Which amd x2 for overclocking?

[danoliver1]

Cheers guys gonna buy a 3800+ and overclock it cant complain for 99quid+vat can you

 

[Impy77]

After all this talk of the x2 3800 being the one to get and it not even worth bothering to consider the 4200 or 4600, can I ask this:

I appear to be limited by my motherboard to a FSB of 278 ish before its unstable (and I've tried many voltage/HT Multi/CPU Multi/RAM dividers on both a winnie 3200 and my current Opty 144 so I know its the max I can get stable at - yes this sucks but anyway).

It is worth me going for the 4200, so I get 11*278 = 3058Mhz rather than 10 * 278 = 2780 Mhz as my theoretical maximum, or is there no chance any of these chips will clock as high as 3Ghz anyway?


Thanks,

Impy

 

[realscot]

as far as X2-4200+ overclocking, i've got mine at 2.5 on stock volts with a 40C full load temp (using AC64pro) and that's about as far as it will go with the stock volts it looks like. (230x11) Someone told me that the 1mb core's tended to run a wee bit hotter?? might not be true.

 

[A.N.Other]

Go for the X2 3800 definitely. By far the best bang for buck.

Look here for OCs with it - and I'd like to hear from you in the thread if you do get one.

 

[trojan698]

Having had both the toldeo and manchester cores it made no difference to the overclock.

If going for a 512kb model then the only one to get is the x 2 3800 all the others are redundant in my book,waste of money and clock no higher.


If going for a 1mb model then the only one to get is the x 2 4400 again all the others are redundant in my book,waste of money and clock no higher.

A x 3800 can be either a toledo core with half the cache disbabled or a manchester core a 512kb core designed from the ground up.

My old x 2 3800 was a toledo core and this made no real world difference.


Just a tip calling people fools in forums that offer information to others is not big and clever and makes you look like one yourself tbh

In general the manchesters seem to clock better just take a look at the above link. Just because you've had one of each, it doesn't actually mean anything. Also, you're one to talk about acting big and clever TBH.

 

[easyrider]

In general the manchesters seem to clock better just take a look at the above link. Just because you've had one of each, it doesn't actually mean anything. Also, you're one to talk about acting big and clever TBH.

The Toledo in that graph needs less vcore than the manchester at much the same clockspeed.

There are more manchester cores in that graph because the toledo core x 2 3800's were a lot more rare, and the chances are, if you were to get an x 2 3800 today it would be a manchester core anyway.

I'm just stating that calling people fools when they are discussing cpu's is not really helpful is it?

it doesn't help anyone.

Having had experience with both cores the toledo and manchester versions of the 3800 I felt it was valid for this thread.Hands on experience is helpful when making comparisions,I dont see how it does not mean anything as you put it.

You state

It's not about the multis or the higher stock speed, fools. its the manchester vs the toledo core.

Then you say

Just because you've had one of each, it doesn't actually mean anything..

Please make up your mind lol

 

[Angilion]

So they're exactly the same chips passing exactly the same tests? AFAIknew they all go through testing and they are sorted into which are most stable at certain clock speeds.

Sometimes, but not always. If yields of chips that pass standard tests at a relatively high clock speed are very high, some of them will be marked lower anyway because of demand.

Yes, the company could mark them all as higher models and meet demand simply by selling the higher models at lower prices, but that will become expected and negatively impact on future sales.

Purely as an example, I'll have Acme CPUs making their Z4 range, rated on some hypothetical performance scale as 3000 to 4000. Over time, Acme refine the manufacturing process, develop new cores, the usual thing, and extend the range to 5000. Prices range from £75 for the 3000 to £500 for the 5000. Due to the newer cores and refined manufacturing, nearly all of their working chips pass their tests as 4500s. This model sells at £250.

Acme have three choices:

i) Sell their CPUs at the the speed they pass the tests at, so almost all are 4500s or higher, and maintain prices. This effectively removes Acme from most of the market, since they have hardly any CPUs below £250.

ii) Sell their CPUs at the speed they pass the tests at, so almost all are 4500s or higher, and drop prices to meet the market. 4500s would have to drop to about £100 and 5000s couldn't be much more or else they would be crappy value compared to the slightly slower 4500s and hardly any would sell. In the short term, this is great. They will spank the opposition on price/performance and gain market share. After the short term, it's bad for Acme. There will be a price war, they will have less money coming in for R&D and they will have to either sell their next range of CPUs for a similar price or come up with astonishing increases in performance, or else their next range won't be popular. Why pay £500 for a new Z5 rated at, say, 6000, when you can get a Z4 rated at 4500 for £100? Under those circumstances, multi-CPU server-type boards might end up in the mainstream market because enough people would want one to fit two Z4s in instead of one Z5.

iii) Sell their CPUs at the speeds required to meet market demand at the usual prices. So what if most of them pass the tests as 4500s? Mark as many as wanted down as, say, 3500s and sell them that way.

(iii) is the best course of action for the company, so that's what they do. Under some circumstances, it is indeed the case that CPUs rated at quite different speeds have in fact passed the same tests for the same speed. Usually on a mature range after a few revisions, like Athlon 64 (X2 or otherwise).

 

[trojan698]

Please make up your mind lol

Do you not understand? I'm saying that the 2 different cores clock differently in general. I'm not interested in an isolated case where you personally found there no difference between the clocking ability of those 2 particular chips you had.

 

[easyrider]

Do you not understand? I'm saying that the 2 different cores clock differently in general. I'm not interested in an isolated case where you personally found there no difference between the clocking ability of those 2 particular chips you had.

Whatever lol

People on XS found the toledo x 3800's more desirable as its the same core as the 4400 and 4800 that run at higher clocks and that all that is different is the disbaled cache.

 

[A.N.Other]

All testing has proved that the machester is a far better clocker. Can't remember the exact figures, but the toledo record stands at about 3350 or something, but the manchesters can go quite a bit higher.