Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

Hi,

My second post ever to the group, and another question. Maybe some
day, I'll have some answers, instead.

I'm currently running a T-Bred B, 2400+ 133MHz, and just upgraded to a
gig of DDR 400 RAM. I had hoped to do the "wire trick", and unlock the
multi on the T-Bred, so I could run the RAM at its full potential.

Woe is me. After all that fiddling, I learn that I have a newer
"superlocked" T-Bred. (Its date code is 0417) I can now change multi
in BIOS but, it has zero effect.

So... I'm shopping for a replacement in the form of an XP-M, which will
not be multiplier locked (none of them are, right?). In my travels I
found this (AXMA3000FKT4C):

http://www.newegg.com/app/ViewProd [...] 482&depa=1

At first glance it looks like it would be ideal... but then I saw
someone say that the chip won't run in a desktop Socket-A:

http://forums.devhardware.com/archive/t-29305

(It's down at the bottom of the page, a comment by "Sokur" )

I haven't found any other reference to such a thing on the web. The
reviewers at NewEgg say they have it running in their desktops but,
heh... NewEgg customer reviews? I don't put much stock in them. Can
anyone tell me if that chip really is incompatible with a Socket-A ATX
board? Better yet, have a link to a reasonably authoritative web page
on the matter?

If I may be so presumptuous... one more question. If that chip *will*
run, it will surely be a hot one, for a "mobile". (2.2GHZ@1.65V)
As I'm obviously interested in running a higher bus speed than
its rating, and probably will push it as much as it will tolerate
with air-cooling... do you think it would outperform a 35W 2400+

http://www.newegg.com/app/ViewProd [...] 439&depa=1

if both were running at max overclock? I've read good things in this
group about the 35W 2400+ and overclocking. And it *is* cheaper.

Thanks for any input...

Jonathan

--
Don't just hit reply. Email address is broken. Thank
your friendly neighborhood spammer. Email replies to:
user: jnthn1 domain: earthlink<dot>net

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

newegg doesn't publish some of the negative reviews, but you can peruse the
ones they have and gets lots of ideas. They published a couple of mine.

I'm running the 2200+ mobile 35w. See what they say, it may be enough for
you. I can run it as a 3000+ at 166/333, 13x, at the default voltage, which
my asus a7v880 set at 1.58v (it says - it has weird sensors). Above that I
have to give it more volts.

It's just a Barton, they say, same as the desktop chip.

I'm using CrystalCPUID and set it to run at 500 mhz and, when cpu demand
goes up, cycle through 1250 to 2083. It tops out at 2083 because the maximum
multiplier set in the bridges is 12.5x, and I'm running standard 166/333 bus
and memory, in synch.

Anyhow, it sure runs cool at 500mhz. And not so bad at 2150 (3000+). (My
motherboard bios can take the multiplier up past the maximum set in the
bridges).

This kind of explains it:

http://www.cpuheat.wz.cz/html/AXP_ [...] iplier.htm

CrystalCPUID:

http://crystalmark.info/soft/CrystalCPUID4-e.html

--
Ed Light

Smiley :-/
MS Smiley :-\

Send spam to the FTC at
uce@ftc.gov
Thanks, robots.

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

On Tue, 21 Dec 2004 15:22:22 -0800, Ed Light
<nobody@nobody.there> appears to have said:

> newegg doesn't publish some of the negative reviews, but you can peruse the
> ones they have and gets lots of ideas. They published a couple of mine.

Oh, I'm not saying I disregard the reviews at NewEgg altogether. Not
everyone knows what they're talking about though, and I've seen lots
of reviews for this or that product at NewEgg in which the reviewer is
actually talking about an altogether different product! In fact, there's
a review on the description page for the XP-M 3000+ in which the guy
says "this 2600 does 2.4ghz at 200*12". Not too reliable if you know
what I mean.

> I'm running the 2200+ mobile 35w. See what they say, it may be enough for
> you. I can run it as a 3000+ at 166/333, 13x, at the default voltage

Enough? There's *never* enough! Seriously, though... what I'm looking
for is something that I can reasonably hope to run at 2.4GHz or higher.
Higher would be better. If I do get something that's rated lower than
the 3000+, I think it would be the XP-M 2400+, as I have seen lots of
folks say that they've gotten 2.4GHz and better with that one (and
its 35W power rating makes it seem to me it would be a much better
overclocker than the 45W 2600+ listed at NewEgg).

Then again, 2.4 GHz would only be a 9% overclock for the 3000+, while it
would be a whopping 22% overclock for the 2400+.

I just wonder if the XP-M 3000+ for DTR will work in my KT6V Socket-A
mainboard, (I see someone else at NewEgg has written a review saying he
has it in a GA-7N400 board. I begin to really believe that the guy who
said it wouldn't be compatible with Socket-A ATX was blowing smoke.) and
if I can reasonably expect it to outperform the XP-M 2400+ if both were
pushed to their respective max clocks. Anyone have experience with that
DTR chip?

