Whose waiting until dual-cores?

[Guest]

Archived from groups: alt.sys.pc-clone.dell (More info?)

Anyone holding back from buying/upgrading their dinosaur until dual-cores
come
out towards the end of this year? I am, however, if they drop prices to
clear
stock when the new chips come out, I might just jump in and find a bargain.

 

[Guest]

Archived from groups: alt.sys.pc-clone.dell (More info?)

"Buck Rogers" <who@cares.com.au> wrote in message
newspsngvcskpkru5fy@mycomput-fdagxy...
> Anyone holding back from buying/upgrading their dinosaur until dual-cores
> come
> out towards the end of this year? I am, however, if they drop prices to
> clear
> stock when the new chips come out, I might just jump in and find a
> bargain.

New features are neat, but there's also the issue of refinement. For
example, the newest (600-series) Prescotts look like a nicer package than
the original version. Maybe the dual procs will be wonderful out of the
gate. OTOH, this might be a good time to buy a modern Prescott, and let the
dual procs evolve for one or two iterations.

I wonder what the real-world typical-user improvement will be for a dual
processor?

 

[Guest]

Archived from groups: alt.sys.pc-clone.dell (More info?)

"Talkin Horse" <davidrolfeN0SP&AM@earthlink.net> wrote:

>"Buck Rogers" <who@cares.com.au> wrote:

>> Anyone holding back from buying/upgrading their dinosaur until dual-cores
>> come
>> out towards the end of this year? I am, however, if they drop prices to
>> clear
>> stock when the new chips come out, I might just jump in and find a
>> bargain.

>New features are neat, but there's also the issue of refinement. For
>example, the newest (600-series) Prescotts look like a nicer package than
>the original version. Maybe the dual procs will be wonderful out of the
>gate. OTOH, this might be a good time to buy a modern Prescott, and let the
>dual procs evolve for one or two iterations.
>
>I wonder what the real-world typical-user improvement will be for a dual
>processor?

You seem to be making the assumption that there actually is such
a thing as a "real-world *typical*-user" instead of what I
suspect is a continuum of "typical" users ranging from the most
clueless noob to the gamer-cum-master computer geek desirous of
always being at the cutting edge, with bulges in the curve at
various points, e.g., typical home/small office user doing
word-processing, spreadsheets, some simple, presentation-type
graphics work, email and mild websurfing, maybe some Usenet,
chatting, BBing, and/or blogging; a) plus a gaming nut in the
household; a) and mebbe b) plus an MP3/movies/whatever nut
downloading everything in sight in the household; or a) plus a
professional engineering/scientific type doing computationally
intense work at work or at home; etc., etc., etc.

I am solidly in the a) category. My usage is plain vanilla and
boring - my "living on the edge" days are two decades back, at
least - which leads to my question, since I have not been
following CPU developments since the first Pentium. WITH is this
"dual-core" thing [and, now that you mention it, that Prescott
thing that I ignored before], and where in the user continuum
does one have to be before one would see any benefit, beyond
"incremental" [e.g., like were I to move from 1.7/1.8 to 3.0 ghz
in a PIV with my usage] and essentially un-noticable, by moving
from ones Buck Rogers-despised "dinosaur" computer to this dual
core stuff.?
--
OJ III
[Email to Yahoo address may be burned before reading.
Lower and crunch the sig and you'll net me at comcast.]

 

[Guest]

Archived from groups: alt.sys.pc-clone.dell (More info?)

"Ogden Johnson III" <oj3usmc@yahoo.com> wrote in message
news:64d6319u6hvq1vunaim57d1imtmbgfsi1n@4ax.com...
> My usage is plain vanilla and
> boring - my "living on the edge" days are two decades back, at
> least - which leads to my question, since I have not been
> following CPU developments since the first Pentium. WITH is this
> "dual-core" thing [and, now that you mention it, that Prescott
> thing that I ignored before], and where in the user continuum
> does one have to be before one would see any benefit, beyond
> "incremental" [e.g., like were I to move from 1.7/1.8 to 3.0 ghz
> in a PIV with my usage] and essentially un-noticable, by moving
> from ones Buck Rogers-despised "dinosaur" computer to this dual
> core stuff.?

