A computer components & hardware forum. HardwareBanter

If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below.

Go Back   Home » HardwareBanter forum » Processors » Overclocking AMD Processors
Site Map Home Register Authors List Search Today's Posts Mark Forums Read Web Partners

Help with overclocking Duron on gigabyte mobo!



 
 
Thread Tools Display Modes
  #1  
Old June 2nd 04, 04:43 AM
matiii
external usenet poster
 
Posts: n/a
Default Help with overclocking Duron on gigabyte mobo!

Hi,

I got Duron 700 on Gigabyte nForce2 GA-7N400 mobo (I know, $h1tty cpu
on good mobo - I'll buy a new one when I have some cash). I got myself
a better cooling and I want to overclock D700 to 1GHz, because AFAIK
now it runs in asynchronous mode with system bus 100, and memory at
200 (got PC3200).

So the reasonable combination seems 5x200, right? Correct me if I'm
wrong. Now, I have to unlock multiplier in Duron, so I've bought
conductive paint, I closed L1 bridges and I wanted to set jumpers on
mobo to 5x, then change FSB to 200 in BIOS. Unfortunately system
refuses to POST, when it has multiplier set to manual, so I had to
switch it back to auto and additionaly clear CMOS. The only exeption
is when I set multi to 7x (a default one in Duron), it works then too.

Another thing I don't get is FSB setting in BIOS - no matter if I
choose 100, or 200, CPU-Z reports that it runs at 100 and, of course,
CPU runs at 700. I want FSB (or system clock as WCPUId calls it) to
run at 200. There is another jumper on mobo, clk_sw - manual says it
should be OFF, when using FSB 100 or 200. It *is* off, and when I
tried changing it, the system wouldn't POST either. Duron
specification says maximum FSB is 200, so what's the story?

I'm pretty confused. I don't know if:

1. I screwed the paint job (though I did it really carefully, I knew
what
to do, and what *not* to do);
2. the jumpers on mobo should be set in certain way;
3. or I missed some BIOS settings.

There is another way, on
http://www.ocinside.de/go_e.html?/ht.../workshop.html there's
an *interactive painting guide...* for Duron, and they give also a way
to alter multiplier - not only unlock it, but change, and voltage,
too. Maybe the mobo is no good for overclocking, and I got to set new
values on CPU itself by cutting/joining bridges?

Help will be *really* appreciated, thanks in advance
Matt
  #2  
Old June 2nd 04, 08:10 AM
Wes Newell
external usenet poster
 
Posts: n/a
Default

On Tue, 01 Jun 2004 20:43:45 -0700, matiii wrote:

I got Duron 700 on Gigabyte nForce2 GA-7N400 mobo (I know, $h1tty cpu
on good mobo - I'll buy a new one when I have some cash). I got myself


Help will be *really* appreciated, thanks in advance


I couldn't get my Duron 700 to run a 1000MHz. Topped out about 950. With
what you have. I'd raise vcore to 1.85 and set the FSB to 133 for 7x133.
Easy and should work ok unless the cpu is a dog.

--
Abit KT7-Raid (KT133) Tbred B core CPU @2400MHz (24x100FSB)
http://mysite.verizon.net/res0exft/cpu.htm
  #3  
Old June 3rd 04, 01:00 PM
matiii
external usenet poster
 
Posts: n/a
Default

Wes Newell wrote in message news:pan.2004.06.02.07.13.30.421575@TAKEOUTverizo n.net...

.... With
what you have. I'd raise vcore to 1.85 and set the FSB to 133 for 7x133.
Easy and should work ok unless the cpu is a dog.


I'm getting really ****ed. Dunno why, hardware totally ignores what I
set in BIOS. No matter if I set FSB to 133, 200 - it always runs
7x100 - WTF!!?!?

