E6320 & E6420... :D

MeatLoaf · 11 Apr 2007 at 07:33

I assume these chips will work in my DS4? Maybe with a BIOS update?

Minstadave · 11 Apr 2007 at 08:06

MeatLoaf said:
I assume these chips will work in my DS4? Maybe with a BIOS update?

Yup no probs

MeatLoaf · 11 Apr 2007 at 08:17

Hmm, i was considering switching my E6300 for an E6400 next month as im hoping its higher multi will get me to 3.4 - 3.6Ghz

Will the E6420 have the same 8x multiplier?

Minstadave · 11 Apr 2007 at 08:19

MeatLoaf said:
Hmm, i was considering switching my E6300 for an E6400 next month as im hoping its higher multi will get me to 3.4 - 3.6Ghz

Will the E6420 have the same 8x multiplier?

It will

Participant · 11 Apr 2007 at 13:40

Journey - you are an idiot. Read Nathan's [useful] post. Conroes do not all have 4Mb cache.

Journey · 11 Apr 2007 at 14:26

Robbie G said:
Journey - you are an idiot. Read Nathan's [useful] post. Conroes do not all have 4Mb cache.

Personal insults are not welcome on these forums. If you missed what I read I was trying to highlight the simple difference, for those that have no use or interest in V.T.

Fair enough the 'Allendale' will have 4MB and the 'Conroe' will have 4MB so the differnce is VT or not VT. I cannot see why you feel the need to call someone an idiot over this.

Justintime · 11 Apr 2007 at 15:00

Why argue over some pieces of silicon that'll be obselete in months?

Big.Wayne · 11 Apr 2007 at 15:11

It looks like the new 6320/6420 are actually [b2] Conroe chips, i.e take an e6600 with good working 4Mb cache and just set the default multi lower (in this case x7 and x8).

INTEL already have an outlet for Conroe chips that are produced with faulty cache, they just disable half the cache and sell them as e6300/e6400, so it looks like they found a way to sell of Conroe chips that have been produced with a good 4Mb cache but maybe the core isn't up to running e6600/e6700 speeds, so they just drop the default multi and rebrand them as 6320/6420.

Am I right? or am I right?

Minstadave · 11 Apr 2007 at 15:15

Big.Wayne said:
It looks like the new 6320/6420 are actually [b2] Conroe chips, i.e take an e6600 with good working 4Mb cache and just set the default multi lower (in this case x7 and x8).

INTEL already have an outlet for Conroe chips that are produced with faulty cache, they just disable half the cache and sell them as e6300/e6400, so it looks like they found a way to sell of Conroe chips that have been produced with a good 4Mb cache but maybe the core isn't up to running e6600/e6700 speeds, so they just drop the default multi and rebrand them as 6320/6420.

Am I right? or am I right?

Depends on their yields, as the process improves they'll have plenty of chips that'll run E6600 speeds, so will end up taking good chips and dropping their multis and selling them as E6420s or E6320s or disabling half their cache and flogging them as E6300/E6400s, just so they can fill the demand. This is where us overclockers benefit

Big.Wayne · 11 Apr 2007 at 15:29

Minstadave said:
This is where us overclockers benefit

Ok thats a valid point but I'm not so sure about that. In theory what you are saying is that INTEL are taking a perfectly good e6600 (or e6700?) and hamstringing (<is that a real word?) them? Not impossible but armed with that knowledge why would anyone now buy an e6600?

I'm guessing the chips that are released as e6320/e6420 are somehow less 'pure' than full blown e6600/e6700's?

We will found out soon enough though

P.S: I still didn't quite get my head around the B2 vs L2 steppings yet and I'm wondering if INTEL are gonna phase-out the classic B2 steppings?

(L-R) B2 and L2

Minstadave · 11 Apr 2007 at 15:50

Big.Wayne said:
Ok thats a valid point but I'm not so sure about that. In theory what you are saying is that INTEL are taking a perfectly good e6600 (or e6700?) and hamstringing (<is that a real word?) them? Not impossible but armed with that knowledge why would anyone now buy an e6600?

This is why locked multipliers came out, in the good old Socket A athlon days nobody on here bought a barton 3200+ because a barton 2500+ would do 3200+ speeds by just upping the multi. It got to the point where dodgy people were buying 2500s, replacing the stickers with 3200+ stickers and flogging them on as 3200s.

Even locking the multi didn't help that much though because the 2500 and 3200 had the same 11x multi, it was just the FSB was different (200/166) so you could just stick the 2500 in a decent mobo and whack the FSB up to 200, instant 3200

The C2Duos are no different, when a lowly E6400 like mine can do 3.2ghz on stock volts, you know they must be getting good yields, so I'm sure there's a good chance decent 4MB cores are being binned to fill the demand for 2MB parts. Its just a shame no one has found a way to reactivate the cache.

