Processor question...

[old.Pure Science]

cheers people

i was thinking of a higher end PIII

cmon TUG post sumfink

 

[Embattle]

I just got a PIII 700E and I've been told its one that has cbo stepping which means it ocs well.

 

[old.Pure Science]

AMD or Pentium?

which should i get?

ta

 

[old.MeddlE]

Celeron II, then overclock like mad.

Should, by chance, TUG peek in here, I've got an intel sample Celeron II 667 on it's way from a friend that does 1GHz!! Wahey, just got to get a Slotket 3 from Abit.

 

[old.Gib]

Or u could get a Duron guaranteed overclocked, whips the Celeries ass these days
Or even better if u have the cash a P3, or Thunderbird overcloked , both round about the same speed for most stuff.

(all nice not overcloekd as well )

 

[stu]

Embattle is right - a 700E with stepping is what you want. If you can't get that, get a 600E.

Celerons are for cheapasses. AMDs are for stupid people. Durons are therefore for stupid cheapasses.

 

[old.TUG]

Right... AMD Durons, v. nice in an overclocked state but I'd only consider 2 mobo's: the Abit KT7/KT7 RAID and Asus A7V. Problem? Mobo's are not cheap. Lots faster than Celeron II's tho.

Celeron II's - excellent stop gaps if you already have an Intel BX mobo or similar with voltage manipulation and support for the said chip. Best bang for buck as you wouldn't need a mobo, just a better cooler and maybe a slotket. Speed wise, Celeron II 600@900 for example is about as fast as a PIII @ 700-750MHz.

PIII's - still not that cheap but pretty good. Faster than celeron II's by a fair whack but harder to overclock I guess too. Stu is almost making sense when he refers to steppings. All of 'em have steppings You just want the newer one (i.e. cb0, the old like my PIII 650E is ca2, and another new one is due for release - cc0). Newer steppings translate to lil' modifications where any imperfections are ironed out etc meaning higher overclocking potential as the CPU core is often 'purer'. My ca2 PIII 650 only does 815, even with top end cooling

Thunderbirds... not worth the cash over durons IMO as video cards these days are too much of a bottleneck so you wouldn't notice any difference from a significantly cheaper duron.

Soooooooo, starting from scratch I'd go either Duron 650@900ish MHz or 950MHz, or PIII 700@933.

If you make a choice and need mobo recommendations, supplier details etc then gimme a shout.

Meddle, I know a guy that can get loadsa unlocked PIII's too even unlocked 933's!!! l33t

[This message has been edited by TUG (edited 22 August 2000).]

 

[old.Pure Science]

nice one m8

thats TUGtastic

a PIII 700@933 sounds best, so any details u mentioned would be good m8. i will need a new motherboard i think?

mail me on anything@pure-science.co.uk or post it here if you want

ta everyone

 

[Embattle]

Good mine has cbo steping

 

[stu]

Anyone happen to know the theoretical safe maximum of the 0.18 die?

Intel have announced P4 on a 0.13 process apparently, with a bus speed of 400(!)Mhz.

 

[old.TUG]

Fuck safe Safety means nothing. Limit is the key. It's totally different than a safe operating frequency. Safety with CPU's comes in when you talk about safe temps, voltages etc... and the answer to those is 1.9v and 55 deg C before worrying.

ATM tho, the best cb0 stepping chips aint doing much more than 1000MHz with air cooling... peltiers, watercooling and so on can see anything upto 1.2GHz I spose

Mobo: I quite like the Abit BE6-II-100: ATA100 controller on board, BX chipset, voltage manipulation, bus speeds ranging from 84-200MHz in 1MHz increments and other features: www.abit.com.tw

i815E chipset mobo's: Asus CUSL2, lotsa FSB's, socket 370, excellent stability, fast, user friendly etc... quite expensive tho.

As for where to obtain such mobo's: www.dabs.com

Oh yeah, something useful you'll wanna know if you use a BX based mobo - your AGP bus speed as standard is 66MHz. Now if you run a 700@933 the AGP speed goes to 89MHz which is not the best for stability for older AGP cards. However, seeing as the BX chipset is only AGP2X capable, the AGP bus speed doesn't matter as much on AGP4X capable cards as really you're just running @ AGP3X with a 133MHz FSB and 89MHz AGP bus speed... Therefore for a BX based system, I specifically recommend GeForce DDR's upwards. NVIDIA make the fastest gaming cards around for the best prices too so stick wid them!

If you need more details on AGP dividers, bus speeds, PCI dividers and PCI speeds gimme a shout: dave@clockspeed.co.uk

Also gimme a shout if you want a source for PIII 700's @ 933 (I don't like to post competitors names here )

[This message has been edited by TUG (edited 22 August 2000).]

 

[old.MeddlE]

I'll admit to being a cheap ass if that's what getting a free (yes FREE) 1GHz processor means. It will drop straight into my existing rig with the aid of a slotket, no new mobo, no new RAM, and it has a globalwin fkp cooler with it too. So for the cost of a slotket I get a 1GHz CPU. That's cheap.