Yes and no. Of course each user is a unique individual, but that doesn't
mean we can't toss around a few general principles, and you can decide
whether they apply to your situation. I like to keep certain things going
until they fall apart. My car is 10 years old and my telephone is 15.
Computers aren't in the same category, since the computer is usually used in
a way that it must remain in synch with the rest of the world. So your 100
MHz Pentium Windows95 clunker may run fine as a stand-alone word processor,
but if you attach it to the Internet or exchange files with other people or
install any modern software or have any tech support questions, you'll
likely run into a wall. You might squeeze 5 years of useful life out of a
PC, but you won't try for 10 unless it's really peripheral to your life --
in which case you're not likely to be perusing this newsgroup. My PC is
approaching 3 years old, and it can still keep up, but the age is starting
to show: Not quite as fast, or as much storage, as newer software expects.
So the question is, when is a good time to get the next machine? When has
(or will) a milestone of performance been reached that will not be quickly
obsoleted? For example, I put off buying my current PC until USB2 became a
standard feature, because this was clearly going to become essential. Of
course, it's often possible to patch a system with minor upgrades, and many
people like to do this, but it's nice to have the essential features built
into the core. So your 1.7 GHz PC may have some life left in it, but it
perhaps doesn't have DVI, writeable DVD, USB2, or some other features. Such
things evolve from luxuries to essential features. So if someone is starting
to think about buying a new PC at this time and hopes to get several years
use out of it, the question is: Is it better to buy a refined Prescott
(along with the other modern hardware), or hold out and wait for the dual
processor? Some of us find that question fun as well as useful.

 

[Guest]

Archived from groups: alt.sys.pc-clone.dell (More info?)

"Ogden Johnson III" <oj3usmc@yahoo.com> wrote in message news:64d6319u6hvq1vunaim57d1imtmbgfsi1n@4ax.com...

> ... WITH is this "dual-core" thing

For starters, think dual processor system, but instead of having two physical
chips, you have one chip that contains both processors. A more specific and
relevant example would be two 90nm Prescott class "cores" plus some
glue/arbitration logic built onto a single die.

> [and, now that you mention it, that Prescott thing that I ignored before]...

Doh.

During the recent IDF (Spring 2005 Intel Developers Forum), Intel provided
more information about their future processors. All of the major tech news
sites have some coverage. Google keywords: IDF smithfield presler

 

[Brian]

Archived from groups: alt.sys.pc-clone.dell (More info?)

Buck Rogers <who@cares.com.au> wrote:
> Anyone holding back from buying/upgrading their dinosaur until dual-cores
> come
> out towards the end of this year? I am, however, if they drop prices to
> clear
> stock when the new chips come out, I might just jump in and find a bargain.

I'm waiting for the IBM 970 myself.

 

[Guest]

Archived from groups: alt.sys.pc-clone.dell (More info?)

Talkin Horse wrote:
> "Buck Rogers" <who@cares.com.au> wrote in message
> newspsngvcskpkru5fy@mycomput-fdagxy...
>
>>Anyone holding back from buying/upgrading their dinosaur until dual-cores
>>come
>>out towards the end of this year? I am, however, if they drop prices to
>>clear
>>stock when the new chips come out, I might just jump in and find a
>>bargain.
>
>
> New features are neat, but there's also the issue of refinement. For
> example, the newest (600-series) Prescotts look like a nicer package than
> the original version. Maybe the dual procs will be wonderful out of the
> gate. OTOH, this might be a good time to buy a modern Prescott, and let the
> dual procs evolve for one or two iterations.
>
> I wonder what the real-world typical-user improvement will be for a dual
> processor?

