Old processor opinions

AMD K6s usually could overclock at least 10% without
problems, so 100 x 5 be well within that range and the
other busses would be back up to normal, 66 and 33.

"CK" <CatKiller.RemoveThis@nospamplease.Bigfoot.com> wrote in message
news:ca7fg6$l38$1@newsg4.svr.pol.co.uk...
 >
 > "CK" <CatKiller.RemoveThis@nospamplease.Bigfoot.com> wrote in message
 > news:ca7621$73k$1@news8.svr.pol.co.uk...
  > > Just got a processor for my g/f's computer. Managed to find a K6-2
475.
 > This
  > > is 95 x 5, apparently, which the motherboard can do - it's a no-name
MVP3,
  > > if you're interested - but according to the manual using 95MHz FSB
will
 > drop
  > > the other bus speeds out of spec - AGP to 60, PCI to 30 and,
crucially,
  > > memory to 63. wtf?
  > >
  > > Anyway, do you think that I should run the processor at 95 x 5,
slowing
  > > everything else down, or should I have a try at 100 x 4.5 or 5 to
start
  > > with? It's five or six jumpers to set the FSB and multiplier, so I'd
 > rather
  > > not have to do it too many times. Fiddly little things, I'm glad we
don't
  > > have to rely on those any more.
  > >
 > Well I've put it in as 5 x 95, and it seems to be working OK. Sound
and
 > video are working fine. I'll try the networking later and run
memtest86 for
 > a bit to make sure the memory's happy. I might try o/cing it to 5 x
100 at
 > some point in the future for the increased memory and video
performance, but
 > tbh the 475 itself is probably enough of an increase over her K6 300
that it
 > probably won't be a priority for a while. Especially with all those
pesky
 > jumpers.
 >
 > While I had the case open I noticed a cracked capacitor between the
 > processor and memory slots. The computer seems to be working fine
 > regardless, so I'm not too worried, but has anyone got any idea what
this
 > would be used for?
 >
 > TIA
 >
 > CK
 >
 ><!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Old processor opinions 

On Wed, 9 Jun 2004 17:55:04 +0100, "CK"
<CatKiller.TakeThisOut@nospamplease.Bigfoot.com> wrote:


 >Well I've put it in as 5 x 95, and it seems to be working OK. Sound and
 >video are working fine. I'll try the networking later and run memtest86 for
 >a bit to make sure the memory's happy. I might try o/cing it to 5 x 100 at
 >some point in the future for the increased memory and video performance, but
 >tbh the 475 itself is probably enough of an increase over her K6 300 that it
 >probably won't be a priority for a while. Especially with all those pesky
 >jumpers.
 >
 >While I had the case open I noticed a cracked capacitor between the
 >processor and memory slots. The computer seems to be working fine
 >regardless, so I'm not too worried, but has anyone got any idea what this
 >would be used for?
 >
 >TIA
 >

Early on in the production of the K6-2 AMD was having problems getting
sufficient quantity of chips that'd do 100FSB... wasn't the total
clockspeed that was the problem, it was the higher FSB interface speed.
If your CPU isn't an early model the odds are high it'll do 5 x 100 fine,
perhaps it would need a voltage increase though, they generally run fine
up to at least 2.5V, but likely if it defaulted at 2.2V, 2.3V is enough...
you'd just have to try it and see... relatively speaking, the difference
is minor, maybe not worth changing it at all.

The capacitor is a problem though, you should replace the motherboard or
at least that capacitor ASAP. There might also be several other locations
on board using same capacitor, many of them near failure.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Old processor opinions 

 > Early on in the production of the K6-2 AMD was having problems getting
 > sufficient quantity of chips that'd do 100FSB... wasn't the total
 > clockspeed that was the problem, it was the higher FSB interface speed.

I thought that that was probably the case. In particular I wondered whether
it was the on-chip cache that might have the problems with the speed. I
remember from a year or so before this chip was made that the Celerons
became really popular with overclockers because it didn't have the on-chip
cache, which meant that the cache didn't hold back the overclock. It was a
while ago, though, so it's all a bit hazy...