My Duron has those symbols on its co

D700AUT1B

Which, according to AMD specs means:

D700 obvious, A is a package type - doesn't matter;
U - nominal core voltage - 1.60 V
T - max. die temp. - 90 deg C
1 - level 2 cache size - 64 KB
B - *maximum system bus speed* - *200 MHz*

stepping code AKCA, manufactured in 2000, 49th week (AKCA0049MPMW)

So? What's wrong? Is BIOS somewhat screwed? Maybe I should flash it,
as someone already has suggested it to me. Or maybe this Duron is too
old to run at higher bus? But I heard of a guy (on #IRCNet,
#overclocking) that runs D600@1000

Really hope to read some good news
Matt
  #4  
Old June 3rd 04, 08:52 PM
Wes Newell
external usenet poster
 
Posts: n/a
Default

On Thu, 03 Jun 2004 05:00:09 -0700, matiii wrote:

Wes Newell wrote in message
news:pan.2004.06.02.07.13.30.421575@TAKEOUTverizo n.net...

.... With
what you have. I'd raise vcore to 1.85 and set the FSB to 133 for
7x133. Easy and should work ok unless the cpu is a dog.


I'm getting really ****ed. Dunno why, hardware totally ignores what I
set in BIOS. No matter if I set FSB to 133, 200 - it always runs 7x100
- WTF!!?!?

Check MB manual. Some boards have to enable the bios control via jumper on
board. I suspect that's what you'fve got.

My Duron has those symbols on its co

D700AUT1B

Which, according to AMD specs means:

D700 obvious, A is a package type - doesn't matter; U - nominal core
voltage - 1.60 V
T - max. die temp. - 90 deg C
1 - level 2 cache size - 64 KB
B - *maximum system bus speed* - *200 MHz*

That's default speed, not maximum.

So? What's wrong? Is BIOS somewhat screwed? Maybe I should flash it, as
someone already has suggested it to me. Or maybe this Duron is too old
to run at higher bus? But I heard of a guy (on #IRCNet, #overclocking)
that runs D600@1000

I could get 1000Mhz out of both my 600 and 700 duron, but not stable.

--
Abit KT7-Raid (KT133) Tbred B core CPU @2400MHz (24x100FSB)
http://mysite.verizon.net/res0exft/cpu.htm
  #5  
Old June 4th 04, 01:06 PM
Gene Puhl
external usenet poster
 
Posts: n/a
Default

(matiii) astounded us with:
om:

Hi,

I got Duron 700 on Gigabyte nForce2 GA-7N400 mobo (I know, $h1tty cpu
on good mobo - I'll buy a new one when I have some cash). I got myself
a better cooling and I want to overclock D700 to 1GHz, because AFAIK
now it runs in asynchronous mode with system bus 100, and memory at
200 (got PC3200).

So the reasonable combination seems 5x200, right? Correct me if I'm
wrong. Now, I have to unlock multiplier in Duron, so I've bought
conductive paint, I closed L1 bridges and I wanted to set jumpers on
mobo to 5x, then change FSB to 200 in BIOS. Unfortunately system
refuses to POST, when it has multiplier set to manual, so I had to
switch it back to auto and additionaly clear CMOS. The only exeption
is when I set multi to 7x (a default one in Duron), it works then too.

Another thing I don't get is FSB setting in BIOS - no matter if I
choose 100, or 200, CPU-Z reports that it runs at 100 and, of course,
CPU runs at 700. I want FSB (or system clock as WCPUId calls it) to
run at 200. There is another jumper on mobo, clk_sw - manual says it
should be OFF, when using FSB 100 or 200. It *is* off, and when I
tried changing it, the system wouldn't POST either. Duron
specification says maximum FSB is 200, so what's the story?


The Duron is "double pumped", an FSB of 100mhz is really 200mhz to the
system. The Spitfire (and Morgan cores up to 1400mhz, IIRC)run at 100mhz,
that's it...If you've unlocked the processor, raise your multiplier, not
the FSB.






--
Gene P
  #6  
Old June 4th 04, 03:47 PM
Michael Brown
external usenet poster
 
Posts: n/a
Default

matiii wrote:
Hi,

I got Duron 700 on Gigabyte nForce2 GA-7N400 mobo (I know, $h1tty cpu
on good mobo - I'll buy a new one when I have some cash). I got myself
a better cooling and I want to overclock D700 to 1GHz, because AFAIK
now it runs in asynchronous mode with system bus 100, and memory at
200 (got PC3200).