This one, I mean:
http://www.newegg.com/app/ViewProd [...] 482&depa=1

Thanks,
Jonathan

--
Don't just hit reply. Email address is broken. Thank
your friendly neighborhood spammer. Email replies to:
user: jnthn1 domain: earthlink<dot>net

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

"Jonathan A." <me@privacy.net> wrote
> Oh, I'm not saying I disregard the reviews at NewEgg altogether. Not
> everyone knows what they're talking about though, and I've seen lots
> of reviews for this or that product at NewEgg in which the reviewer is
> actually talking about an altogether different product! In fact, there's
> a review on the description page for the XP-M 3000+ in which the guy
> says "this 2600 does 2.4ghz at 200*12". Not too reliable if you know
> what I mean.

But when you see about 5 people getting 2.4 then you know it's likely you
could.


>
>> I'm running the 2200+ mobile 35w. See what they say, it may be enough for
>> you. I can run it as a 3000+ at 166/333, 13x, at the default voltage
>
> Enough? There's *never* enough! Seriously, though... what I'm looking
> for is something that I can reasonably hope to run at 2.4GHz or higher.

If you read these reviews, you can see that 2.4 with the 2200+ is pretty
common! It's the same chip as the 3000+. But do they test each one and grade
them? Who knows.

http://secure.newegg.com/app/Custr [...] 19-103-440


> Higher would be better. If I do get something that's rated lower than
> the 3000+, I think it would be the XP-M 2400+, as I have seen lots of
> folks say that they've gotten 2.4GHz and better with that one (and
> its 35W power rating makes it seem to me it would be a much better
> overclocker than the 45W 2600+ listed at NewEgg).

That one does look good.


>
> Then again, 2.4 GHz would only be a 9% overclock for the 3000+, while it
> would be a whopping 22% overclock for the 2400+.

Again, they're the same chip, so you can check those reviews to see if
people are getting the same sort of overclocks, unless the 3000+ doesn't
have enough reviews yet.

> I just wonder if the XP-M 3000+ for DTR will work in my KT6V Socket-A

What does DTR stand for?



> mainboard, (I see someone else at NewEgg has written a review saying he
> has it in a GA-7N400 board. I begin to really believe that the guy who
> said it wouldn't be compatible with Socket-A ATX was blowing smoke.)

You do see user reviews by people who are misinformed, everywhere. But you
never know. I wondered about how most people were saying Samsung hd's are
quiet, and some complaining about the noise. Then it turnes out that some
have Nidec motors - quiet - and some have JVC - loud - and it's kind of
random which one you get.

Here's another article on overclocking the XP-M's:

http://www.tweakers.com.au/article [...] /page1.asp


--
Ed Light

Smiley :-/
MS Smiley :-\

Send spam to the FTC at
uce@ftc.gov
Thanks, robots.

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

On Wed, 22 Dec 2004 14:51:57 -0800, Ed Light
<nobody@nobody.there> appears to have said:

>
> "Jonathan A." <me@privacy.net> wrote
>> Enough? There's *never* enough! Seriously, though... what I'm looking
>> for is something that I can reasonably hope to run at 2.4GHz or higher.
>
> If you read these reviews, you can see that 2.4 with the 2200+ is pretty
> common! It's the same chip as the 3000+. But do they test each one and grade
> them? Who knows.

There's another thing. I would certainly hope that AMD would grade the
cores before giving them their rating. Then again, this page:

http://www.ciao.co.uk/AMD_AthlonXP [...] ew_5358799

suggests that they may be selling cores of the same quality, simply
repackaged, and rebadged for marketing purposes. That's a relatively old
page, but there may be something to it.

>> I just wonder if the XP-M 3000+ for DTR will work in my KT6V Socket-A
>
> What does DTR stand for?

DeskTop Replacement... from what I can gather, it's a "standard" for
making notebooks for people who don't want a laptop as much as they want
a smaller desktop solution. The power requirements are different... less
emphasis on power saving.

That's what makes it tough to decide between the 3000+ DTR and the
plain vanilla 2400+ XP-M, AIUI. The DTR chip runs a higher standard
clock, but has the disadvantage of also having much higher voltage
requirements "out of the box". It runs at 1.65V at its rated clock,
while the 2400+ runs at 1.35V. I may not be an expert, but it don't take
a rocket scientist to figure that (less voltage)=(less wattage)=(less
heat). Less heat is good.

> Here's another article on overclocking the XP-M's:
>
> http://www.tweakers.com.au/article [...] /page1.asp

Thanks for the link. Looks like lots of good info there.

