Why do E6400/E6300s clock so well??

[thilanliyan]

Shouldn't they be binned for a lower speed grade?? It seems as though retail chips are hitting a wall around 3.5-3.6GHz even for the E6600/E6700s at least according to xtremesystems.org members. So if an E6600 takes the same amount of volts to get to 2.4GHz (stock) as an E6400 takes to get to 2.13GHz, shouldn't the E6600s get a lot higher when you pump up the volts or are they hitting the architecture limits??

How are they binned anyway?? By the voltage required??

 

[aldamon]

I could be wrong, but I think all current Conroes are being binned by how much good cache they have and not by speed. Cores with 4MB of cache become E6600+ and up and everything else becomes either an E6300 or E6400. Intel doesn't care how high they go. They can't be 6600s, 6700s, or 6800s without that cache.

 

[stevty2889]

The smaller cache on the 6300 and 6400 probably help them to reach the better speeds. They are manufactured with 2mb of cache, not with 4mb of cache with half of it disabled. Just like the X2's with 512kb of cache tend to OC a little better than the ones with 1mb of cache.

 

[shabby]

Doesnt seem like they're binned at all, intel hit the jackpot with these chip and almost all of them hit the same speeds.
Like aldamon said, looks like they only get binned by cache.

 

[lopri]

Was that theory verified yet? Whether E6300/E6400 are cache-disabled E6600/E6700?

 

[stevty2889]

Originally posted by: lopri
Was that theory verified yet? Whether E6300/E6400 are cache-disabled E6600/E6700?

It's not a theory, Allendale is manufactured with 2mb of cache, and I have seen it myself

 

[theteamaqua]

it has nothing to do with cache... for some reason 965 baords especially GA-965 series loves the 7x multiplier, i have seen E6300 do 3.5GHz on P5B Deluxe with tower 120, the higher the multi , the lower the fsb ... 7x multi is the key i believe, i tried that on my 6400 didnt work, i think 6300 actually OC better b/c of its 7x multi .. 8x multi doesnt work as well as 7x multi

thats why u will see a lot fo people who burn $1000, $500 on 6700 , 6800 and only get to 3.8GHz max on air with any board

and yet i spend $250 got to 3.4GHz , not bad for my money

 

[harpoon84]

I think the reason E6300 and E6400s are overclocking so well is because they are clocked WAY under the architecture ceiling. I would hazard a guess that if you picked any random C2D chip before they are binned, most would reach 3GHz on stock voltage.

 

[Madellga]

Originally posted by: stevty2889
The smaller cache on the 6300 and 6400 probably help them to reach the better speeds. They are manufactured with 2mb of cache, not with 4mb of cache with half of it disabled. Just like the X2's with 512kb of cache tend to OC a little better than the ones with 1mb of cache.

I've seen this before....Pentium 2 450 x Celeron 300A...early Anand's article, I jumped on the Celeron 300A - it would also o/c 50%.

The 300A had half of the 450 cache.

 

[theteamaqua]

ha for those who got the week 24 which is the batch #L624 = year 06 , week 24 on the lable of core 2 duo, u all gonna get lower clock ...

week 25 E6400 can easily reach 3.6GHz, mine stopped after 425fsb 3.4GHz

i have week 24 by the way ... it sux...

but yeah i think the silicon they use to make conroe can perform at the same frequency as pentium d's, maybe thats why, cuz after OCed, evry conroe is about the same, near that 3GHz to 3.8GHz area ...

PS: E6300 can go over 3GHz easily ... with good 965 baords like P5B Deluxe, GA-965-DS3

 

[Payner44]

theteamaqua, I just made a new PC with the P5B Deluxe and E6400...also have 2G of 667 RAM. Can you offer any advise on what settings to use to get this thing to the 3Ghx mark? I'm new to overclocking...thanks!

 

[gobucks]

Originally posted by: stevty2889
The smaller cache on the 6300 and 6400 probably help them to reach the better speeds. They are manufactured with 2mb of cache, not with 4mb of cache with half of it disabled. Just like the X2's with 512kb of cache tend to OC a little better than the ones with 1mb of cache.

isn't the die size on an allendale the same as on a conroe? That would indicate that half the cache is indeed disabled.

anyways, as for the binning issue, i think there is some degree of it, since X6800s seem to clock better than the others. still, it is pretty obvious that a typical OC for any model is north of 3GHz. the thing is, Intel's core 2 production is clearly going really well, and most chips are capable of greatly exceeding their required speeds. unfortunately for them, just because they CAN clock their chips at 3GHz, doesn't mean there is a market for them. So they have to sell a lot of perfectly fine chips as e6300s and e6400s, since that is where the majority of demand is.

the same exact thing happened with the 90nm athlon 64 chips. i remember my old 3000+ was capable of 2.55GHz, and most others were as well. Sure they could have sold them as 4000+ or higher, but no one would buy them unless they lowered prices substantially, and there is no point in that from their point of view.

 

[Hulk]

Originally posted by: Madellga

Originally posted by: stevty2889
The smaller cache on the 6300 and 6400 probably help them to reach the better speeds. They are manufactured with 2mb of cache, not with 4mb of cache with half of it disabled. Just like the X2's with 512kb of cache tend to OC a little better than the ones with 1mb of cache.

I've seen this before....Pentium 2 450 x Celeron 300A...early Anand's article, I jumped on the Celeron 300A - it would also o/c 50%.

The 300A had half of the 450 cache.

Actually that was a little different. The Celeron 300a had half the cache of the PIII 450 but the Celly had the full speed on-board cache whereas the PIII 450 had a larger cache running at half the core speed. The Celly was faster in quite a few apps.