Conroe vs. AM2: Memory & Performance @ Anand

[brucedeluxe169]

http://www.anandtech.com/memory/showdoc.aspx?i=2800&p=1

From the article:

Probably the hardest conclusion for many will be the fact that increasing memory speed, increasing clock speed, and increasing CPU speed alone will not be enough for AM2 to catch up to Core 2 Duo in performance. The performance gap that remains when overclocking AM2 to 2.93GHz at 266 clock speed with DDR2-1067 is still huge. A die-shrink from 90 to 65nm and the additional cache that will allow will definitely help, but we are even skeptical there with Core 2 Duo already overclocking to 4GHz and beyond. No doubt AMD will find a solution, but it is now clear this will not be an easy fix for AMD.

Article does a great job of comparing memory performance on the AM2 and Conroe boards. Seems that in the short term, Intel's lack of an integrated mem controller is no real loss, but in the future, unless they have something else up their sleeves, they might have to integrate one.

 

[Donnie27]

Article does a great job of comparing memory performance on the AM2 and Conroe boards. Seems that in the short term, Intel's lack of an integrated mem controller is no real loss, but in the future, unless they have something else up their sleeves, they might have to integrate one.

We've been trying to tell AMD followers that since Yonah launched.

 

[brucedeluxe169]

yea, but they have their heads in the sand, they will for a long long time, I now know that

 

[GoldenTiger]

Frankly the gap that remains in Super Pi results when comparing AM2 and Core 2 Duo at the same speed was something of a shock. Clock for clock, with all other variables the same, Core 2 Duo is still almost 60% faster than AM2. There is nothing complex about calculating the value of Pi to 2 million places, but it does show the true power of Conroe in computation-intensive tasks.

and

We can now say with authority that Core 2 Duo is the faster performer clock-for-clock across the board. At the same 2.93GHz Far Cry is 27.7% faster, Half-Life 2: Lost Coast is 12.4% faster, and Quake 4 is 22.2% faster on Core 2 Duo. Of course AMD does not currently have a 2.93GHz CPU, so we tested by overclocking FX62. This suggests that FX64, or whatever it will be called, will not help much at 3.0GHz with a 200 clock speed.

Wow!!!!!

 

[brucedeluxe169]

i *know* its all just blind speculation and all, and its a far away battle, but am i the only one that is getting the feeling that intel will not have a problem handling the K8L?

People always say that the other company's next gen part always beats the current gen's top, but are we to assume that intel and amd will just continue to sprint out in front of each other to only let up and wait for their competitor to catch and eventually overtake them?

the way I see it, Intel has taken off, and is gonna try its hardest to not look back.

 

[Donnie27]

yea, but they have their heads in the sand, they will for a long long time, I now know that

Yup!

 

[savantu]

What will they fight K8L with ?

If K8L arrives in the summer , Intel will have the following :

Conroe DC
-3.33/1333MHz or 3.46GHz/1066MHz for XE ( that's conservative )
-3.2GHz E6900 ( again conservative since this is/was a Q1 2007 part )
Kentsfield QC
-2.66GHz for XE
-2.4GHz normal

If K8L arrives in oct-nov , Intel will have the following:
-The previous parts +
-Conroe 2 or Penryn is close to release (nov-dec 2007 or january 2008 )
-45nm 6MB L2 1333FSB
-3.33/1333MHz or 3.46/1066MHz top desktop
-3.66/1333MHz XE

If something really really bad happens with K8L and it slips into 2008 it will face Penryn until the summer when Nehalem appears.
What do we know about Nehalem ?

Not much , brand new microarch ( completly different from Conroe IIRC ) , IMC ( dual channel DDR3 ) , CSI with at least 9.6GBs I/O BW , 4 cores .Most likely it will have a shared cache ( 6-12MB ) and special purpose accelerators in it.

Even if by some miracle K8L recaptures the performance crown between summer and Penryn launch , I doubt it will hold for more then 4 months.

 

[brucedeluxe169]

What will they fight K8L with ?

If K8L arrives in the summer , Intel will have the following :

Conroe DC
-3.33/1333MHz or 3.46GHz/1066MHz for XE ( that's conservative )
-3.2GHz E6900 ( again conservative since this is/was a Q1 2007 part )
Kentsfield QC
-2.66GHz for XE
-2.4GHz normal

If K8L arrives in oct-nov , Intel will have the following:
-The previous parts +
-Conroe 2 or Penryn is close to release (nov-dec 2007 or january 2008 )
-45nm 6MB L2 1333FSB
-3.33/1333MHz or 3.46/1066MHz top desktop
-3.66/1333MHz XE

If something really really bad happens with K8L and it slips into 2008 it will face Penryn until the summer when Nehalem appears.
What do we know about Nehalem ?

Not much , brand new microarch ( completly different from Conroe IIRC ) , IMC ( dual channel DDR3 ) , CSI with at least 9.6GBs I/O BW , 4 cores .Most likely it will have a shared cache ( 6-12MB ) and special purpose accelerators in it.

