Nexgen Intel Chipset (Bearlake-X) details revealed

Grimmy

Splendid
Feb 20, 2006
4,431
0
22,780
How do they come up with these names. :lol:

I remember they were running out of names, like for hurricanes.

Guess there's plenty of names left for system upgrade parts.
 

mpjesse

Splendid
I think the end of "low latency" is at an end... for now. I'm curious to know what the latencies are for the new Radeon x1950xtx 1GB DDR4 memory. They have to be really high. I mean... 1Ghz?! (no DDR bullsh*it either). True 1Ghz clock speed. Amazing.

Regardless it doesn't really matter anymore. Pure bandwidth seems to be the priority. And don't forget that RDRAM had horrible latencies, yet it still managed to destroy DDR226 @ CAS 2.5 back in the day.

And speaking of which, didn't RDRAM hit 1Ghz first?
 

1Tanker

Splendid
Apr 28, 2006
4,645
1
22,780
This is probably the best era I've experinced since Barton and Coppermine.
It's back to OC by high 1FSB:1RAM again just like in the old days. :D

Just only a month ago before C2D was released I had doubt whether DDR2-1066 would ever reach it's full potential, but now with Allendales clocking beyond 535Mhz FSB we need even more from RAM just to run 1:1.

So as long as they keep the latency of DDR3 under control and keeping going beyond 533Mhz it'll be a welcoming tech this time compared to DDR2 when it was introduced.

Now I can't wait for Bearlake-X (other than it's price if nVidia isn't able to keep up).
Also, notice that we're back to lower multis like in PIII days....7x,8x. None of these 18x-24x multis. :wink:
 

1Tanker

Splendid
Apr 28, 2006
4,645
1
22,780
How do they come up with these names. :lol:

I remember they were running out of names, like for hurricanes.

Guess there's plenty of names left for system upgrade parts.

Intel uses geographical names of locations near the sites where the parts are developed. Bearlake is a place in Oregon I suspect (without going to an atlas). The reason they do this is because there is no fear of trademark violations on an unbranded product as a company or individual cannot register a trademark for a name of a location.

AMD uses names of location as well for lesser publicized projects, however they also use common names for objects --- usually derogatorily aimed at Intel --- such as Sledgehammer, Hammer, or the most recent one indicating 'domination', Bulldozer.

JackNobody's ever been able to figure out where Sempron came from though. Still sounds weird. :wink:
 

mpjesse

Splendid
That means New Mexico landmarks are next. Intel has a large 65nm FAB here in Albuquerque.

So don't be surprised when you start seeing codenames like "Tamale" or "Taco."

Just kidding. More likely are names like Monzano, San Mateo, or Sandia.
 

1Tanker

Splendid
Apr 28, 2006
4,645
1
22,780
How do they come up with these names. :lol:

I remember they were running out of names, like for hurricanes.

Guess there's plenty of names left for system upgrade parts.

Intel uses geographical names of locations near the sites where the parts are developed. Bearlake is a place in Oregon I suspect (without going to an atlas). The reason they do this is because there is no fear of trademark violations on an unbranded product as a company or individual cannot register a trademark for a name of a location.

AMD uses names of location as well for lesser publicized projects, however they also use common names for objects --- usually derogatorily aimed at Intel --- such as Sledgehammer, Hammer, or the most recent one indicating 'domination', Bulldozer.

JackNobody's ever been able to figure out where Sempron came from though. Still sounds weird. :wink:

Sempron is a branded name, thus trademarked by AMD.

The name is derived from the Latin word, Semper (meaning "always"), and the tech suffix -ron (which apparently means "budget processor"). Sempron should fall at a lower performance point than equivalent rated parts from other AMD processor lines.
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2139I like the Tech Reports explanation of Sempron better. :D

Yep, that's the name: Sempron. AMD has elected to stick with its "fake subatomic particle" naming scheme rather than veer into Intel's "fake member of the periodic table of elements" naming scheme. Sempron is largely a branding exercise, so the name is important. The Sempron name is intended to evoke phrases like "semper fidelis" and other such tokens of solidity and steadfastness. Roughly translated from a mix of Latin and leet-speak, though, Sempron means "always pornographic," and I fear the little CPU will never fully escape that connotation of its recently fabricated moniker. As for why AMD decided not to reuse the Duron name, all I can say is that the intricacies of product branding strategies escape me.

hehehe

http://techreport.com/reviews/2004q3/sempron/index.x?pg=1
 

Nitro350Z

Distinguished
Apr 19, 2006
416
0
18,780
Regardless it doesn't really matter anymore. Pure bandwidth seems to be the priority. And don't forget that RDRAM had horrible latencies, yet it still managed to destroy DDR226 @ CAS 2.5 back in the day.

And speaking of which, didn't RDRAM hit 1Ghz first?

Yes, RDRAM did hit 1Ghz first, 1066mhz, was really expensive at the time, but like you said, it was good.

Can't wait to get my hands on DDR3 and this new chipset, I wonder if it will be a good overclocker.
 

1Tanker

Splendid
Apr 28, 2006
4,645
1
22,780
That means New Mexico landmarks are next. Intel has a large 65nm FAB here in Albuquerque.

