Differences between OEM and retail PIII 450 ???
Mark Powell
to get it to overclock to more than 527MHz (4.5x117) on an ABit
BX6 v2. However, if I slot in my brothers retail boxed PIII 450
that was bought about two months ago exactly the same hardware will
clock up to 581MHz (4.5x129) with a slight voltage increase.
Both CPU cases have the same detail on top, except that the OEM
version was manufactured in the Phillipines and the retail was
made in Malay. The part no. of the cache chips on the OEM version
end with -4H (which I'm assuming means 4ns) and just underneath
say 250MHz. I've not been able to check out the cache on the retail
version as the heatsink looks difficult to take off and it's not
my CPU :)
Anyone else noticed such behaviour? Should I take my OEM version back
for a retail or have the days of the vastly overclockable PIII 450
ended? I hope not.
Cheers.
--
Mark Powell - System Administrator (UNIX) - Clifford Whitworth Building
A.I.S., University of Salford, Salford, Manchester, UK.
Tel: +44 161 295 5936 Fax: +44 161 295 5888 www.pgp.com for PGP key
M.S.P...@ais.salfrd.ac.uk (spell salford correctly to reply to me)
Vincent Feyaerts
Oem from Philipines does 5...@2.0v. (600 also but my fricking memory
doesnt allow it). You just have to be lucky - perhaps another OEM from
the same plant allows much higher fsb speeds.
Mark Powell heeft geschreven in bericht
<7mhklb$8du$1...@plato.salford.ac.uk>...
Mark Powell
Vincent Feyaerts <xvin...@pandora.bex> wrote:
>I had a OEM from Malaysia that would only do 558 at a high voltage. My
>Oem from Philipines does 5...@2.0v. (600 also but my fricking memory
>doesnt allow it). You just have to be lucky - perhaps another OEM from
>the same plant allows much higher fsb speeds.
So it's definately not an OEM vs. retail issue. It's just pot luck
whether you get a CPU that'll be stable at the higher speeds? After
reading all the stuff on the web about the great overclocking
qualities of the PIII 450, I assumed that they could all get up to
581MHz.
Do all the PIII 450's have the same cache as I detailed. i.e.
"-4H" markings and "250MHz" on the cache ICs?
Mark Powell
Mark Powell <ma...@salfrd.ac.uk> wrote:
> Do all the PIII 450's have the same cache as I detailed. i.e.
>"-4H" markings and "250MHz" on the cache ICs?
For interests sake, the OEM PIII 450, I've just returned to the shop, had
the following markings:
On plastic case:
450/512/100/2.0V S1
19210412-0051 Philippines
imc '98 SL364
On cache chips:
M5M5V2236BGP
917BE0X-4H
250MHz
Looks like I'm going to have to take a chance and get hold of another OEM
PIII this evening. Fingers crossed. Is this really the only way to do it?
<sigh>
Antos Dragon
those chips are rated as -250 it could be the core holding the
overclock back...or it could be a crappy L2 manufacturer... the only
way to really know would be to disable the L2 cache and try to boot
higher (even at 124 and 133)...and if it was stable at 558/581/600,
then that l2 manufacturer isnt up to snuff... if it still was
unstable (especially at 581 L2 off, any voltage), then you simply got
a lower quality core....
But since you took the chip back... good luck on the next one...
BTW this is VERY important as far as solving the overclock riddle...
if your new chip also has trouble going at least 558 (or wont do
5...@2.0v), try disabling L2 and seeing how far it will go...and post
your results....
I still personally think that the retails in general may have slightly
better cores but your findings could prove it for sure...
Mark Powell
Antos Dragon <kylearan@my*dejanews.com> wrote:
>But since you took the chip back... good luck on the next one...
>BTW this is VERY important as far as solving the overclock riddle...
>if your new chip also has trouble going at least 558 (or wont do
>5...@2.0v), try disabling L2 and seeing how far it will go...and post
>your results....
Got the new one. Guess what? Exactly the same behaviour. Stable up
to 527MHz (4.5x117) at 2.0v. Not stable any higher, with any voltage
and with L2 cache off. A combination of bad core and bad L2? As I
didn't check the previous one with L2 off, that may have just been
bad core? The -225 would seem to indicate to the casual observer
(me) that this one has slower L2 than the last.
Not having much luck here. I've seen a couple of mentions
on the web of people claiming higher speeds for OEM versions. Maybe
it's a territory thing? UK supplied retail have better cores?
Anyway, I'll take this fella back too. Probably next week now
though.
