

 Is that clear enough, or do I need diagrams?  ;-)

-- Ben
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-16 Thread Jim Kuzdrall


>   I've been asking the question, "How would an end-user significantly
> benefit from x86-64?"
>   Anyone got any other ideas?

End users of portable processing may benefit from clock rate 
reduction.  The 64-bit internal and main memory paths double the 
processor's instruction throughput.  A given 32-bit performance can be 
had at half the clock rate in a 64-bit processor.  (Almost.  There are 
other factors.)

The internally consumed power is proportional to the clock rate for 
CMOS technology.  Laptops and handhelds which are limited by battery 
life or heat removal may find the additional PC board area and cost 
justified.

As For performance, the 64-bit instruction fetch from main memory 
and instruction decoder may be a bigger advantage than the 64-bit ALU.  
Main memory access time is still a big bottleneck.

   Presumably the data for video has to be continually read into cache 
from main memory, 64-bit transfers would effectively double the 
performance at the same internal clock speed.

Jim Kuzdrall
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)

2007-02-16 Thread Tom Buskey

On 2/16/07, Jon 'maddog' Hall <[EMAIL PROTECTED]> wrote:

On Fri, 2007-02-16 at 11:35 -0500, Ben Scott wrote:

  I've been asking the question, "How would an end-user significantly
> benefit from x86-64?"

Define the end user as a home user with hobbies or std office
drone^H^H^H^H^Huser.

>   Last night at Martha's, Matt Brodeur pointed out one thing that
> could actually count: Full motion video editing.  This is something

That's one thing I've thought of.  Some of that is done with 32 bits also.

>   Anyone got any other ideas?

Maybe someone will come up with a way to meaningfully tag and index photos
automatically and it will require 64 bits.

> other things perused by organizations.  I'm talking about what *end
> users* use their PCs for.  The people browsing MySpace, forwarding
> their email chain letters, downloading illegal music with
> Kanapsterwire, and looking at pron.)
Ben,

You seem to be defining every "end user" as "mom-and-pop-home", or "bank
teller".

Yes, the people that drive buying a computer.  The mass market.

In the scientific and engineering world we have things like CAD packages

I'd argue that these users are a boundary case.  In years past, they bougth
super computers and workstations when mom-and pops and std office users
bought Apple ][s and the IBM PC.

Weather forecasting, global modeling, and even Matt's example of video

are all lumped together into what most system vendors call "scientific
and engineering", making up about 16% of the total system market.  The
other 84% belongs to "commercial computing" (transaction based, small
processes that live a short life).

And as 32 bit systems and concepts like Beowulf systems came along, the 84%
systems moved into the space the 16% was in.

I don't think the scientific computing space has as much influence on the
commercial computing hardware as the other way around.  Software might be a
different story.
___
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Re: [OT] End-user uses for x86-64 (was: Why are still not at 64 bits)