So don't be surprised when you start seeing codenames like "Tamale" or "Taco."

Just kidding. More likely are names like Monzano, San Mateo, or Sandia.
TEQUILA
 

Nitro350Z

Distinguished
Apr 19, 2006
416
0
18,780
That means New Mexico landmarks are next. Intel has a large 65nm FAB here in Albuquerque.

So don't be surprised when you start seeing codenames like "Tamale" or "Taco."

Just kidding. More likely are names like Monzano, San Mateo, or Sandia.
TEQUILA

We can only hope. :D
 

m25

Distinguished
May 23, 2006
2,363
0
19,780
http://www.xbitlabs.com/news/chipsets/display/20060822231520.html

Codename: Bearlake-X
DDR3 support
5Ghz PCI-E 2.0 bus (compared to 2.5Ghz currently)
Support for 1333mhz FSB

Not revolutionary, but DDR3 support is nice to see.

Still the 1333 FSB ?!?! Even here I don't see anything revolutionary :roll:
What do they need the Dual, 1333MHz DDR3 if they only use a 1333MHz bus?!
 

Action_Man

Splendid
Jan 7, 2004
3,857
0
22,780
http://www.xbitlabs.com/news/chipsets/display/20060822231520.html

Codename: Bearlake-X
DDR3 support
5Ghz PCI-E 2.0 bus (compared to 2.5Ghz currently)
Support for 1333mhz FSB

Not revolutionary, but DDR3 support is nice to see.

Still the 1333 FSB ?!?! Even here I don't see anything revolutionary :roll:
What do they need the Dual, 1333MHz DDR3 if they only use a 1333MHz bus?!

:roll: Why does it have to be revolutionary?
 

Action_Man

Splendid
Jan 7, 2004
3,857
0
22,780
:roll: Bearlake is a chipset not a cpu dipsh!t. Learn to read gramps.

Intels IMC is due in 2008 with itanium first while bearlake is due next year. Conroe doesn't have an IMC and it wipes the floor with K8.
 

m25

Distinguished
May 23, 2006
2,363
0
19,780
:roll: Bearlake is a chipset not a cpu dipsh!t. Learn to read gramps.

Is there any difference ? It still will make the CPU work @ 1333MHz. They're going to base their future CPUs on a crappy 1333MHz bus,... well.


Conroe doesn't have an IMC and it wipes the floor with K8
That's why it will soon be wiped by K8L. It's architecture and 4X more chache won't help it anu more.
 

darkstar782

Distinguished
Dec 24, 2005
1,375
0
19,280
I'm curious to know what the latencies are for the new Radeon x1950xtx 1GB DDR4 memory. They have to be really high. I mean... 1Ghz?! (no DDR bullsh*it either). True 1Ghz clock speed. Amazing.

Why is DDR bullshit? Would it not be theoretically possible to construct a system whereby the clock was 10GHz but used to control a bus that only transferred 1MT/s, thereby having no more bandwidth than a DDR system with a 500MHz clock?

The speed of the clock generator is not what matters, its how many data transfers are performed per second, Hz is a measure of frequency, how many times something happens in a second, "My wifes menstrual cycle is on a 413nHz frequency". As such, is it not equally accurate, and far more relevent, to use a measure of the frequency of data transfers, meaning 2GHz in this example?
 

Action_Man

Splendid
Jan 7, 2004
3,857
0
22,780
Is there any difference ?

You need to ask that question!?

It still will make the CPU work @ 1333MHz. They're going to base their future CPUs on a crappy 1333MHz bus,... well.

Which is more then conroe needs and since the chipset is coming out next year and they wont be releasing anything new till the start of 2008, well I'll laugh at your attempts to connect the dots.

That's why it will soon be wiped by K8L.

You mean in 2008?
 

m25

Distinguished
May 23, 2006
2,363
0
19,780
Which is more then conroe needs ...

What about Kentsfield and future (but not so much) 8 cores? Will they all have to share a 1333 bus... 166 MHz/core :?: :!:

Don't tell around you're smart, please. :lol:
 

Nitro350Z

Distinguished
Apr 19, 2006
416
0
18,780
If they're going to come out with 8 cores, I think they're going to use a dual fsb, Can't remember where I heard that.

Also what is Intels csi exactly? I remember reading a few things about it but I cant remember. Supposedly going to be ready around 2008 right?

As for the PCI-E 2.0, it seems kinda pointless to speed up PCI-E since there arent even graphics cards that can saturate an 8x link, correct?
 

exit2dos

Distinguished
Jul 16, 2006
2,646
0
20,810
What is PCI-epress 2.0 anyways? And why do we need it?

It's going to increase to 5GBS. It also may have built-in virtualization

http://www.eetasia.com/ART_8800421062_499491_d828caf0200606.HTM

Graphics chips need the new 5Gbps speeds to drive higher performance as well as open the door to using the fast channels to eliminate graphics memory in favor of using the systems main memory—even when graphics are on a card off the motherboard.

In servers, both serial ATA and serial-attached SCSI standards are preparing a move up from 3Gbps to 6Gbps speeds that will want Express 2.0. In addition, multiport controllers for Ethernet, Infiniband and Fiber Channel will want the faster system link.