Scott Reams wrote:

To be fair... Intel has no 64bit chips in the consumer pipeline, and there
are none on their roadmaps. The only 64bit chip Intel offers is the
ultra-expensive Itanium for servers.

Scott, how do you see 64 bit chips benefiting audio? From
having worked in CPU architecture many years ago all I see
is a potential for requiring a higher data and instruction
bandwidth to accomplish the same work.

If and when there is a nonvolatile ram technology of
sufficient capacity I can see a gain but I feel it can only
benefit us now by forcing device scaling which may improve
the performance of the 32 bit subset of the architecture by
default and then only if that subset can operate without the
overhead penalty of supporting 64.


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

Bob Cain wrote:
Scott Reams wrote:

To be fair... Intel has no 64bit chips in the consumer pipeline, and there
are none on their roadmaps. The only 64bit chip Intel offers is the
ultra-expensive Itanium for servers.

Scott, how do you see 64 bit chips benefiting audio? From
having worked in CPU architecture many years ago all I see
is a potential for requiring a higher data and instruction
bandwidth to accomplish the same work.

I can see 64 bit coefficients being a handy thing for some kinds of processing.
Then again, if you don't care about running realtime, there is no reason you
can't do 64 bit arithmetic with an 8-bit machine.

The other thing I see is that some of the machine with a 64 bit architecture
might have faster 32 bit floating point operations if stuff is put into the
ALUs to do simultaneous halfword operations. The SPARC 3 did this and it was
a slight win.

If and when there is a nonvolatile ram technology of
sufficient capacity I can see a gain but I feel it can only
benefit us now by forcing device scaling which may improve
the performance of the 32 bit subset of the architecture by
default and then only if that subset can operate without the
overhead penalty of supporting 64.

Again, address space and word length are independant.... you can get all
the address space you want with the expense of nastiness like segmentation.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Scott Dorsey wrote:

Again, address space and word length are independant.... you can get all
the address space you want with the expense of nastiness like segmentation.

To what, then, do attribute the move to 64 bits?


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

And my Dell Optiplex GX150 running Windows XP Professional hasn't crashed
ONCE in over 2 years I've been using it, day to day.

thats great! my dell inspiron crashes daily...blue screen. this has happened
since it came out of the box. my hp crashes once daily.... locks up. atleast i
can count on its regularity. my mac crashes, on start up, once every couple of
weeks, usually when i am clearing out specific extensions. nothings perfect,
but damn would i love to get my $'s back from hell dell.
scott spelbring | recording + interactive | dragonflyeast.com

Don't feel bad. For years my older son couldn't touch a computer without
killing it. I mean, I'd build a computer, burn it in for days doing stress
tests and such, give it to him, and within two days it would be dead.

So recoverable crashes are MUCH better than that! g

--


Roger W. Norman
SirMusic Studio

301-585-4681




"scott spelbring" wrote in message
...
And my Dell Optiplex GX150 running Windows XP Professional hasn't crashed
ONCE in over 2 years I've been using it, day to day.

thats great! my dell inspiron crashes daily...blue screen. this has
happened
since it came out of the box. my hp crashes once daily.... locks up.
atleast i
can count on its regularity. my mac crashes, on start up, once every
couple of
weeks, usually when i am clearing out specific extensions. nothings
perfect,
but damn would i love to get my $'s back from hell dell.
scott spelbring | recording + interactive | dragonflyeast.com

Personally I don't know what all the "fear" is in a Mac user contemplating a
PC or vice versa. It's just a tool, and like any tool the use is mostly
dependant upon the application. Obviously anyone looking to buy a new
computer today should look at the application they wish to run and buy the
appropriate platform, and buy the best damned one they can afford. I'd
rather opt for stability and reliability over a period of 18+ months (almost
getting to be an unrealistic requirement anymore) than upgradability.

This system is on it's last 6 months at the Athlon 1600+ level and already
there are now 3200+ CPUs. But with multiple SCSI cards (2940 and a
fibrechannel), 180 gigs, dual head, etc., I won't necessarily be upgrading
until I really find it necessary.

At least the new G5 looks to be a solid machine. Were I inclined to move to
Pro Tools it wouldn't concern me the least in making the move to a Mac.

--


Roger W. Norman
SirMusic Studio

301-585-4681




"Scott Reams" wrote in message
m...
Ahh, but I want OS-X, so it´s irrelevant.
I think this is part of the argument that get´s lost.

It doesn't get lost. It's just one of several factors.

What OS someone prefers is up to them. Some might choose Apple... some MS.
The performance and price relationships are still important to those not
absolutely determined to use one OS or the other (as in... Nuendo or
CubaseSX users, where OS has little impact on how the app functions).

-S

Bob Cain wrote:
Scott Dorsey wrote:

Again, address space and word length are independant.... you can get all
the address space you want with the expense of nastiness like segmentation.

To what, then, do attribute the move to 64 bits?

