Quad Core Intel Q6600 - upcoming G0 stepping can run at much higher temps and.............

.... is thus highly likely to overclock a lot better with the same
cooling system as the current stepping of the processor..... !!!!
================================================== =======

From Tom's Hardware:-

Santa Clara (CA) : Intel has informed its customers that the Core 2
Quad Q6600 desktop quad-core processor as well as the Xeon X3220 and
X3210 server quad-core CPUs have been transitioned to a new stepping,
which will allow customers to integrate the processors in computer
systems using lower power flexible motherboards (FMB).

According to a product change notification released on Friday, the
three processors were transitioned from the B-3 to the G-0 processor
stepping and can now run not only on 105 watt performance FMBs, but on
95 watt mainstream FMBs as well.

The new CPUs, which can be identified by the CPUID "06FB" (B-3
processors were labeled "06F7"), also increase the Tcase, a value that
describes the maximum temperature a CPU can sustain, by 11 degrees
Celsius. According to Intel, this change will allow system builders to
scale back cooling requirements and achieve lower system noise levels
overall.

All three G-0 CPUs require BIOS updates. Samples of the processors are
expected to be available on May 11. General availability of the units
is scheduled for July 16.

====================================

Seems as if one should plan purchase of a quad-core desktop after the
current stepping has completely worked through the distribution/retail
chain.. Also Intel is planning very sharp price reductions of all
Core2 desktop processors in Q3... the price of the Q6600 is then
expected to fall to around $275. In the transition between the B3 and
G0 stepping, the wise purchaser will be closely looking for the new
sSpec code(s) which is usually printed somewhere amongst the processor
ID information on the box. The sSpec code is also always printed on
the processor heat-spreader.

Find the Q6600 G0 stepping sSpec code(s) here:-

http://processorfinder.intel.com/Default.aspx

when the part is formally available.

The only stepping code currently listed for the Q6600 is indeed "B3"
for which the sSpec is SL9UM.

Seems as if Intel is genuinely pushing all the right performance and
cost buttons to make the launch of the quad desktop version of
Barcelona as hard as possible for AMD. I would not be at all surprised
if Intel started shippng the quad desktop Penryn CPU by the beginning
of Q4 (2007). Especially if Barcelona performance really exceeds that
of the Core 2 quads. The best way to attack AMD financially is to deny
them every opportunity of charging premium prices for performance,
thus stifling their ability to quickly recover their development
costs.

John Lewis

Instead of just running faster, why won't they make
these systems use more ram? They could make
each core use its own 2 gig bank of ram. And that
could be made to run much cooler again. Seems
just dumb to bottleneck a quad processor with a
single bank of ram ... and only 2 gig at that.
There's no way a quad processor system has any
advantage without it.

johns

johns <johns321@moscow.com> writes:

> There's no way a quad processor system has any
> advantage without it.

Err, right Johns. Thanks for that stunningly crap precis. You *are*
joking? You do know about on board cache and so forth? Right?

On Sun, 29 Apr 2007 12:36:05 +0200, Walter Mitty <mitticus@gmail.com>
wrote:

>johns <johns321@moscow.com> writes:
>
>> There's no way a quad processor system has any
>> advantage without it.
>
>Err, right Johns. Thanks for that stunningly crap precis. You *are*
>joking? You do know about on board cache and so forth? Right?
>


Methinks that "johns" only understanding about the term 'cache' is the
under-bed one where he keeps his savings.

John Lewis

"Walter Mitty" <mitticus@gmail.com> wrote in message
news:87647fa2ne.fsf@gmail.com...
| johns <johns321@moscow.com> writes:
|
| > There's no way a quad processor system has any
| > advantage without it.
|
| Err, right Johns. Thanks for that stunningly crap precis. You *are*
| joking? You do know about on board cache and so forth? Right?
|

I'd like to hope he's joking, but sadly that's probably not the case... I'm
thinking he's just a little misguided.

It's always a bit amusing when people who obviously have absolutely no clue
how and why hardware works decide to pipe up with their views, and they
always seem to word them like they are just absolute truths...

Ah well, it's Usenet, what can ya do?

