128 cache vs 256 cache / celerons

[grubazaba]

i see that the fcpga2 celalatins ccome in 1.2, 1.3 & 1.4 with 256Kb L2 and the 1.7 & 1.8 come with only 128Kb...

where will i notice the difference in L2 cache?

...also i hear lots of talk of the 1.2 but not much higher. is that the best overclocker?? i've heard of plenty of ppl getting that to 1.5...

(virtual) penny for your thoughts!

[Celemine1Gig]

Well, you're right, they're all FCPGA Celerons, but there's one big difference:

The 1.2, 1.3, 1.4 GHz FCPGA2 Celerons are based on the 0.13µ Tualatin P3 core and have 256KB L2-Cache which makes them about as fast as a P3(not the P3-S with 512KB L2-Cache, which is much faster!) at the same clockspeed and FSB(there are also 1.0 and 1.1 Ghz versions of these Celis). In contrast to these CPUs the 1.7 and 1.8 Ghz Celerons are based on the 0.18µ Pentium 4 Willamette core and only have 128KB L2-Cache. These P4 Cels are severely crippled P4s and have neither good performance, nor good overclocking potential.

First of all, you won't be able to run any of these chips in your current system(in your sig.) without modifications.

For the P4 Celis 1.7Ghz and 1.8Ghz you would have to buy a P4 mainboard.
For the Tualatin Celis 1.0, 1.1. 1.2, 1.3 and 1.4 GHz you have 3 options:

1)buy a Tualatin compatible mainboard

2)get a Powerleap Converter Card or Socket to run the FCPGA2 Tualatin Celi in your current board

3)do the socket modification yourself as described in many threads here or in an article on the frontpage or in some user articles

So, it's up to you,to decide!

Regards

Ingo;)

[grubazaba]

Thanks ingo!

i guess should change my sig - i now am the proud owner of a TUSL2-C/A (still in it's box!!) ready for my first overclock.

just a few things: how did you get 140fsb without having to raise your voltage? ;)

i have seen the 1.0/1.1 & 1.0a/1.1a what's the difference - is the a version the tualatin with 256?

and these P3-S with 512KB... the s stand for server? aren't they a bit expensive! like dbl the price!!

oh... first i'll overclock my cumine then i've seen a 1.2 for £54 inc vat (17.5%) and a 1.4 for £62 (inc) - although the shop says it's 128 L2 so i have just written to check.

i'd like a good p3 but they are just too expensive....

[deez]

Many of the Tualatin chips are excellent overclockers and do not require much voltage increase especially with a good heatsink like the AX-7

The Tualatin celeron is designated by A and has 256K L2 cache and is basically a coppermine P3 with an extra instruction set...costs less and overclocks better. The 1.0A and 1.1A are pretty much definite for 133FSB+

and yes the P3-S is a server level chip and it has 512K L2 cache which makes it perform much better in CPU intensive activities like SETI, folding, rendering etc... my SETI times with a P3-S @1434 are about 1.5 hours less per work unit than a cele 1.0A@1450. Doesn't do much for gaming though.

[grubazaba]

thanks deez... would you say an A chip is better than a 1.2? also i just noticed that *OUR* boards (cos i just got a tusl2-c/a) won't allow us to up the vcore for a tualatin - doesn't that cause probs when overclocking a celeron II ??

...oh , just found some coolermaster HHC-L61 HSF for £9 !!

thanks for clarifying the p3-s - i do a lot of 2d graphics and gaming. the 2d is more memory intensive than processor i think.

[deez]

All of the 1.2 celerons are A chips but the 1.0 and 1.1 were made in both coppermine and Tualatin. While the 1.2A is generally a good overclocker some have not been able to hit 133FSB thats why I suggest a 1.0A or 1.1A instead

and the issue with the TUSL-2C has been resolved in recent bios releases. Not changing the vcore was only a problem when the board first came out so dont worry.

