New Dual 2 Core CPU

[SamuelF566]

After reading many articles on the new Dual 2 Core, it's obvious that
this could be a major reason to uprade now if you really want the
greatest/fastest computer on the block. Unfortunately AMD Is announcing
a Quad chip too claiming it will be even faster. Personally my
Pentium 4 2.4 ghz with DSL Is plenty fast for everything I do and will
not upgrade at least till prices drop drastically till prices are
around $150.00 for MB And CPU.
So who will buy these new chips or systems with this new technology?
Gamers ...serious gamers or people into serious video editing needs.

 

[Mark]

After reading many articles on the new Dual 2 Core, it's obvious that
this could be a major reason to uprade now if you really want the
greatest/fastest computer on the block. Unfortunately AMD Is announcing
a Quad chip too claiming it will be even faster. Personally my
Pentium 4 2.4 ghz with DSL Is plenty fast for everything I do and will
not upgrade at least till prices drop drastically till prices are
around $150.00 for MB And CPU.
So who will buy these new chips or systems with this new technology?
Gamers ...serious gamers or people into serious video editing needs.

Anyone who needs more processing power than was available before the Core 2
duo's were released. This includes gamers, video editors, dvd encoding, and
scientific or business software that needs to process a lot of data. I am
getting one mostly for games and Matlab.

 

[Paul B]

Thus spake Mark:

Anyone who needs more processing power than was available before the
Core 2 duo's were released. This includes gamers, video editors, dvd
encoding, and scientific or business software that needs to process a
lot of data. I am getting one mostly for games and Matlab.

And people like me who haven't upgraded for years - I bought an XP-M2500 &
run it reliably @ 200x11.5x to replace an XP1800+ 18 months ago. The extra
speed was underwhelming. As for quad cores, many have suggested that's just
AMD's panicked response to having been left with their pants down rather
than a serious technical solution. My hope is that AMD will never be so
complacent again - though I sure a few pundits were also surprised that
Intel actually managed to produce something so good! As for gaining an extra
turn of speed, newer (read faster) h/ds can often give more than a
incremental CPU u/g does.

 

[daytripper]

Thus spake Mark:

And people like me who haven't upgraded for years - I bought an XP-M2500 &
run it reliably @ 200x11.5x to replace an XP1800+ 18 months ago. The extra
speed was underwhelming. As for quad cores, many have suggested that's just
AMD's panicked response to having been left with their pants down rather
than a serious technical solution. My hope is that AMD will never be so
complacent again - though I sure a few pundits were also surprised that
Intel actually managed to produce something so good! As for gaining an extra
turn of speed, newer (read faster) h/ds can often give more than a
incremental CPU u/g does.

wrt to your take on AMD and quad-core packages, does the codename "Clovertown"
mean anything to you?

 

[Paul B]

Thus spake daytripper:

wrt to your take on AMD and quad-core packages, does the codename
"Clovertown" mean anything to you?

If A is better than B, n.A will always be better than n.B!

Intel's foray into quad cores, of course. Whatever AMD currently does, Intel
can copy. AMD, IMO are themselves now in the position they put Intel in -
playing catch-up with red faces but probably good for consumers & AMD in the
long run. Just how much s/w consumers are likely to run is really going to
benefit from 4 cores?

 

[Tony Hill]

After reading many articles on the new Dual 2 Core, it's obvious that
this could be a major reason to uprade now if you really want the
greatest/fastest computer on the block. Unfortunately AMD Is announcing
a Quad chip too claiming it will be even faster.

Quad-core chips are primarily going to target the server market for
the first little bit. The differences in the nature of what servers
do vs. what desktops do tend mean that servers will benefit MUCH more
from quad-core chips than desktops. Similar to the way that SCSI
tends to be a lot faster on servers but SATA tends to be just as fast
or faster on desktops.

Personally my
Pentium 4 2.4 ghz with DSL Is plenty fast for everything I do and will
not upgrade at least till prices drop drastically till prices are
around $150.00 for MB And CPU.

That should be in about a years time, maybe a year and a half tops.

So who will buy these new chips or systems with this new technology?
Gamers ...serious gamers or people into serious video editing needs.

Anyone who needs a new computer would probably be considering it.
LOTS of people out there (such as myself) have slower systems than
your 2.4GHz P4 and are finding that they are growing a little long in
the tooth for some tasks. Others have more demanding work that could
benefit from the processor. Certainly video editing and gamers are
two examples that jump to mind, but they are hardly the only ones.