CoinSpin

CoinSpin wrote:
> "Walter Mitty" <mitticus@gmail.com> wrote in message
> news:87647fa2ne.fsf@gmail.com...
> | johns <johns321@moscow.com> writes:
> |
> | > There's no way a quad processor system has any
> | > advantage without it.
> |
> | Err, right Johns. Thanks for that stunningly crap precis. You *are*
> | joking? You do know about on board cache and so forth? Right?
> |
>
> I'd like to hope he's joking, but sadly that's probably not the case... I'm
> thinking he's just a little misguided.
>
> It's always a bit amusing when people who obviously have absolutely no clue
> how and why hardware works decide to pipe up with their views, and they
> always seem to word them like they are just absolute truths...
>
> Ah well, it's Usenet, what can ya do?
>
> CoinSpin
>

"... he's just a little misguided." the largest understatement since the
Pope was described as wearing a slightly different hat.

You forget the past ... NT4.0 and 1000 machines running
parallel calcs over shared connections ????? Where do
you guys constantly get off pretending to know jack about
the time of day ???? Where the hell have you been all
these years, that you don't seem to grasp even the
simplest concepts. Why do you think they are putting
4 processors on the same mobo to tackle the same
job ??? Holy Cat! It is parallel processing, and that
ain't done in cache. It's a running program you pack
of morons. It uses big stacks, and pulls little pieces
to cache. It is a ram starved operation on any single
machine, and that is why it worked 15 years ago
in NT 4.0 . I can share 4 machines in my office
and do the same thing ... only better because each
box has 2 gigs of ram. Oh yeah! For those of you
who suddenly think you can use the "they ain't
connected fast enough" bull****, You are wrong
again. How do you think a CRAY works.

johns

I helped invent the entire industry. You guys are
a joke. You don't have a clue about this level of
hardware. Don't worry. I'll keep trying to help
you learn.

johns ( a chip level gamer )

On 30 Apr 2007 00:00:36 -0700, johns <johns321@moscow.com> wrote:

>I helped invent the entire industry. You guys are
>a joke. You don't have a clue about this level of
>hardware. Don't worry. I'll keep trying to help
>you learn.
>
>johns ( a chip level gamer )
>
Er, buffalo chips ??

John Lewis

On 29 Apr 2007 23:55:07 -0700, johns <johns321@moscow.com> wrote:

>You forget the past ... NT4.0 and 1000 machines running
>parallel calcs over shared connections ????? Where do
>you guys constantly get off pretending to know jack about
>the time of day ???? Where the hell have you been all
>these years, that you don't seem to grasp even the
>simplest concepts. Why do you think they are putting
>4 processors on the same mobo to tackle the same
>job ??? Holy Cat! It is parallel processing, and that
>ain't done in cache. It's a running program you pack
>of morons. It uses big stacks, and pulls little pieces
>to cache. It is a ram starved operation on any single
>machine, and that is why it worked 15 years ago
>in NT 4.0 . I can share 4 machines in my office
>and do the same thing ... only better because each
>box has 2 gigs of ram. Oh yeah! For those of you
>who suddenly think you can use the "they ain't
>connected fast enough" bull****, You are wrong
>again. How do you think a CRAY works.
>
>johns
>

Er, DVD encoding on Premiere Pro runs 80% faster on my dual-core
machine with both processor-cores enabled than with one. One of
several metrics that I performed when I built my dual-core system.
Must complain to Adobe about their too-efficient use of the dual 1Meg
caches on the 939-pin 4400+.

BTW, I shall be building a quad-core machine some time in the Fall of
this year for both professional and gaming applications ( the quad
should speed up the compute-intense protions of Premier Pro by about
another 50% over the dual-core process-times). Nice for my Pro video
work, where shorter times represent real money savings... also certain
non-real-time functions suddently become realtime with the addition of
more parallel processing power and the associated large L2 caches.
Johns, would you like some benchmark results from programs that are
fully multicore-aware? I have quite a selection of such programs. No
doubt by that time Crysis will have joined this mix of programs... not
that "johns" will be interested in any multicore testing of Crysis,
since according to him this particular game does not exist. Er, does
"johns' really live on the same planet as the rest of us?

John Lewis