 > If your CPU isn't an early model the odds are high it'll do 5 x 100 fine,
 > perhaps it would need a voltage increase though, they generally run fine
 > up to at least 2.5V, but likely if it defaulted at 2.2V, 2.3V is enough...
 > you'd just have to try it and see... relatively speaking, the difference
 > is minor, maybe not worth changing it at all.
 >
AMD put out both 2.4V and 2.2V parts. From a bit of reading through the
overclocker archives it seems that the 2.2V part was the better for
stability at higher speeds. This one, unfortunately, is the 2.4V part.

 > The capacitor is a problem though, you should replace the motherboard or
 > at least that capacitor ASAP. There might also be several other locations
 > on board using same capacitor, many of them near failure.

While I agree that the motherboard is probably on its last legs, it might
not be too much of a problem. Hopefully by the end of the year I'll have
upgraded my computer, which will mean that my other half gets my cast-offs
and her (now looking a bit flakey) motherboard will be surplus to
requirements. If it holds out till then, no problem. If it doesn't then it
means that she gets the newer computer and I get to hang on to this one for
a bit longer. Either way, the £20 that I spent yesterday on the ~ 60%
performance increase for the meantime seems like a reasonable investment.

Thanks for the reply.

CK<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Old processor opinions 

CK wrote:
  >> Early on in the production of the K6-2 AMD was having problems
  >> getting sufficient quantity of chips that'd do 100FSB... wasn't the
  >> total clockspeed that was the problem, it was the higher FSB
  >> interface speed.
 >
 > I thought that that was probably the case. In particular I wondered
 > whether it was the on-chip cache that might have the problems with
 > the speed. I remember from a year or so before this chip was made
 > that the Celerons became really popular with overclockers because it
 > didn't have the on-chip cache, which meant that the cache didn't hold
 > back the overclock. It was a while ago, though, so it's all a bit
 > hazy...
 >
  >> If your CPU isn't an early model the odds are high it'll do 5 x 100
  >> fine, perhaps it would need a voltage increase though, they
  >> generally run fine up to at least 2.5V, but likely if it defaulted
  >> at 2.2V, 2.3V is enough... you'd just have to try it and see...
  >> relatively speaking, the difference is minor, maybe not worth
  >> changing it at all.
  >>
 > AMD put out both 2.4V and 2.2V parts. From a bit of reading through
 > the overclocker archives it seems that the 2.2V part was the better
 > for stability at higher speeds. This one, unfortunately, is the 2.4V
 > part.
 >
  >> The capacitor is a problem though, you should replace the
  >> motherboard or at least that capacitor ASAP. There might also be
  >> several other locations on board using same capacitor, many of them
  >> near failure.
 >
 > While I agree that the motherboard is probably on its last legs, it
 > might not be too much of a problem. Hopefully by the end of the year
 > I'll have upgraded my computer, which will mean that my other half
 > gets my cast-offs and her (now looking a bit flakey) motherboard will
 > be surplus to requirements. If it holds out till then, no problem. If
 > it doesn't then it means that she gets the newer computer and I get
 > to hang on to this one for a bit longer. Either way, the £20 that I
 > spent yesterday on the ~ 60% performance increase for the meantime
 > seems like a reasonable investment.

60% CPU speed increase doesn't equal 60% performance increase, sorry to say.
More like 15-20%. (From experience).
--
~misfit~


---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.700 / Virus Database: 457 - Release Date: 6/06/2004<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Old processor opinions 

<snip>
  > > Either way, the £20 that I
  > > spent yesterday on the ~ 60% performance increase for the meantime
  > > seems like a reasonable investment.
 >
 > 60% CPU speed increase doesn't equal 60% performance increase, sorry to
say.
 > More like 15-20%. (From experience).
 > --
You could well be right. But from scanning some of the old articles on the
K6 series, people were very much of the opinion that the K6-2 performed a
lot better clock-for-clock than the original K6. Haven't looked at the
architectural changes yet to establish whether this is feasible or not,
though. Also, she had reasonable other parts in the system - 128MB of RAM
and a GeForce 256 on Win98SE with a 5,400 rpm hard drive. The processor was
a clear bottleneck with regards to performance in most things that she uses
the computer for. So most of the increase in processor speed should go
towards increasing the performance overall, since that was the main limiting
factor. Her chances are good, I think, of getting a noticably better
experience from using her computer as a result of this upgrade and if she
doesn't, hey, it's only £20...