Jonathan

--
Don't just hit reply. Email address is broken. Thank
your friendly neighborhood spammer. Email replies to:
user: jnthn1 domain: earthlink<dot>net

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

"Jonathan A." <me@privacy.net> wrote
> That's what makes it tough to decide between the 3000+ DTR and the
> plain vanilla 2400+ XP-M, AIUI. The DTR chip runs a higher standard
> clock, but has the disadvantage of also having much higher voltage
> requirements "out of the box". It runs at 1.65V at its rated clock,
> while the 2400+ runs at 1.35V. I may not be an expert, but it don't take
> a rocket scientist to figure that (less voltage)=(less wattage)=(less
> heat). Less heat is good.

To run the 2400+ at 3000+ you would have to raise the voltage. Probably to
1.55 or 1.6. Conversely, the 3000+ might run ok as a 2200+ or 2400+ at
1.35.

Since desktop motherboards read a 1.35v cpu as 1.52 or 1.58, to actually run
at 1.35 you'd need a motherboard that could set that in the bios. Some are
listed at http://www.silentpcreview.com/
among lots of articles on undervolting, also
http://www6.tomshardware.com/cpu/20041001/index.html.


--
Ed Light

Smiley :-/
MS Smiley :-\

Send spam to the FTC at
uce@ftc.gov
Thanks, robots.

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

I ran a test just now

http://members.cox.net/edlight5/amd.gif

I started out with my 2200+ mobile, always at an indicated 1.58v, at
2150mhz, a 3000+, bus at 166/333. All temps under the socket, so 5-10C
cooler than in the core. I got 41C.

Next, ran prime 95 torture on hottest choice. Up to 48C.

Then, let it cool all the way down at 500mhz, to 32C.

Finally, put it on 1250mhz. Wasn't rising, so started p95 again. Rose to
39C.

As I type this it's back on the CrystalCPUID routine, running at 500, and
when needed going through 1250 to 2083. It's at 31C.


--
Ed Light

Smiley :-/
MS Smiley :-\

Send spam to the FTC at
uce@ftc.gov
Thanks, robots.

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

"Ed Light" <nobody@nobody.there> wrote in message
news9wyd.469$yW5.309@fed1read02...
>I ran a test just now
>
> http://members.cox.net/edlight5/amd.gif

BTW the lower line is the system temp. Note how it's not affected by the cpu
because of the ducting the cpu heat out the back.

I used this thing:

http://store.yahoo.com/casecooler/suovairdu.html

Had to mount the case fan on the back of the case to fit it (fan grille cut
out). No cpu fan. Case fan 80mm panaflo high speed at 2350 rpm on a zalman
fanmate. Heatsink is Thermal Integration TI-V77L. Arctic Ceramique.

--
Ed Light

Smiley :-/
MS Smiley :-\

Send spam to the FTC at
uce@ftc.gov
Thanks, robots.

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

On Thu, 23 Dec 2004 01:25:04 -0800, Ed Light
<nobody@nobody.there> appears to have said:

> I ran a test just now
>
> http://members.cox.net/edlight5/amd.gif
>
> I started out with my 2200+ mobile, always at an indicated 1.58v, at
> 2150mhz, a 3000+, bus at 166/333. All temps under the socket, so 5-10C
> cooler than in the core. I got 41C.
>
> Next, ran prime 95 torture on hottest choice. Up to 48C.
>
> Then, let it cool all the way down at 500mhz, to 32C.

Well... my current XP-2400+ runs about 58C under burnk7 with the stock
AMD cooler. I saw your other post about the duct thing. My case has a
"chimney" for airflow to the CPU. I've often wondered if I could get
better temps by somehow making it airtight.

Anyhow, if the 2200+ will run 2150MHz at 1.58v, I'd bet I can get
that 2400 to do 2.4GHz at 1.7v or lower. (With better cooling than I
currently have, of course.) Thanks for the input. I'm thinking more and
more that, whatever I do, the extra $50.00 for the 3000+ just wouldn't
be justified.

ciao,
Jonathan

--
Don't just hit reply. Email address is broken. Thank
your friendly neighborhood spammer. Email replies to:
user: jnthn1 domain: earthlink<dot>net

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

Jonathan,

Let us know how it turns out!


--
Ed Light

Smiley :-/
MS Smiley :-\

Send spam to the FTC at
uce@ftc.gov
Thanks, robots.

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

On Thu, 23 Dec 2004 19:11:37 -0800, Ed Light
<nobody@nobody.there> appears to have said:

>
> Jonathan,
>
> Let us know how it turns out!

It'll be a few weeks yet, but I'll have to drop by and show off
the new hot rod.

Thanks for the pointers.

Jonathan

--
Don't just hit reply. Email address is broken. Thank
your friendly neighborhood spammer. Email replies to:
user: jnthn1 domain: earthlink<dot>net