Stu-, how's your mp3 collection muppet?

 

[Embattle]

I'm going to punt mine to 933Mhz today, as long as the fan arrives

 

[old.TUG]

Why didn't you buy from me you c*nt?

I am da ch33p3sT (for some stuff)

spam time: www.clockspeed.co.uk plus I give free tech support so

 

[Embattle]

Nice story about fan.... http://forums.gameplay.com/showthread.php?threadid=170269

Sorry TUG would of but I just ordered from OCUK, the orb is cheaper from them any way

 

[Embattle]

Not accepting credit/debit cards is kind of backwards

 

[old.TUG]

o dear I sorta work for powercomputing but to generate sales plus other tasks and I used to work @ OcUK



Sorry you had problems with your order though, I'll have a word with the boss to ensure he sends the right stuff out in the future.

[This message has been edited by TUG (edited 31 August 2000).]

 

[Embattle]

You do that or I'll spank you

 

[stu]

Wow TUG you are so incredibly l33t. Although, if you'd actually read what I posted I asked what the theoretical safe maximum of the process is. Not "how fast does Intel recommend" or "how fast have people clocked their current chips up to". Ie physically, what is the maximum frequency before degradation in the die and interference sets in. Obviously you don't know. That would have been fine as a response.

Meddle: how's your collection of eastern european goat scat porn?

[This message has been edited by Stu- (edited 31 August 2000).]

 

[old.TUG]

Ok then. Frequency before degradation... err, too many factors that cannot pinpoint one frequency alone - heat being the main factor... others inc voltage, variances in a multitude of CPU's themselves (nothing is COMPLETELY uniform in every aspect) and so on. Therefore I cannot answer and I doubt anyone else could either.

 

[old.TUG]

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Stu-:
Anyone happen to know the theoretical safe maximum of the 0.18 die?
<HR></BLOCKQUOTE>

Actually, that question is not conclusive in what you're trying to ask anyway, you could stick a load of words on the end to give it a proper meaning.

i.e. Anyone happen to know the theoretical safe maximum *VOLTAGE* of the 0.18 die?

or

Anyone happen to know the theoretical safe maximum *CPU SPEED* of the 0.18 die?

Thats why I was maybe a little wayward in my reply to your question... Fair enuff?

 

[stu]

Aye. The point I was making (in a roundabout way) was that the conversion to .18 micron from .25 was made in the end because the .25 micron process just wasn't fine enough to support a chip of high enough MHz. It seems surprising that they've needed to switch down to a lower process for P4 already.

Although I guess the increase in speed from the fastest Katmais (600MHz iirc) to the currently fastest Coppermines (1233MHz and climbing) is pretty massive, it's just it's happened so quickly.

 

[old.[GA] Shovel]

nice to see you 2 getting along

SO this P4... - what speeds we talking this time? and will they be featuring another useless gimmik with them? or have Intel finaloly decided that no one cares about the catchy letters they can cram onto the end?

 

[old.TUG]

You're totally right there Stu, but I think you can now see that 'safety' has nothing to do with it in terms of max speed, it's just a limit of the manufacturing process. I think the Pentium IV's are 0.13? I forget...

Anyway, Shov, forget about the P4 *for now!* as it seems to be underperforming a PIII of the same clockspeed by upto 25%!

Might as well start learning about it in detail when it comes out, thats what I usually do with most products

 

[old.MeddlE]

I heard, and cannot confirm, that Intel have demonstrated the P4 running at 1.4GHz and also, and this one stretches the belief, at 2GHz!! Alpha will have to come up with some nifty new coolers for that baby!!

 

[bodhi]

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by TUG:
Anyway, Shov, forget about the P4 *for now!* <HR></BLOCKQUOTE>

Wouldnt a better reply have been "Forget them for now, because they arent out yet"?

 

[old.TUG]

Nope

 

[Durzel]

I'm not sure cooling is the "be all and end all" of microprocessor overclocking. Sure, any fool will tell you that it would be possible to overclock any chip of any factory Mhz to 3GHz simply by cooling it sufficiently. I believe (I'm not a science expert so I can't attest that this is necessarily the truth) that there are a number of factors - build quality, quality of parts, etc that contribute to the theoretical "maximum" a processor could be clocked at.

Incidentally, I ran one of my two Pentium II's at 504Mhz (from 450, 112 FSB) for about a year in Windows 98, upgraded to Windows 2000 (wherein both processors are utilised) and I have no end of timeing problems, lockups etc. I'm beginning to wonder if one of the processors is "burnt in" at the higher speed (or something)..

 

[old.TUG]

Cooling is joint most important factor in overclocking attempts IMO, next thing is motherboard quality and stability.

Yes, there are lotsa scientific details outlining limitations, I think www.overclockers.com have some articles on it.

As for poor stability in win2k, got better coolers for ya chips? Case cooling is important too...

Also, if possible, knock the voltage up to 2.1v and that should solve ya stability probs.