Even if by some miracle K8L recaptures the performance crown between summer and Penryn launch , I doubt it will hold for more then 4 months.

yeeeouch! I never looked at it that way! wow, thats alot of pwnage coming from the blue camp.

those 17 dollar intel shares are looking mightly tempting....

 

[chrisf6969]

What will they fight K8L with ?

If K8L arrives in the summer , Intel will have the following :

Conroe DC
-3.33/1333MHz or 3.46GHz/1066MHz for XE ( that's conservative )
-3.2GHz E6900 ( again conservative since this is/was a Q1 2007 part )
Kentsfield QC
-2.66GHz for XE
-2.4GHz normal

If K8L arrives in oct-nov , Intel will have the following:
-The previous parts +
-Conroe 2 or Penryn is close to release (nov-dec 2007 or january 2008 )
-45nm 6MB L2 1333FSB
-3.33/1333MHz or 3.46/1066MHz top desktop
-3.66/1333MHz XE

If something really really bad happens with K8L and it slips into 2008 it will face Penryn until the summer when Nehalem appears.
What do we know about Nehalem ?

Not much , brand new microarch ( completly different from Conroe IIRC ) , IMC ( dual channel DDR3 ) , CSI with at least 9.6GBs I/O BW , 4 cores .Most likely it will have a shared cache ( 6-12MB ) and special purpose accelerators in it.

Even if by some miracle K8L recaptures the performance crown between summer and Penryn launch , I doubt it will hold for more then 4 months.

Thats the big one (45nm). Intel has always used it larger size, to master the process shrinks to win every battle before. AMD's biggest problem is that Intel is back on the normal track, with a similar architecture (Netburst was just radically different) that as long as they're ahead on the the processes it will be NEARLY IMPOSSIBLE for AMD to catch up, much less surpass it.

 

[savantu]

yeeeouch! I never looked at it that way! wow, thats alot of pwnage coming from the blue camp.

That is the general picture ( which few AMD fans seem to notice ) , things might end up a little different , but not by much ( most likely in a positive way ).It all comes down to execution , and after all the flaws in 2004-2006 period I expect Intel to do it right this time.

 

[aldamon]

yea, but they have their heads in the sand, they will for a long long time, I now know that

AMD's CPU speeds have been so stagnant for so long that I think what you're seeing is just complacancy. We saw this before in the final days of the AthlonXP. The chips got ridiculously cheap, and it was an easy upgrade, so people (like me) downplayed Intel's performance advantage.

 

[Donnie27]

yeeeouch! I never looked at it that way! wow, thats alot of pwnage coming from the blue camp.

those 17 dollar intel shares are looking mightly tempting....

Awee but that's the catch-22. Intel's stock's value isn't limited to just processors like AMD's. It's important but not the only reason. Intel doesn't need mass layoffs to fix the few problems they do have but they have to trim some Fat or overhead.They have to dump some money loosin' Depts. They just laid off a 1000.

 

[brucedeluxe169]

Awee but that's the catch-22. Intel's stock's value isn't limited to just processors like AMD's. It's important but not the only reason. Intel doesn't need mass layoffs to fix the few problems they do have but they have to trim some Fat or overhead.They have to dump some money loosin' Depts. They just laid off a 1000.

laid off 1000 managers, which is alot of fat....

not to mention their tech leadership will not disappear for another 3 years at the minimum, and PC demand will grow after the next quarter (back to school, holidays, vista early next year).

Intel stock is at a very good point right now i'd say.

 

[Donnie27]

laid off 1000 managers, which is alot of fat....

not to mention their tech leadership will not disappear for another 3 years at the minimum, and PC demand will grow after the next quarter (back to school, holidays, vista early next year).

Intel stock is at a very good point right now i'd say.

I agree and I should have said managers.

 

[duby229]

I agree and I should have said managers.

Unfortunatly, those 1000 managers only account for 0.03% In the short term laying them off does very little.

I'm glad to see you guys gloating over these benchmarks tho... It'll just make things even better when you discover it doesnt actually work that way in the real world...

 

[Poncho]

What will they fight K8L with ?

If K8L arrives in the summer , Intel will have the following :

Conroe DC
-3.33/1333MHz or 3.46GHz/1066MHz for XE ( that's conservative )
-3.2GHz E6900 ( again conservative since this is/was a Q1 2007 part )
Kentsfield QC
-2.66GHz for XE
-2.4GHz normal

If K8L arrives in oct-nov , Intel will have the following:
-The previous parts +
-Conroe 2 or Penryn is close to release (nov-dec 2007 or january 2008 )
-45nm 6MB L2 1333FSB
-3.33/1333MHz or 3.46/1066MHz top desktop
-3.66/1333MHz XE

If something really really bad happens with K8L and it slips into 2008 it will face Penryn until the summer when Nehalem appears.
What do we know about Nehalem ?

Not much , brand new microarch ( completly different from Conroe IIRC ) , IMC ( dual channel DDR3 ) , CSI with at least 9.6GBs I/O BW , 4 cores .Most likely it will have a shared cache ( 6-12MB ) and special purpose accelerators in it.

