Results - 400MHz FSB Celeron in Dimension 8300
William R. Walsh
I'm not sure what Dell was thinking when they designed the particular
heatsink used in this system, but it was a chore to get those little green
retention arms to come loose. After nearly cutting myself on the heatsink
fins, I got them out. Everything else was a piece of cake.
After slapping the Celeron in, applying some fresh heatsink compound and
putting the system back together, I turned it on. Nothing dramatic
happened--it just worked. System Setup reports a 2.2GHz Celeron CPU with a
400MHz FSB speed.
What's so startling is how much *faster* the response time of the system is.
Applications finally come up almost instantly. Windows move around much
faster. In high computing load situations, it is obvious that the Celeron
isn't running anywhere near as quickly, and the smaller L1 cache is also
noticeable. I'll take that over a draggy response, though.
Also notable is the vastly reduced heat output. Even at 100% CPU load, the
fans don't hit "warp speed" any more.
Okay, I'll try a Northwood P4 and see what that does. Clock speed/FSB
suggestions?
William
William R. Walsh
> load, the fans don't hit "warp speed" any more.
The power being drawn is also much lower. With the Prescott P4 in place, I
could easily hit 250-270 watts per the meter built into my UPS. The Celeron
at 100% load is pulling about 197, and that includes draw from the monitor,
printer, speakers and scanner.
William
S.Lewis
"William R. Walsh" <newsg...@idontwantjunqueemail.walshcomptech.com>
wrote in message news:sTkDj.79146$yE1.16320@attbi_s21...
Congratulations on that interesting find.
3.0GHz or 3.2GHz Northwood should work just fine. Both are 800mhz FSB with
512 cache.
Get some fresh thermal compound if you need it.
So is it your opinion (right now) that the 3.4GHz Prescott has been the
issue the entire time?
Timothy Drouillard
what happens just for grins.
What are the odds that there was improper or inadequate thermal transfer
with the original 3.4 and the heatsink?
At the least it would confirm that the previous performance is indeed caused
somehow by the 3.4 chip.
"William R. Walsh" <newsg...@idontwantjunqueemail.walshcomptech.com>
wrote in message news:sTkDj.79146$yE1.16320@attbi_s21...
William R. Walsh
>I would be real tempted to re-install the 3.4 using fresh compound
> and see what happens just for grins.
I already did. It made no difference. Contact between the heatsink, compound
and 3.4 seemed to be very good. I cleaned the heatsink completely anyway,
put some fresh compound down and tried again.
I also tried a larger power supply (450 watt over the 300 or so watt unit
that Dell installed) in case the system was somehow being starved for power
and running slowly as a result. That made no difference either, although the
larger supply ran somewhat cooler.
William
Timothy Drouillard
Well it was worth a shot anyway if for no other reason but to confirm the
3.4 is 'sluggish'.
I'm assuming you've already used something like Intel's CPU id utility
simply to confirm the ID of the 3.4?
(dumb question, but I had to ask..)
"William R. Walsh" <newsg...@idontwantjunqueemail.walshcomptech.com>
wrote in message news:nRlDj.26031$TT4.16579@attbi_s22...
wm_w...@hotmail.com
> I'm assuming you've already used something like
> Intel's CPU id utility simply to confirm the ID
> of the 3.4?
I did not use an Intel utility to determine the ID of the CPU. I used
the CPU-Z tool, which has always worked well for me in the past. (It
is also not Intel specific.)
While the chip itself doesn't say "Prescott" on it (nor did I expect
it to), the printing on top clearly identifies it is an Intel Pentium
4 at 3.40GHz clock speed.
> (dumb question, but I had to ask..)
Never hurts to have all the facts.
William
wm_w...@hotmail.com
> Get some fresh thermal compound if you need
> it.
I have a tube of thermal compound from Radio Shack handy. There's
nothing special about it, but it has always worked well when I needed
it--mostly for voltage regulators, 486 CPUs (!!!) and things like
that. It seems to work fine in newer computers as well. (I'm not a big
believer in fancypants heatsink compounds like Arctic Silver. Simple,
cheap and effective will do.)
> So is it your opinion (right now) that the
> 3.4GHz Prescott has been the issue the entire
> time?
Time will tell. I'm hopeful, but I'm also wary of the potential for
psychological influence to make things seem better than they really
are. (What a mouthful!) I'll run this system for a while and see how
it behaves to make sure that there really is more of an improvement
and that I'm not imagining things.
Besides, waiting will hopefully make the entry price for a Northwood
P4 go down. :-)
William (can you tell I'm cheap?)
HDRDTD
ISTR that some time ago (and maybe still today), there were counterfit chips
being sold, that someone in japan was taking CPU's, then remarking the
printing on top so they could be sold as faster chips.
<wm_w...@hotmail.com> wrote in message
news:e9d064d4-83b4-490c...@x41g2000hsb.googlegroups.com...
S.Lewis
<wm_w...@hotmail.com> wrote in message
news:6b622e97-dd09-46e7...@m34g2000hsc.googlegroups.com...
You're not "cheap". I prefer to think of it as being "practical". heh
I'm the same way.
Have you priced the CPU's over at www.pricewatch.com ?
I can tell you that (perception-wise) I can't really tell any difference
between the 3.0GHz and 3.2GHz Northwoods. (Since I didn't own them
simultaneously I can't give you apples/apples PCMark2002 or other numbers).
BUT.....I can absolutely tell you that both Northwoods beat the pants off my
3.0GHz Prescott with that benchmark.
Good luck.....
Stew
S.Lewis
"HDRDTD" <HDR...@comcast.net> wrote in message
news:bsednYA0CdWUCkPa...@giganews.com...
> As long as CPU-Z identified it correctly.
>
> ISTR that some time ago (and maybe still today), there were counterfit
> chips being sold, that someone in japan was taking CPU's, then remarking
> the printing on top so they could be sold as faster chips.
>
Judging from my own CPU/system, I think the Prescotts are just that bad
(generally).
Heat......heat.........and more heat.
wm_w...@hotmail.com
> As long as CPU-Z identified it correctly.
It's the latest version available...the one that was recently updated
to cover Intel's "Skulltrail" platform and matching processor. (Maybe
there's one newer than that by now, but it should still cover a
Pentium 4 system that was put together in October 2004.)
> ISTR that some time ago (and maybe still today),
> there were counterfeit chips being sold, that
> someone in japan was taking CPU's, then remarking
> the printing on top so they could be sold as
> faster chips.
It wouldn't surprise me at all to hear that. What would surprise me is
if a generally reputable vendor such as Dell was buying chips like
that and including them in their systems. This system came with the
already mentioned Prescott processor when it was built by Dell.
William
Ben Myers
middle man... Ben Myers