The fact that being able to move data in 64 bit chunks is an enormous win
if you are doing 64 bit arithmetic, and that can be a big deal for some
stuff. Code that ran in double precision mode now runs in single, with
a considerable speed improvement.

But the 32-64 jump isn't going to be anywhere near as big as the 16-32
jump was, just because there is less of a need to do that for typical
applications.

The wider data word also allows a big improvement in the speed of block
operations, where you're moving arrays from place to place in memory,
since you can now do it in 64 bit chunks.

And if you are really good, or your compiler is really good, you can do
things like store adjacent 32 bit values into a single 64 bit word and
then use 64 bit operations to load and store them both at the same time.
This is less of a win than it might seem at first.

But address space is the least of your worries.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Scott Dorsey wrote:

But address space is the least of your worries.

On the contrary, from my time in architecture years ago I'm
pretty certain that the movement to 64 bit addressing and
the attendant enhancements to storage management and access
is the sine qua non for the next level of performance in
database management and transaction processing applications
which is where the bulk of business and the bulk of computer
profit is still entrenched. The direct addressability of
huge amounts of RAM is almost entirely what's driving it.
See:

http://www.eweek.com/print_article/0...a=40723,00.asp

For audio it offers almost no perceivable benefit. For
calculation we just don't need it and wider busses for the
more rapid movement of blocks of data don't require 64 bit
integer architectures. Busses wider than 32 bits have been
around for a very long time for the purposes of higher
memory bandwidth.


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

(a b) wrote in message . com...

intel and amd have several 64bit chips in the consumer pipeline. in
all liklihood, they will be more powerful and cheaper than the apple
offerings. and once again apple will be priced out of the mass market.


Right, but if they started making poop so as not to be priced out of
the MASS market...ALL computers would suck as opposed to just 95% of
them.

Or API could always start making Behringer quality preamps and get
into the bigtime!

-BH

Well, there would be speed advantages by offering huge amounts of
addressable memory space. Instant track bouncing, no pre-loaded caching,
instantiations of plugins that simply sit there waiting without crapping out
a cpu. Little things like that! g

Isn't it ridiculous that what was perceived as a major advantage just a
couple of years ago is now a time waster! g

--


Roger W. Norman
SirMusic Studio

301-585-4681




"Bob Cain" wrote in message
...


Scott Dorsey wrote:

But address space is the least of your worries.

On the contrary, from my time in architecture years ago I'm
pretty certain that the movement to 64 bit addressing and
the attendant enhancements to storage management and access
is the sine qua non for the next level of performance in
database management and transaction processing applications
which is where the bulk of business and the bulk of computer
profit is still entrenched. The direct addressability of
huge amounts of RAM is almost entirely what's driving it.
See:

http://www.eweek.com/print_article/0...a=40723,00.asp

For audio it offers almost no perceivable benefit. For
calculation we just don't need it and wider busses for the
more rapid movement of blocks of data don't require 64 bit
integer architectures. Busses wider than 32 bits have been
around for a very long time for the purposes of higher
memory bandwidth.


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

The obvious speed advantage comes from addressable memory. Unless a quatum
leap in hd speeds hits inexpensively, the current bottleneck of hard disk
access is going to play an even more significant part. And I don't mean
faster bussing, although that plays a part. I mean pure raw writing and
reading speed. So by virtue of having huge improvement in addressable
space, memory becomes the arena whereby any reasonable improvements in
system speed will be seen initially. With something like Samplitude, and
it's ability to do full multitrack recordings in memory, I could easily see
having 10s of gigs of memory and a system with the responsiveness the likes
of which have yet to be seen.

The new Opteron that Scott Reams keeps talking about has up to 19.2 GB/s of
bus speed, which means a lot of memory would be an ideal condition to use
for recording. Couple it with a VERY IMPORTANT UPS and one could have a
tremendous system, not unlike AT&T's ridiculously expensive DISC system.

Now if IBM could just get that holographic memory down pat and put out an
inexpensive blue laser setup, we could afford the terabytes of memory we are
all going to need to have! g

--


Roger W. Norman
SirMusic Studio

301-585-4681




"Scott Dorsey" wrote in message
...
Bob Cain wrote:
Scott Dorsey wrote:

Again, address space and word length are independant.... you can get
all
the address space you want with the expense of nastiness like
segmentation.

To what, then, do attribute the move to 64 bits?

The fact that being able to move data in 64 bit chunks is an enormous win
if you are doing 64 bit arithmetic, and that can be a big deal for some
stuff. Code that ran in double precision mode now runs in single, with
a considerable speed improvement.

But the 32-64 jump isn't going to be anywhere near as big as the 16-32
jump was, just because there is less of a need to do that for typical
applications.

The wider data word also allows a big improvement in the speed of block
operations, where you're moving arrays from place to place in memory,
since you can now do it in 64 bit chunks.

And if you are really good, or your compiler is really good, you can do
things like store adjacent 32 bit values into a single 64 bit word and
then use 64 bit operations to load and store them both at the same time.
This is less of a win than it might seem at first.

But address space is the least of your worries.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."