View Single Post
  #2  
Old June 24th 11, 07:56 AM posted to alt.comp.hardware.overclocking
Paul
external usenet poster
 
Posts: 13,364
Default e2180, Abit ip-35e

area51 wrote:
I've never overclocked, but when I purchased this PC, I knew that it
would be good to overclock. Also has an excellent CPU fan.

Intel e2180
Abit IP-35e
4GB DDR2-680 (according to BIOS)
CPU Core Voltage says 1.325 OV
CPU VTT Voltage says 1.200v

Can anyone suggest what to put in for values?
I'm only interested in going up to about 3.0GHz or less.
I've read very little, and I don't understand what people mean when they
say, "stock" or "up a notch". So plain English please.

Or a website that explains this. I've googled overclocking, but there's
too much information out there, when all I should have to do is change a
few settings.

Thanks


Intel Pentium Processor E2180
(1M Cache, 2.00 GHz, 800 MHz FSB)
65 nm (so you can use a bit extra voltage).

The 45nm processors are limited to 1.4V long term. The 65nm can
take more than that, but I don't know how much more. I boosted my
65nm processor by 0.100 volts and got a 33% overclock on it. But
that wasn't enough voltage to make it stable 24/7. I dropped
back to stock, after doing a quick test.

Going from 2GHz to 3GHz is ambitious. It can be done, but
is subject to the limits of the individual processors (did you
get a "good" one or not).

Your processor has an "external clock" setting. It will be 200Mhz.
200*4 = quad pumped FSB800. That's where the FSB speed rating
comes from. It's the CPU input clock times four.

Your processor is locked with a 10X multiplier. 200*10 = 2000MHz core speed.

To hit 3GHz, you need 300*10 = 3000MHz.

The same input clock, also affect the memory. The BIOS has a
"DRAM Speed (CPURAM)" setting. It sets a ratio between the
input clock, and the DRAM clock. As you crank your CPU, you're
also accidentally cranking the DRAM. You have to change
the DRAM Speed (CPURAM) setting to compensate.

For example, say CPURAM was 5:4. That ratio is 1.25X.
Now, say I was overclocking the CPU by 25 percent. To compensate,
if I set CPURAM to 1:1, that is a reduction from my previous
setting by a factor of 25%. Then, the input clock "bump" up of 25%,
followed by the reduction in DRAM setting from 5:4 to 1:1, leaves
the DRAM at its original speed.

OK, so first, set a reasonable target. Say a 33% overclock. First,
work out the necessary DRAM compensation setting. Drop the ratio
so the RAM won't run too fast when you reach your max predicted
speed. (Note - I'm giving this simplified method, so I don't
have to write a section about CAS and friends.)

Then, "scan upwards" and watch the system stability. Don't use
your Windows hard drive for this. Mistake. Use a memtest86+ floppy,
or a Linux LiveCD, as a test mechanism. They can't be corrupted
like Windows can. You don't want to boot Windows on top of flaky
hardware, because your registry can end up corrupted.

So, at stock, you're at 200MHz. Make your DRAM correction (the
one that would be accurate, if you made it all the way to a 33%
overclock). Now, set the input clock to 205MHz. Boot and
test stability as you see fit. Now, say the computer crashes
at 205. Now, go to Vcore. Say Vcore was 1.325V. Bump Vcore to 1.350V
and retest. Is it stable ? OK, now increase input clock to 210MHz,
boot your Linux CD and test. Is it stable ? If not, increase voltage
to 1.375V.

By collecting that info, you'll see a "ramp".

Freq | __/
| __/
|/
+-------- VCore voltage

The ramp allows you to estimate what it'll take to go higher
and higher.

Some processors have a brick wall like this. My old Pentium 4
did this.

Freq | ___________
| __/
|/
+-------- VCore voltage

I could increase VCore all day long, and I'd get no more
frequency from it. If you go too high, you'll burn it out.
And the motherboard design, and BIOS, may not actually
give good advice on how much voltage is safe. This
is... all part of the "fun".

I did 33% with my 2.6GHz Core2, with 0.1V of Vcore bump,
so I would expect you could get at least that much without
damage. Any more than that, do some more research.

A good place to look, is the "Customer Review" section
for the processor in question (Feedback tab). Note that
each processor is different, so different people get
different results. Expect a statistical spread. Intel
bins the processors (sorts them), so you don't always
get a heroic one.

"E2180 2.00 GHz FSB800"
http://www.newegg.com/Product/Produc...82E16819116052

"Overclocks to 2.8 for me. Wont go up any higher due to
instability. Got a bad chip I guess. running 2.8 @ 1.28 volts
right now @ 45c load. More voltage wont work with this chip
though " --- [Guy hit a brick wall... too bad, so sad]

"Easily pushed to 2750 MHz with no voltage increase.
Starts to get a little hot"

"I like it. 3.0Ghz 1200mhz Fsb, with only minor voltage changes"

"Overall still king of the 65nm budget chips. I've seen even
the newer revisions take punishment with 1.6v+ clocking in
at 3.7GHZ+" --- [ Um, OK, suicide run I guess :-) ]

It's a matter of how much of a risk you want to take with the
processor, whether you'd contemplate higher voltages.

Also, a word about how the VCore voltage is added. The
VID code, comes from the processor. Intel "range locks" the
VID, to a certain range. VID goes up and down, due to Intel
SpeedStep (EIST), which is why the motherboard manufacturer doesn't
bypass it entirely. If the motherboard is to easily support
EIST, they use the VID pins that come from the CPU.

And that means, any large VCore boost, is added after the
fact. On my motherboard, I used a resistor mod, to get the
extra voltage. But a decent motherboard, offers a "bump" from
the BIOS setup.

CPU ---- range ------------ (+) ----- final VCore voltage
locked |
VID Boost addition
term from BIOS

So the "boost", tends to be outside the so-called valid
range that Intel would prefer you use. My motherboard
didn't offer "boost" from the menu, and I used a resistor mod.
But most enthusiast boards, add that boost (somehow). And
boosting it, makes the processor hotter.

Note - I'm not good at this stuff, and there are a few excellent
tutorials on the web, which go into a lot more minute details.
(Anandtech had at least one article of some merit.)
The above is simple minded, because I don't know how to
set terminator voltages, NB voltages, for max overclock.
If you look hard enough, you'll get a better article than
what I just wrote.

BTW - Since Abit is out of business, it isn't easy to find manuals
and the like. To answer the above, I was lucky to have a copy
of "Abit_IP35_Manual.pdf" to verify the necessary basic settings
are in the BIOS screens.

Paul