Another example was the 6800 cores, loads of them were disable to run with just 8 or 12 pipes in 6800LEs or 6800NUs whilst still having perfectly fully functional 16 pipes.

Big.Wayne · 11 Apr 2007 at 17:59

Minstadave said:
The C2Duos are no different, when a lowly E6400 like mine can do 3.2ghz on stock volts, you know they must be getting good yields, so I'm sure there's a good chance decent 4MB cores are being binned to fill the demand for 2MB parts.

Yes I remember, I had an XP2500+ in Aug 2003 nice chip at XP3200+ speeds. On the whole I'm aware of the speed binning process but I don't just assume thats the case every time a chip manufacturer releases something new, also figure in the soon to be released e6650 and e6750, are these just gonna be overclocked e6320/e6420's (simply running 333MHz-FSB instead of 266MHz-FSB)or are the e6320/e6420 just underclocked e6650/e6750's?

e6320 (266x7= 1862MHz) 4MB
e6420 (266x8= 2128MHz) 4MB

e6650 (333x7= 2331Mhz) 4MB
e6750 (333x8= 2664MHz) 4MB

So take a new e6320, insert it in your mobo, boot up and change the FSB from 266MHz to 381MHz and you have an unreleased e6750 (actually a bit faster but lets not split hairs)

Only way to know is wait and see what people are getting Orthos stable with these new chips, as soon as I see 3.6GHz stable (on air) I'm in!

Participant · 11 Apr 2007 at 19:20

Journey said:
Personal insults are not welcome on these forums

Neither, from what I can gather, are arrogant misstatements...

Quote from you: "Why do people always feel the need to make things more difficult than they are"

In response I'd say - why do people feel the need to take an arsey tone when they are in fact totally wrong?

Participant · 11 Apr 2007 at 19:27

Journey said:
Personal insults are not welcome on these forums. If you missed what I read I was trying to highlight the simple difference, for those that have no use or interest in V.T.

Fair enough the 'Allendale' will have 4MB and the 'Conroe' will have 4MB so the differnce is VT or not VT. I cannot see why you feel the need to call someone an idiot over this.

PS, you're still wrong

NOT all Conroes have 4Mb cache. I suggest you go back over this entire thread and read it start to finish.

NathanE · 11 Apr 2007 at 20:13

No Allendale's have 4MB, they are all 2MB at most. They're fabricated from the get-go to be 2MB parts. Rather than wasting time at the end of the production line testing the whole fragile 4MB that a Conroe has and laser burning out half of it if its defunct. Cost savings 'n all that... Like I said the difference between Allendale and Conroe is simply the fabrication process. It's all about production process improvement.

BTW, expect Allendale's to OC very well when they are launched. They will produce less heat because they contain fewer transistors.

easyrider · 11 Apr 2007 at 20:52

NathanE said:
BTW, expect Allendale's to OC very well when they are launched. They will produce less heat because they contain fewer transistors.

I'm hoping the 6320 and 6420 get 3ghz +

Big.Wayne · 11 Apr 2007 at 21:39

NathanE said:
BTW, expect Allendale's to OC very well when they are launched. They will produce less heat because they contain fewer transistors.

Allendales are launched already [e4300] and yes in theory they 'should' run cooler but having used one they do seem to get very hot.

Minstadave · 11 Apr 2007 at 21:40

Big.Wayne said:
Allendales are launched already [e4300] and yes in theory they 'should' run cooler but having used one they do seem to get very hot.

They don't seem to clock that well either. The E4300s have been generally a little disappointing.

NathanE · 11 Apr 2007 at 21:48

Big.Wayne said:
Allendales are launched already [e4300] and yes in theory they 'should' run cooler but having used one they do seem to get very hot.

Yeah the E4300's are a bit limited though as they have a lot of ground to make up before you can even start taking them into new ground.

When Intel release some higher clocked Allendales (i.e. when they replace the E6300/6400 with Allendale) then I expect the OC situation will improve over what we've seen on the E4300.

No idea why they're running hotter though. That one alludes me

Very strange if that is actually the case...

Journey · 11 Apr 2007 at 22:00

Robbie G said:
Neither, from what I can gather, are arrogant misstatements...

Quote from you: "Why do people always feel the need to make things more difficult than they are"

In response I'd say - why do people feel the need to take an arsey tone when they are in fact totally wrong?

I'd love to know how you make text have a tone. If you RE-READ what I said, I simply was pointing out that people tend to feel the need to over complicate things when in most cases a very simple answer will suffice.

A bit like what you are doing now

I'd also like to clarify the word totally means everything, so you saying absolutley everything I wrote was wrong?

On a brighter note and nothing to do with any of this rubbish my 6420 was ordered today

Missed the next day delivery so I'll get it for Friday will post results Sat/Sun when I get time off work to play with it,