CPU case:
450/512/100/2.0V S1
99220434-0333 MALAY
imc '98 SL364
Cache chips:
TOSHIBA L30173
TC55V2377AFF-225
JAPAN 9918 MBD
Core label: L918A178
Antos Dragon
600 unless L2 is off (then its stable) or unless I freeze the chip
first... going to look at ways to cool the cache down...
Mine is a retail, BTW...
You said your friend's P3 did do 600 in your system?
If his DID do 600 and yours wont even do 558 with L2 off, its
definitely a core or L1 cache problem...
Its definitely not 'regional'...its basically the luck of getting the
right lot... there is NO way to tell if an OEM has a good core or not
except by testing it...
Id return that one (if possible) and get a guaranteed 558 @2.0v
chip...
Phat
have you tried upping the voltage? sanding? putting a good heat sink on?
Mark Powell <ma...@salfrd.ac.uk> wrote in message
news:7miubi$d11$1...@plato.salford.ac.uk...
Mark Powell
PC133. I've already said that a retail P3 450 was stable at 581MHz when
slotted into my machine.
>have you tried upping the voltage? sanding? putting a good heat sink on?
Seems no point considering the above? The retail I used has obviously
got better L2 *and* core. I'm looking for something that'll be stable at
558 or preferably 581 without any messing about. Then I'll consider getting it
higher.
Mark Powell
Antos Dragon <kylearan@my*dejanews.com> wrote:
>I have no clue... Mine has -225 cache, and does 5...@2.0v, but wont do
>600 unless L2 is off (then its stable) or unless I freeze the chip
>first... going to look at ways to cool the cache down...
Wouldn't the TenMaxx P3 FT do that? Or do you mean even more cache
cooling?
>Mine is a retail, BTW...
>
>You said your friend's P3 did do 600 in your system?
It was a retail and was stable at 581...@2.1v.
>If his DID do 600 and yours wont even do 558 with L2 off, its
>definitely a core or L1 cache problem...
Yeah, seems that way. Two different dealers with bad cores though.
Things ain't looking good. I think I'll wait until next week until I
replace this one. Maybe the crud 'll've been bought by then.
>Its definitely not 'regional'...its basically the luck of getting the
>right lot... there is NO way to tell if an OEM has a good core or not
>except by testing it...
I was thinking as much. Nevermind, I'll keep searching for the Grail :)
>Id return that one (if possible) and get a guaranteed 558 @2.0v
>chip...
That's difficult in the UK. I've yet to find a dealer that you can
chat to about overclocking. They all just roll their eyes and say
it's "not recommended." Probably because none of the them know
anything about it, except that it's definately a BAD thing :( Thus,
it's not really possible to get them to give you a CPU that works
at anything other than 450 <sigh>
Nick Hall
[snip]
>
> I still personally think that the retails in general may have slightly
> better cores but your findings could prove it for sure...
>
Surely the only difference between an OEM and retail is the box. I'm
sure that not all malay chips are retail, and all Philippines are oem.
Mark, try to buy a Philippines then, if your brothers o/c's better. BTW,
I have an OEM c300a, which is fine at 450, and apparently the retail
300a's were better overclockers. It's a costa rica chip IIRC.
--
Nick
Remove .SPANKYSPAM from my email address to reply......
JustinC
<ma...@salfrd.ac.uk> writes
>
>>Id return that one (if possible) and get a guaranteed 558 @2.0v
>>chip...
>
>That's difficult in the UK. I've yet to find a dealer that you can
>chat to about overclocking. They all just roll their eyes and say
>it's "not recommended." Probably because none of the them know
>anything about it, except that it's definately a BAD thing :( Thus,
>it's not really possible to get them to give you a CPU that works
>at anything other than 450 <sigh>
>
some vital point.
Intel make chips, rate them for clock speed, fine. Some have more
tolerance for over clocking than others, this is a result of the
manufacturing process... But the rating Intel give *is* attainable.
There is nothing wrong with the chip, nothing is being sold that doesn't
perform as it should. If you can over-clock, fine, if not tough. How you
can expect (or hope) to find a dealer to swap CPUs for you because 'they
do *exactly* what the manufacturer promised' I don't know.
Analogy:
You sold me a car. It does exactly what you said it would but I thought
I could make it go faster running it on nitro, it didn't work as
planned, I want to exchange it for a different one.
The chip manufacturers are not going to honour returns on faulty goods
if the dealers accept chips back from people who void the warranty.
--
Justin C by the sea. mailto:9907....@j-catter.demon.co.uk
Mark Powell
JustinC <9907....@j-catter.demon.co.uk> wrote:
>I don't understand the logic of this thread; either that or I'm missing
>some vital point.
It is in an overclocking group. I crossposted it to a uk group in a
hope of getting some local feedback.