The heatpipe got very good reviews and some forum members here use them. Personally I love my AX-7 and thats what I recommend. The SLK-800 Kicks butt too but its twice as much $$$

I would only suggest a P3-S if you do extensive graphics work that requires a lot of CPU usage like rendering...for gaming you will notice very little difference. Personally I only got it b/c I'm a SETI fanatic and SETI takes up 100% of your unused CPU cycles.

[grubazaba]

hey deez... it arrived: 1.1a tualatin:
01 philippines
1100A/256/100/1.475
7137A474-0892 (SL5ZE)

a big chip compared to my cumine p3

oh, and...it has a little hole in the heat spreader!

why is that? and do i have to be careful with the thermal grease?

[looktall]

Originally posted by grubazaba
hey deez... it arrived: 1.1a tualatin:
01 philippines
1100A/256/100/1.475
7137A474-0892 (SL5ZE)

a big chip compared to my cumine p3

oh, and...it has a little hole in the heat spreader!

why is that? and do i have to be careful with the thermal grease?
that appears to be a rather old cpu.
made almost a year ago. (unless i've read the FPO/batch code wrong)

i believe the chip is actually slightly smaller than on the coppermines, but you can't tell because of the heatspreader.

the hole in the heatspreader is a remnant from manufacturing, it no longer serves a purpose. no need to worry about thermal paste getting inside it.

[grubazaba]

i meant the whole chip... as in the green board.

yep, it looks to me like a 2001 chip....
and there's a bloody great warranty sticker over the heatspreader that looks like it'll be a messy job getting off.

thanks for that about the hole!

[old is ok if it aint used]

[deez]

congrats on the new chip...you should hit 133FSB no problem.

Warranty sticker on the heatspreader? That must have been put there by the dealer cuz intel doesn't do that. Try cleaning it off with some rubbing alcohol or electronics cleaning solution

[looktall]

what sort of idiot puts a warranty sticker on the damned heatspreader?
great way to make you need to use the warranty. :rolleyes:
you're going to have to remove that sticker for sure. leaving it on will make the temps higher and possibly kill the chip.

what ever you use to remove the sticker, try not to scratch the heatspreader.

[grubazaba]

...i nearly screamed! - but the time to wait for a replacement - can i be bothered? nope.

as soon as i get my pci modem i'll rip this pc-chips board apart and put my tusl into play.

got 2 coolmaster cooldrives as well! look very nice - testing will see.

what's all this about burning in?? - i'll have to start at 100Mhz anyway till i upgrade the bios. won't allow me to adjust the vcore for tualatin (according to the manual)

[looktall]

my 1.1A has never even seen 100mhz. it's very first boot was at 133mhz.
it does it at stock volts too, you might like to try it first.
mmm 1.4ghz of tualatin goodness. :D

i've never burned in a cpu yet. i certainly never burnt in my 1.1A and it can run at 1.6ghz.

[deez]

grubazaba....just flash the bios to the most recent version 1012 and you can adjust the vcore with no problem. And since your planning to overclock you will probably want to use the tweaked bios from here http://www.x86-secret.com/articles/tweak/i815twken.htm its 100% rock stable

[grubazaba]

nice one deez! checking it out now.

[deez]

any luck?

[funnyperson1]

good luck...keep us updated...you can tell im very happy with my purchase :)

[Jola]

I'm running a 1.1A in a CUSL2 (wire tricks!)

No problems - straight into FSB=133 and never been lower.

Also ran it at FSB=140, but slightly unstable there - with some work it would probably be stable there.

[grubazaba]

Originally posted by deez
any luck?

i'm just waiting for my v92 modem to arrive!

this pcchips has onboard modem... i don't want to switch boards and not have a modem for a week or so...

but i will keep you all posted thanks for the help!!