I'm inclined to say "not much" at first. As far as I can tell from
reading things, a dual-core CPU will work much like an existing
motherboard with dual single-core CPUs. Not much software currently
takes advantage of dual cpus, so little performance gain is seen.
It all depends on exactly what you do, of course. Perhaps if you run a
number of apps at once you might see more benefit than if you only do
one main thing at a time.
My feeling would be to wait a while and let the software catch up with
dual-core. The more dual-core machines out there, the more incentive the
software developers will have to make their products use it.

 

[Guest]

Archived from groups: alt.sys.pc-clone.dell (More info?)

John Fryatt wrote:
> Talkin Horse wrote:
>
>> "Buck Rogers" <who@cares.com.au> wrote in message
>> newspsngvcskpkru5fy@mycomput-fdagxy...
>>
>>> Anyone holding back from buying/upgrading their dinosaur until
>>> dual-cores come
>>> out towards the end of this year? I am, however, if they drop prices
>>> to clear
>>> stock when the new chips come out, I might just jump in and find a
>>> bargain.
>>
>>
>>
>> New features are neat, but there's also the issue of refinement. For
>> example, the newest (600-series) Prescotts look like a nicer package
>> than the original version. Maybe the dual procs will be wonderful out
>> of the gate. OTOH, this might be a good time to buy a modern Prescott,
>> and let the dual procs evolve for one or two iterations.
>>
>> I wonder what the real-world typical-user improvement will be for a
>> dual processor?
>
>
> I'm inclined to say "not much" at first. As far as I can tell from
> reading things, a dual-core CPU will work much like an existing
> motherboard with dual single-core CPUs. Not much software currently
> takes advantage of dual cpus, so little performance gain is seen.
> It all depends on exactly what you do, of course. Perhaps if you run a
> number of apps at once you might see more benefit than if you only do
> one main thing at a time.
> My feeling would be to wait a while and let the software catch up with
> dual-core. The more dual-core machines out there, the more incentive the
> software developers will have to make their products use it.
>
You're correct, designing programs to utilize multiple threads is fairly
complex, and until recently almost never implemented in consumer
programs (servers on the other hand are another story). The main
advantage will be load balancing, and it should be interesting to see
how these dualcores compare to true SMP systems (like will the shared
cache have an effect, etc.).

 

[Brian]

Archived from groups: alt.sys.pc-clone.dell (More info?)

Nicholas Andrade <sdnick484@nospam.yahoo.com> wrote:
>> I'm inclined to say "not much" at first. As far as I can tell from
>> reading things, a dual-core CPU will work much like an existing
>> motherboard with dual single-core CPUs. Not much software currently
>> takes advantage of dual cpus, so little performance gain is seen.
>> It all depends on exactly what you do, of course. Perhaps if you run a
>> number of apps at once you might see more benefit than if you only do
>> one main thing at a time.
>> My feeling would be to wait a while and let the software catch up with
>> dual-core. The more dual-core machines out there, the more incentive the
>> software developers will have to make their products use it.
>>
> You're correct, designing programs to utilize multiple threads is fairly
> complex, and until recently almost never implemented in consumer
> programs (servers on the other hand are another story). The main
> advantage will be load balancing, and it should be interesting to see
> how these dualcores compare to true SMP systems (like will the shared
> cache have an effect, etc.).

I feel like something's changed and I missed out on it.

Nearly everything in the past decade should be programmed with at least
two threads. One thread handles the user interface, the other threads
handle "anything else." Printing, automatic spell checking, POP mail
checking, etc etc are all helper threads. Explorer is claiming a few
dozen threads right now on my system, from tooltips to form elements.

Only a few programs have NOT used multithreading properly. Notepad
is a good example. It grinds to a halt whenever anything over 1MB is
opened. UltraEdit by comparison works like a champ. And surprisingly
(to me anyway) iTunes is a threading nightmare. I don't recall any
program grinding my system to a halt as effectively and consistently
as iTunes.

But anyway, of the dozens of threads in the run queue or block queue,
I agree, getting 100% utilization is a lot like bin stacking problems.
It still doesn't mean the efficiency doesn't settle in at a favorable
number. Even hyper-threading benchmarks look attractive.