>Intel make chips, rate them for clock speed, fine. Some have more
>tolerance for over clocking than others, this is a result of the
>manufacturing process... But the rating Intel give *is* attainable.
Agreed.
>The chip manufacturers are not going to honour returns on faulty goods
>if the dealers accept chips back from people who void the warranty.
So how do they tell the warranty has been voided?
Mark Powell
Nick Hall <ni...@vw-beetle.freeserve.co.uk.SPANKYSPAM> wrote:
>Antos Dragon wrote:
>[snip]
>> I still personally think that the retails in general may have slightly
>> better cores but your findings could prove it for sure...
>
>Surely the only difference between an OEM and retail is the box. I'm
>sure that not all malay chips are retail, and all Philippines are oem.
>Mark, try to buy a Philippines then, if your brothers o/c's better. BTW,
I originally had an OEM Philippines, which attained 527. I got a
replacement OEM Malay which also could only attain 527. I'm talking about
stably here. They will boot at higher speeds, but will eventually throw-up
an error during soak testing. My brother's CPU is a retail Malay. That
was the good one that'd do 581 stably.
Steve
> for a retail or have the days of the vastly overclockable PIII 450
> ended? I hope not.
Come on be honest - you have left on the OEM fan and not used thermal
compound haven't you :)
--
Steve - UK Gamers
st...@ukgamers.com
http://www.ukgamers.com
JustinC
<ma...@salfrd.ac.uk> writes
>In article <s7gjZSAT...@j-catter.demon.co.uk>,
>JustinC <9907....@j-catter.demon.co.uk> wrote:
>>I don't understand the logic of this thread; either that or I'm missing
>>some vital point.
>
>It is in an overclocking group. I crossposted it to a uk group in a
>hope of getting some local feedback.
>
>
>So how do they tell the warranty has been voided?
by a higher voltage than it is designed for.
Conor Turton
catter.demon.co.uk says...
> >So how do they tell the warranty has been voided?
> Pass, I'm no techie but I'm sure it must be evident if a chip is damaged
> by a higher voltage than it is designed for.
>
They can't tell unless you are stupid enough to modify the chips as some
japanese ones have been.
--
________________________
Conor
conor....@bigfoot.com
ICQ:31909763
________________________
manny
Philippines is very stable at 600 and 2.1 volts. It even gets into
windows at 630 and 675 but hangs. I haven't taken off the plastic
case and I am using a generic PIII cooler with 2 fans. I did take out
the "plastic" square on the crappy hs and put some thermal paste from
computernerd. I got my 450 from microsource (www.invado.com) and i use
corsair pc133 ram.
If yours won't overclock, try using higher voltages. I bought 3
c300a's last year. One was stable at 2v, the other at 2.1 and the 3rd
won't run stable at anything less than 2.4v even with a sandwich
cooler and after 7 months of continuous use!
What reason do you tell the retailers for returning non-overclockable
processors?
-Manny
J.T. Wenting
CPU sold as (for example) a PIII-450 can attain is therefore not a fixed
number, but varies between 450MHz and at 499MHz (while remaining within the
design parameters of the CPU). When overclocking, these numbers increase but
there is still a range (which may be even wider).
I know someone who purchased 2 Celeron 300As from the same source, at the
same dat, and was able to get 375MHz out of one, 450 out of the other. As
they came from the same batch, they will have come from the same production
line also.
--
J.T. Wenting
www.hornet.demon.nl
Fight war*z!
Steve
> chat to about overclocking. They all just roll their eyes and say
> it's "not recommended." Probably because none of the them know
> anything about it, except that it's definately a BAD thing :( Thus,
> it's not really possible to get them to give you a CPU that works
> at anything other than 450 <sigh>
We specialise in overclocking gear.
Mark Powell
JustinC <9907....@j-catter.demon.co.uk> wrote:
>>So how do they tell the warranty has been voided?
>Pass, I'm no techie but I'm sure it must be evident if a chip is damaged
>by a higher voltage than it is designed for.
But you can overclock without having to increase the voltage at all.
Just increasing the FSB speed is all that's often needed.
Mark Powell
manny <remove...@ix.netcom.com> wrote:
>If yours won't overclock, try using higher voltages. I bought 3
>c300a's last year. One was stable at 2v, the other at 2.1 and the 3rd
>won't run stable at anything less than 2.4v even with a sandwich
>cooler and after 7 months of continuous use!
I have tried higher voltages. I went up to the max that my BX6 v2 allows
2.3v. I didn't try any taping etc. to go higher. I've a Tennmax P3 TF on
order, so I'll try it with that fitted, before I return this one.