Anyone give me some pointers on this?
http://forum.oc-forums.com/vb/showthread.php?s=&threadid=62822&perpage=30&pagenumber=4

[grubazaba]

sugar! i wrote about an A4 page and it crashed sending!

to make it shorter... i need some help here guys!

new mobo. new chip. extra hard drive. More efficient heatsink. Extra case fan.

my *bios* says 27(system) 37(cpu) *asus probe* says 27, 40 *MBM5* says 27, 16, -32 (what's this third sensor?)

now the asus probe reading of 40 NEVER changes!!! i asume it can only read temps of over 40* or sth crap like that?

...and the coolermaster heatsink (designed for 2.2 chips) is blowing cool air. so mbm could be right.

but i'm not sure as the bios is usually correct!

my volts are rock steady at 1.50, 3.49, 4.99, 11.73. all three tests produce EXACTLY these results - which was shocking. [except in bios where 12v line reads 12.28]

based on my old system temps and the additional cooling (and the hard drive in a cool drive!) i'm a bit worried about overclocking.

but i slapped it into manual: 133,133,33 just to see what would happen.... and it wouldn't post. so i also tried (manual setting off) increasing the target frequency of the chip and it won't do anything over 1.1.

my bios is the 1011 (mid 2002). will try the 86bios when i know i'll be safe!

what is my real chip temp? and where do i go from here... should the new tualy cpu run so hot? or is it a case of re-arctic silver! - i used a little less than my old HS because that was alu and really rough and this is copper and smooth as a baby's bottom ;) (not that i'd know about that)...

deez or someone can you help? thanks!

[6502kid]

I dont thing you will an OC out of it until you pop that
x86 bios in it. Get it running at stock, and flash that bad boy.

[grubazaba]

+----------+----------+----------+-------+-------+--------+--------+
| | | | S1 | S2 | S3 | C0/1 |
+----------+----------+----------+-------+-------+--------+--------+
| 26/09/02 | 20:34:06 | 1474 MHz | 32° C | 22° C | -32° C | 1.70 V |
| 26/09/02 | 20:33:36 | 1474 MHz | 32° C | 22° C | -32° C | 1.70 V |
| 26/09/02 | 20:33:06 | 1474 MHz | 32° C | 22° C | -32° C | 1.70 V |
| 26/09/02 | 20:32:36 | 1474 MHz | 33° C | 22° C | -32° C | 1.70 V |
| 26/09/02 | 20:32:06 | 1474 MHz | 32° C | 22° C | -32° C | 1.70 V |
| 26/09/02 | 20:31:36 | 1474 MHz | 32° C | 22° C | -32° C | 1.70 V |
| 26/09/02 | 20:30:21 | 1532 MHz | 32° C | 22° C | -32° C | 1.70 V |
| 26/09/02 | 20:29:51 | 1532 MHz | 32° C | 22° C | -32° C | 1.70 V |
| 26/09/02 | 20:29:21 | 1532 MHz | 32° C | 23° C | -32° C | 1.70 V |
| 26/09/02 | 20:21:09 | 1532 MHz | 32° C | 22° C | -32° C | 1.65 V |
| 26/09/02 | 20:20:39 | 1532 MHz | 32° C | 21° C | -32° C | 1.65 V |
| 26/09/02 | 20:20:09 | 1532 MHz | 32° C | 21° C | -32° C | 1.65 V |
| 26/09/02 | 20:19:39 | 1532 MHz | 32° C | 22° C | -32° C | 1.65 V |
| 26/09/02 | 20:15:48 | 1479 MHz | 31° C | 22° C | -32° C | 1.65 V |
| 26/09/02 | 20:15:18 | 1479 MHz | 31° C | 21° C | -32° C | 1.65 V |
| 26/09/02 | 20:14:48 | 1479 MHz | 31° C | 20° C | -32° C | 1.65 V |
| 26/09/02 | 20:14:19 | 1479 MHz | 31° C | 20° C | -32° C | 1.65 V |
| 26/09/02 | 20:13:48 | 1479 MHz | 31° C | 21° C | -32° C | 1.65 V |
| 26/09/02 | 20:13:18 | 1479 MHz | 31° C | 22° C | -32° C | 1.65 V |