Even if by some miracle K8L recaptures the performance crown between summer and Penryn launch , I doubt it will hold for more then 4 months.

This man right here knows what's up. Can't see a single thing innacurate with that post. Well done!

 

[MrGuvernment]

Unfortunatly, those 1000 managers only account for 0.03% In the short term laying them off does very little.

I'm glad to see you guys gloating over these benchmarks tho... It'll just make things even better when you discover it doesnt actually work that way in the real world...

It does when you use your system for more then just gaming, like more people then people on here seem to think do.

 

[Poncho]

I'm glad to see you guys gloating over these benchmarks tho... It'll just make things even better when you discover it doesnt actually work that way in the real world...

Kinda like how AMDs fans gloated over their "non real world" benchmarks and laughed when people told them that they didn't matter much in real world performance? LOL...


Ahhhh.... gotta love the short term memory.

 

[savantu]

Unfortunatly, those 1000 managers only account for 0.03% In the short term laying them off does very little.

I'm glad to see you guys gloating over these benchmarks tho... It'll just make things even better when you discover it doesnt actually work that way in the real world...

I didn't know that Intel has 3,333,333 managers .Damn , if for every manager they have 2 workers it means Intel has 10 million employees . Woot!!

Oh , wait ! It's the DUBY-MATH , math of the future now !

 

[Donnie27]

Unfortunatly, those 1000 managers only account for 0.03% In the short term laying them off does very little.

I'm glad to see you guys gloating over these benchmarks tho... It'll just make things even better when you discover it doesnt actually work that way in the real world...

Prove it with a Math Equation? 1000 of roughly 100,000 is 1% and since those 100,000 aren't all managers, that percentage is useless. Those 100 managers not getting paid is more meaningful for than laying off 8 to 12 thousand lower level employees.

Still holding on to hope for your beloved AMD uh? Dude I feel sorry for ya'.

 

[brucedeluxe169]

dude, you really dont think there is a difference in the real world for conroe?

 

[Donnie27]

dude, you really dont think there is a difference in the real world for conroe?

He believes what he's saying!

 

[chrisf6969]

managers often make 50% - 100% more than the people below them, and sometimes have more stock options or other perks. so 1000 "managers" could mean some MAJOR cost savings.

Assuming they were glut. Now if the people below them are now not supervised as well and don't work as efficiently it could cost Intel some of the money they saved. But as with many large corp.s they were probably "top heavy" ie: over-managed.

 

[brucedeluxe169]

managers mean 50-100% bigger salaries.

bonuses on top of that

management perks... better computers, corporate spending account that they can whore at will

personal offices

many have company cars/travel allowances

etc
etc
etc

management jobs suck up tons of money.

good riddance

 

[Donnie27]

managers often make 50% - 100% more than the people below them, and sometimes have more stock options or other perks. so 1000 "managers" could mean some MAJOR cost savings.

Assuming they were glut. Now if the people below them are now not supervised as well and don't work as efficiently it could cost Intel some of the money they saved. But as with many large corp.s they were probably "top heavy" ie: over-managed.

Seems like "top heavy" is what some investors complained about. At least that's what one told me

 

[Poncho]

managers mean 50-100% bigger salaries.

bonuses on top of that

management perks... better computers, corporate spending account that they can whore at will

personal offices

many have company cars/travel allowances

etc
etc
etc

management jobs suck up tons of money.

good riddance

Not at intel it doesn't. This may apply to some companies and SOME high high level managers at Intel..... but it doesn't fit for most of the managers affected. Depending on the level of management they may not make any more money than the people who report to them. The biggest thing this action did was trim some fat and redundency.

 

[brucedeluxe169]

ahh wells, trimming fat is always a good thing I guess.....

hopefully they replace some of those with more engineers engineers rock

 

[nst6563]

managers mean 50-100% bigger salaries.

bonuses on top of that

management perks... better computers, corporate spending account that they can whore at will

personal offices

many have company cars/travel allowances

etc
etc
etc

management jobs suck up tons of money.

good riddance

doesn't work that way at AT&T either if you're looking at 3rd level and below. Depending on what department you're in, management makes LESS than the people that report to them...by a significant amount (sometimes in the $30,000 range).

3rd level and above...I whole heartedly agree

 

[jvrobert]

Not at intel it doesn't. This may apply to some companies and SOME high high level managers at Intel..... but it doesn't fit for most of the managers affected. Depending on the level of management they may not make any more money than the people who report to them. The biggest thing this action did was trim some fat and redundency.

Somewhat true for first level managers, but above that it amplifies. Your manager may make only say 10-15% more than you do in base salary, but once you add in the much larger bonuses and stock you're probably looking at more like 40-50%.

 

[forcefed]

I'm glad to see you guys gloating over these benchmarks tho... It'll just make things even better when you discover it doesnt actually work that way in the real world...

You've been watching too much of the matrix...

The conroe smokes anything and everything from amd in apps and games... assuming theres no bottleneck *cough* [H] conroe review *cough*.