Really not much has changed now as compared to when Intel released the
386 ~20 years ago. Yeah, it was more expensive than the 286 and lots
of people could get by with the older chip, they were happy and didn't
see who would be buying this new super-duper 386 chip. Same was true
when the 486 came out, then the Pentium, the PII, PIII, P4 and now the
Core 2 Duo. And you know what? The same will be true when Intel and
AMD bring out their next big thing.

The biggest difference between now and when the 386 was released is
that the Core 2 Duo is *MUCH* cheaper for the increase in performance
it offers. A new 386 system, when accounting for inflation, would
probably have been around $5000 for about a 2-fold improvement in
performance. A new Core 2 Duo system would offer roughly a 2-fold
improvement in overall performance vs. your current P4 setup, and
it'll only set you back about $1000.

 

[nobody]

Thus spake daytripper:

If A is better than B, n.A will always be better than n.B!

If A already saturates the bus with memory access, nA will be
bandwidth-starved.
If B scales linearly (OK, almost linearly), nB will likely beat nA.

Intel's foray into quad cores, of course. Whatever AMD currently does, Intel
can copy. AMD, IMO are themselves now in the position they put Intel in -
playing catch-up with red faces but probably good for consumers & AMD in the
long run. Just how much s/w consumers are likely to run is really going to
benefit from 4 cores?

A=Core2;B=K8;

NNN

 

[George Macdonald]

Thus spake Mark:

And people like me who haven't upgraded for years - I bought an XP-M2500 &
run it reliably @ 200x11.5x to replace an XP1800+ 18 months ago. The extra
speed was underwhelming.

What did you expect out of 300Mhz or so difference in core clock?

As for quad cores, many have suggested that's just
AMD's panicked response to having been left with their pants down rather
than a serious technical solution.

That's just rabble-babble. It's been obvious for a while that the
Pentium-M core was Intel's "development" platform for new cores... and that
they'd leapfrog past AMD by some smallish but significant amount... if they
re-focussed there and used the 65nm transistor budget wisely.

There's no "panicked" mode at AMD - quad core is just another part of their
65nm strategy which they are obviously working hard on; how well it might
do on the performance desktop is all down to just how much use game-makers
can make of multi-cores... something we don't know yet. It certainly has a
place in servers where Intel is apparently scrambling out another twin-dual
core compromise to compete.

My hope is that AMD will never be so
complacent again - though I sure a few pundits were also surprised that
Intel actually managed to produce something so good! As for gaining an extra
turn of speed, newer (read faster) h/ds can often give more than a
incremental CPU u/g does.

Assuming complacency on AMD's part is to miss what's been happening in the
process technology and its relationship with micro-architecture. AMD has
always been a few months behind Intel on process "node" - it's only been
18months since AMD got 90nm working properly and the expectation is that
they'll have 65nm at the turn of the year. I don't expect there to be much
of a leap-frog in general processing, unless they get some clock speed ramp
which nobody outside AMD knows about yet. We may get some new (updated ?),
more realistic benchmarks which highlight Intel's "cheat " on latency
metrics... to reinforce AMD's claims for the on-die memory controller.
Somebody might even write a benchmark which causes memory disambiguation to
crap all over itself.;-)

 

[George Macdonald]

Thus spake daytripper:

If A is better than B, n.A will always be better than n.B!

Where do you get the idea that Intel's A is allowed to make progress vs.
previous processors and AMD's B is not?

 

[chrisv]

After reading many articles on the new Dual 2 Core, it's obvious that
this could be a major reason to uprade now if you really want the
greatest/fastest computer on the block. Unfortunately AMD Is announcing
a Quad chip too claiming it will be even faster. Personally my
Pentium 4 2.4 ghz with DSL Is plenty fast for everything I do and will
not upgrade at least till prices drop drastically till prices are
around $150.00 for MB And CPU.
So who will buy these new chips or systems with this new technology?
Gamers ...serious gamers or people into serious video editing needs.

We upgrade... "because it's there" 8)

Seriously, I upgrade about every 2.5 years, for a variety of reasons.
That seems to be about how long it takes for (affordably-priced) new
hardware to advance significantly over what I'd built before (with the
exception of video cards, which move along somewhat faster).

This rate of upgrading has worked pretty well in my household, where
my daughter gets the hand-me-down PC whenever I upgrade. Her computer
is then 5 years old at worst, which means she always has a "modern"
computer, if never a "state of the art" computer".

 

[Rthoreau]

If A already saturates the bus with memory access, nA will be
bandwidth-starved.
If B scales linearly (OK, almost linearly), nB will likely beat nA.

A=Core2;B=K8;

NNN

That's hilarious almost fell off my chair reading that, reminds me of
all those geeky math jokes!

Does this mean I will get more frags in my favorite first person
shooter! I am still waiting for software to use my Opty 165. Although
playing quake4 on linux comes pretty close, also utk4 also is fun. It's
not like it's going to improve those games much. I would probably gain
more by a good mouse, or more practice.