Hopefully the delay will mean a different batch of CPUs, too.
>What reason do you tell the retailers for returning non-overclockable
>processors?
Currently I've purchased them from different places and blagged that I
actually wanted a retail CPU and didn't realise the difference. Just
played dumb really. Easy to get a full refund that way. Running out
of dealers in the Manchester area though :)
Mark Powell
Steve <st...@ukgamers.com> wrote:
>> Anyone else noticed such behaviour? Should I take my OEM version back
>> for a retail or have the days of the vastly overclockable PIII 450
>> ended? I hope not.
>
>Come on be honest - you have left on the OEM fan and not used thermal
>compound haven't you :)
How did you guess :) It's a crappy Jamicon P3 cooler. Like I said
in a recent post, I'm going to try this one with a Tennmax P3 TF,
before I give up on it. The retail CPU that'd do 581MHz wouldn't
have had any thermal compound on it though? I did see someone mention that
their Intel heatsink *did* have some compound on it though. So maybe.
JustinC
<ma...@salfrd.ac.uk> writes
>
>Currently I've purchased them from different places and blagged that I
>actually wanted a retail CPU and didn't realise the difference. Just
>played dumb really. Easy to get a full refund that way. Running out
>of dealers in the Manchester area though :)
>
OEM and the retail version?
JustinC
<jai...@firepile.demon.co.uk> writes
>On Mon, 19 Jul 1999 23:30:41 +0100, JustinC
><9907....@j-catter.demon.co.uk> wrote:
>
>>In article <7mussc$12j7$1...@plato.salford.ac.uk>, Mark Powell
>><ma...@salfrd.ac.uk> writes
>>>
>>>Currently I've purchased them from different places and blagged that I
>>>actually wanted a retail CPU and didn't realise the difference. Just
>>>played dumb really. Easy to get a full refund that way. Running out
>>>of dealers in the Manchester area though :)
>>>
>>OK my turn to actually *be* dumb... What is the difference between the
>>OEM and the retail version?
>
>be 450-able, whether through design or accident I don't know.
>
In that case I can't understand the justification in requesting a
swap/refund, the retailer simply tells you "there is no difference now
go away, I know what you are up to you little over-clocker you... I
don't want you sort in here, gives people ideas about having software
run at a reasonable speed and I won't have it! Not from the processors I
sell. Tsk, kids of today, don't know they're born... I remember when a
computer was something you only saw in the movies.....etc"
Sorry, too much booze. I seriously can't see why the retailer would
swap/refund on these grounds. I am just curious if someone would care to
enlighten me.
--
Justin C by the sea. mailto:9907....@j-catter.demon.co.uk
Usenet rule 1: Don't post while under the influence of alcohol.
BlueBoy
excess heat from high vcore or simply a badly fitted heatsink. If there was
there would be no point doing it as it would cost money. The objection I
have is a moral one. If you choose to run your processor out of spec then
you should agree to the fact that your warranty will be void. Intel is
spiteful and if they see a large amount of money being payed out for
warranty replacements they will simply put a frequency lock in the chips.
If you fry a chip take the consequences for your own actions.
--
Blueboy aka Happy Fragger
while(dead_horse)
{
flog();
}
I support the campaign for unmetered telecommunications, do you?
http://www.unmetered.org.uk/
Due to a recent influx of spam my address is witheld from all but the most
worthy. it begins with marek, follows with the definite, indefinite or vague
place and concludes with Intercreations.com
Mark Powell
JustinC <9907....@j-catter.demon.co.uk> wrote:
>Sorry, too much booze. I seriously can't see why the retailer would
>swap/refund on these grounds. I am just curious if someone would care to
>enlighten me.
Worked for me. I've now got an OEM PIII 450 that overclocks to
581MHz (129x4.5) at 2.1v. It will overclock to 600MHz with the L2
cache off, but I'm not interested in that.
For interests sake the markings are:
On plastic case:
09210493-0269 COSTA RICA
imc '98 SL35D
On cache:
M5M5V2236BGP
913BE0G-4H
250MHz
On CPU:
3920A046
Zeus
and make of memory?
Mark Powell <ma...@salfrd.ac.uk> wrote in message
news:7os5q2$h9k$1...@plato.salford.ac.uk...
Mark Powell
ABit BX6 rev. 2
>and make of memory?
PC133 Winbond W986408BH-75
Before I got hold of the PC133 RAM, I did try my brother's retail
450 at 581MHz and that too was stable with RAM rated at "PC125".
That was LGS GM72V66841CT-7J. All RAM was purchased from www.scan.co.uk,
next to the Reebok Stadium.