Any direct comparisions of Conroe 4MB cache vs. 2MB cache chips?

[Hulk]

I'm curious as to how much difference the extra 2MB cache makes in most applications/games. Anyone know of any links that show a direct comparison? I'm hoping someone has lowered the multiplier on the E6600 to the setting of the E6400 or E6300.

 

[lopri]

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2795&p=4

The impact of L2 cache size is bigger with C2Ds than with A64s. For E6300/E6400 to be equal to 3.6GHz E6600/E6700, the former will need frequencies up to 3.8~4.0GHz. (Yes, even after considering FSB. Besides, 450+ FSB is a huge stress on the board even for P965)

 

[Hulk]

Don't know how I missed that!

Thanks

 

[Pabster]

Yep. Although some will insist the higher FSB of their Allendale is enough to compensate ... it isn't. A 6300/6400 needs 300-400MHz more clock to do comparable figures to a 6600/6700/6800 at a lower speed, regardless of a high FSB.

Now then the real question is... does it make a difference? Depends on what you're doing with the rig.

 

[Baked]

From that chart, 2MB only gives you 3.5% avg. performance boost. But whether you're gonna buy the 2MB Allendale or the 4MB Conroe really comes down to what programs you're gonna run.

 

[n7]

Originally posted by: Baked
From that chart, 2MB only gives you 3.5% avg. performance boost. But whether you're gonna buy the 2MB Allendale or the 4MB Conroe really comes down to what programs you're gonna run.

That or the price you want to pay.

After seeing how the E6600s didn't clock any better than the E6400s & how they cost over $100 CND more, i was happy to get an E6400, & thus far, i'm not unhappy with my decision.

 

[harpoon84]

Hmm...

So 4MB vs 2MB L2 = 3.5% difference.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2795&p=6
1333MHz FSB vs 1066MHz FSB = 2.4% difference on average.

1333 vs 1066 = 25% difference

E6600 9x multi vs E6300 7x multi = 28.6% FSB difference when equally clocked. The ratio is greater than the 25% AT tested to reach the 2.4% difference.

Therefore, according to ATs results, an E6300 @ 3.6GHz would only be ~1% slower than an E6600 @ 3.6GHz, due to the faster FSB compensating somewhat for the lack of cache.

This is purely a mathematical estimation of course, feel free to contradict my findings.