Rthoreau

 

[Paul B]

Thus spake George Macdonald:

Where do you get the idea that Intel's A is allowed to make progress
vs. previous processors and AMD's B is not?

I don't! AMD has had the upper hand technically for some time & the game of
leaping frogs will surely continue. Had I made such assertions a couple of
months ago, I would have probably been accused of prematurely wishful &
partisan
thinking.

Does AMD need such defending here? My analogy with the equation was
simplistic but was it that far off the mark? Sorry, I just get the
impression that some far more knowledgeable folks on this forum than I'll
ever be, almost resent Intel getting something right for a change.

 

[George Macdonald]

Thus spake George Macdonald:

I don't! AMD has had the upper hand technically for some time & the game of
leaping frogs will surely continue. Had I made such assertions a couple of
months ago, I would have probably been accused of prematurely wishful &
partisan
thinking.

You presume too much. A couple of months ago, was well after the much
publicised IDF/Anand umm, scoop.

Does AMD need such defending here? My analogy with the equation was
simplistic but was it that far off the mark? Sorry, I just get the
impression that some far more knowledgeable folks on this forum than I'll
ever be, almost resent Intel getting something right for a change.

It was not a defense but a statement based on some known facts and
reasonable speculation, which you seem to have missed based on your n.A vs.
n.B canard... and which I mentioned briefly in another post in this thread:
with 65nm the transistor budget allowed Intel to add some extra paths and
features to their CPU design; they also had to add 64-bit.

The AMD quad-core will also be 65nm and will also have new features based
on the the same increased die real estate for the basic core; your
statement seems to presume that the AMD's B will not gain over their
current offerings and therefore will still lag on the fundamental
microarchitecture. I'd also note that AMD will be enhancing, rather than
adding, what has been a native 64-bit implementation for >3 years now; AMD
has also had a better handle on thermal management and a highly successful
true dual-core die at the current 90nm. While it's always possible that
they could stub their toe, AMD is starting from a better position on the
65nm transition.

As for resentment against Intel, there are enough Intel shills around in
other fora to cancel that one out. Between them, the Web site benchmarking
ingenues/idiots and Intel's unscupulous marketing practices there's bound
to be some insistence on reality from people who "know". The fact is that
AMD has had a better product for nearly 3 years and in that time, with
Microsoft's help has dragged Intel out of their Itanium fantasy dream. I
believe that, on the record, they deserve a better return for their
efforts.

 

[SamuelF566]

We upgrade... "because it's there" 8)

Seriously, I upgrade about every 2.5 years, for a variety of reasons.
That seems to be about how long it takes for (affordably-priced) new
hardware to advance significantly over what I'd built before (with the
exception of video cards, which move along somewhat faster).

This rate of upgrading has worked pretty well in my household, where
my daughter gets the hand-me-down PC whenever I upgrade. Her computer
is then 5 years old at worst, which means she always has a "modern"
computer, if never a "state of the art" computer".

I would upgrade every 2.5 years if it did not mean to get new memory,
video card and MB. upgrading to dual core means everything goes, even
the AGP Video card.
If you could just drop in a new CPU ...That would be great.

 

[chrisv]

I would upgrade every 2.5 years if it did not mean to get new memory,
video card and MB. upgrading to dual core means everything goes, even
the AGP Video card.
If you could just drop in a new CPU ...That would be great.

Well, if you could just "drop in a new CPU" it's very likely that you
would be disappointed with the result. In my experience, unless the
upgrade gives an honest-to-gosh doubling of computing power, it's
difficult to even notice, much less a significant improvement to your
life.

Do dual-core CPU's have double the computing power of single-core
CPU's? It must be so, but I'm still not sure how well it pays-off in
the real world...

 

[Nate Edel]

Do dual-core CPU's have double the computing power of single-core
CPU's? It must be so, but I'm still not sure how well it pays-off in
the real world...

Depends entirely on your workload. For some things (certain video encoding
tasks, some software development tasks) the speedup can be pretty well
linear; for other things, there's zero speedup.

They (or pre-dual core, dual-socket systems as well) also tend to be more
responsive subjectively - they *seem* faster, and "freeze" (or slow down in
noticeable UI latency) less often even if the actual process completion time
is not much shorter.

For general office tasks (where a P3 1.2ghz/P4 1.8ghz is probably just fine
in terms of absolute speed) moving to dual core is a no brainer as a result;
a bottom of the line P-D 805 or A64 X2 3600+ is going to FEEL much faster to
most users than a top-of-the-line 3.8ghz Xeon or 2.8ghz single core Opteron.