Thanks for the comment.

CK<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Old processor opinions 

"CK" <CatKiller.TakeThisOut@nospamplease.Bigfoot.com> wrote in message
news:ca7fg6$l38$1@newsg4.svr.pol.co.uk...
 >
 >
 > While I had the case open I noticed a cracked capacitor between the
 > processor and memory slots. The computer seems to be working fine
 > regardless, so I'm not too worried, but has anyone got any idea what this
 > would be used for?
 >
Usually the same reason they are used at each IC on a well-designed circuit
board, to minimize ground bounce. Just replace it...its an easy job. Also
note that a lot of products work either without them or without the value
calculated for each and every application (IC at the board level, board at
the system level), so it isn't the end of the world (probably) if you choose
to do nothing, but if you have any weird behavior in the future remember
that capacitor.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Old processor opinions 

"CK" <CatKiller RemoveThis @nospamplease.Bigfoot.com> wrote:
 >Just got a processor for my g/f's computer. Managed to find a K6-2 475. This
 >is 95 x 5, apparently, which the motherboard can do - it's a no-name MVP3,
 >if you're interested - but according to the manual using 95MHz FSB will drop
 >the other bus speeds out of spec - AGP to 60, PCI to 30 and, crucially,
 >memory to 63. wtf?
 >
 >Anyway, do you think that I should run the processor at 95 x 5, slowing
 >everything else down, or should I have a try at 100 x 4.5 or 5 to start
 >with? It's five or six jumpers to set the FSB and multiplier, so I'd rather
 >not have to do it too many times. Fiddly little things, I'm glad we don't
 >have to rely on those any more >TIA >CK

Holy Christ!! I thought I was becoming computer savvy, but I am
totally completely lost with this thread! My best interpretation is
g/f's = girlfriend's. But even then I first though "grandfather's".

Big Mac<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Old processor opinions 

On Thu, 10 Jun 2004 12:25:14 +1200, "~misfit~"
<misfit61nz.DeleteThis@yahoomung.co.nz> wrote:

 >60% CPU speed increase doesn't equal 60% performance increase, sorry to say.
 >More like 15-20%. (From experience).

huh, you came to same mine conclusions (also mine experiences) in
average ... so I am asking myself lately, why people still OC their
these days powerful (also in terms of heat!) CPUs. Practicaly not much
of difference (just generating more heat than gain, except in
Benchmarks-synthetic ones mostly) ... in the "old" days (few years ago
only!) there was more gaining than these days in "real life" average
use of comps ... well, some applications really benefit of higher
clock if someone uses them frequently (like encoding stuff) ...

Faster HDs & much Ram really makes the difference these days ...

.... when you pass 1GHz & 133fsb & have 0,13u tech Cpu IMHO is enough
for more than 95% of uses .... other is IMHO more marketing strategy
to sell stuff ... (a bit of exeption not to gain much in watts
consumed - mobile Athlons & P-M ´s on desktop MoBo ...)

BTW, did you know: there (in Japan) was released (a month ago or so) a
"standard" desktop MoBo for - P-M !!! ... this would be a nice
uppgrade (latest technology w/o additional heat produced!)

When this would spread more around (availability) in a year or so,
then we could think seriously to uppgrade (& still mantain our
machines quiet ... )


--
Regards, SPAJKY ®
& visit my site @ <a style='text-decoration: underline;' href="http://www.spajky.vze.com" target="_blank">http://www.spajky.vze.com</a>
"Tualatin OC-ed / BX-Slot1 / inaudible setup!"
E-mail AntiSpam: remove ##<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Old processor opinions 

  > > The capacitor is a problem though, you should replace the motherboard or
  > > at least that capacitor ASAP. There might also be several other
locations
  > > on board using same capacitor, many of them near failure.
 >
In investigating further, it turns out that the L2 cache has broken. It's
quite near the cap that I noticed had gone, and some others appear to be
bulging.

I'll get them replaced as soon as the guy with a soldering iron answers the
phone...<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Old processor opinions