[grubazaba]

x86-ed it!
1.475v no post
1.525v freeze on post
1.575v just about makes safe mode then crashes
1.625v whoopeee... dooo! doh! very unstable, virus scan, firewall, mbm, netscape...crash (again & again)
1.675v (last chance) - houston we have ignition - and

the results from mbm5 (above) note that the asus probe still says 40* and my HSF air flow is cool still. I'm 0n 133/133/33.

where's ol' man? he was saying something about having to overclock to 134 to get this board to do anything.

by the way should DMA be checked for hard-drives in my computer properties? I have ATA 100 drives... & should they only be checked at 100FSB?

Anyway... 1466 my first overclock!

[6502kid]

I dont think the 1/4 divisors kick in on that one until you
hit 134 or above.....

[6502kid]

Mine likes 135mhz with 1.6v

[grubazaba]

thanks kid!

what's a 1/4 divisor? and what does it affect?

whatever it is... does it mean that i might be able to run at 134 or 135 at a lower voltage?

[looktall]

1/4 divisors are what makes the PCI bus drop back into spec.
a quarter of 133 = 33mhz.
when you run it at 100mhz the PCI divisor is 1/3.
there is also an AGP divisor. which is 2/3 for 100mhz and 1/2 for 133, which both put the AGP back into spec. (66mhz)

but erm..6502kid, shouldn't that be available for 133mhz and above, not 134mhz? otherwise when you try to run a non overclocked PIII your system is going to be out of spec.

[6502kid]

You would think so, but a number of these 815 chipset boards
do that when you try to crank a 100mhz default CPU up to
133 or above.

1/4 works like it should at 133 with a P3 in them......

Dont ask me why, cause I dont know.

[looktall]

hmmm interesting. do these mobo's use jumper as well as bios or just bios?

if jumpers, does it still not use 1/4 at 133?

it's not much of a problem for some of these lower celertins that can easily run at 133mhz.(1mhz more won't hurt them :p) but when you get some of the higher speed chips struggling to reach 133, that extra mhz could be the difference between posting and not posting.

[grubazaba]

this is what i see.... you basically get two options in the bios to tweak from:

100/100/33 - which goes up to 132/132/44 (fsb/sdram/pci)
or from 133/133/33 plus

I am using 133/133/33 which won't post at default and crashes on windows screen without making desktop until 1.625 voltage.

however i tried the 132/132/44 and this posetd straight away at default but locked on windows screen. 1.525 gave windows protection error and 1.575 took me to desktop (same with 1.625/1.675) - but this way you can't get past the desktop as the onboard sound seems to kill the system (you know windows startup sound).

133/133/33 - is quarter divisor right? ie pci should run at 33.

or was ol man saying that even though it shows 133/133/33 it's not running at 133/133/33? i'm confused.

[grubazaba]

Originally posted by looktall

but erm..6502kid, shouldn't that be available for 133mhz and above, not 134mhz? otherwise when you try to run a non overclocked PIII your system is going to be out of spec.

OK. I sussed it:

ol'man was wrong... you get 1/4 at 133 (or the above would be true!) - but i can see this board would cause problems: if you get a chip that's maybe capable of 129FSB the 44 MHz PCI would probably not let you clock past about 102.
However, if you get a good chip capable of 133+ then you can start at 133/133/33 (1/4 divisor) - i apparently am pushing my chip to the limit - it won't even post at 1.625 volts for 135/135/33.
@133/133/33 mine posted & reached desktop but was was unstable at 1.625 & the next step is 1.675 (which incidentally in reality appears to be 0.25 higher - ie 0.7v) so there's no way my chip would run stable at 135!

now that's what you call jammy! ;)