So the reasonable combination seems 5x200, right? Correct me if I'm
wrong. Now, I have to unlock multiplier in Duron, so I've bought
conductive paint, I closed L1 bridges and I wanted to set jumpers on
mobo to 5x, then change FSB to 200 in BIOS. Unfortunately system
refuses to POST, when it has multiplier set to manual, so I had to
switch it back to auto and additionaly clear CMOS. The only exeption
is when I set multi to 7x (a default one in Duron), it works then too.


There's a couple of gotchas with the CPGA Durons (the ceramic ones, they go
"clink" when you tap em with a screwdriver ) and some motherboards. This
issue manifests itself as the board refusing to POST at any multiplier other
than the default one for the chip. For example, a friend of mine has an
identical Duron to you. He was wondering how high it could go, so I swapped
it for a spare Palomino I had lying around for a few days so that I could do
a few experiments. I rejoined the L1's and could set the multiplier to
whatever I wanted on my Soltek 75DRV5. I found a nice stable point on that
board, pinmodded it, Prime95'd it, and dropped it back off to his place. It
wouldn't POST on his motherboard (a Gigabyte as well, funnily enough).
Dropping it back into my board madde it work perfectly again.

So, I removed the pinmod and is began working (at the default multiplier) in
his board again. Changing the multiplier to anything but stock made it fail
to boot.

The problem is that the CPGA Durons (and TBirds, possibly) have the
multiplier set in TWO places: The L6, which tells the M/B what multiplier to
use, and the L1/L2/L3/L4 setup which tell the CPU what to use. Some BIOSes
check to make sure these are equal, and won't POST if they are different.
With the CPGA chips, the copper traces on the top are fakes. You can remove
them without damaging the chip at all (yes, this was verified on the Duron
in a rather unfortunate accident ...). The real traces are actually under
the ceramic, so you need a diamond drill to cut through them.

Alternatively, you can mod the motherboard to make it do multiplier
adjustments properly. As per usual standards, Fab51 has this fully
documented (albeit only in Japanese).
http://fab51.fc2web.com/pc/duron/duron1.html

With a 7x multiplier, you've got all the (important) L6's already pulled to
ground, unless you're wanting to go straight for a 11.0x multiplier (not
recommended). In this case, you either have to modify the motherboard (sever
the traces, yikes!) or get fairly brutal on the CPU (rip the pins off,
yikes!). I'd personally recommend you don't try either of these, and go for
straight FSB overclocking. With a 7x multiplier you won't have any problems
with the board not being able to handle the speed.


One thing that DOES puzzle me though is that you're using a fairly recent
board, but it doesn't do things correctly. My first step would be to set the
FSB back to the default, 100MHz, then try adjusting the multiplier to 6.5x.
If this works, then it's just the CPU not liking high FSBs or just not
liking the 1GHz you were trying to run it at. Try something a little lower,
such as 5x133 or 5x150. Also, you might want to try running the memory in
async mode which may give you a bit of a boost in in games.

[...]

--
Michael Brown
www.emboss.co.nz : OOS/RSI software and more
Add michael@ to emboss.co.nz - My inbox is always open


  #7  
Old June 5th 04, 04:31 PM
matiii
external usenet poster
 
Posts: n/a
Default

SHHEEESHH!!

1. I set mobo multiplier jumpers to auto, FSB jumper (CLK_SW) to off,
meaning 100MHz "fix 200MHz CPU" as the manual says. It also says "You
must set CLK_SW to 100MHz when you used FSB 200MHz CPU". So it is
done.

2. I reset BIOS to defaults, i.e. FSB 100 as well.

3. Of course, Duron reports at 700 - when I change FSB steb by step,
there's no difference, as I posted before - system doesn't POST when
there's other multi than 7, though I haven't tried all (reasonable,
because list reaches 18 combinations.

So I don't got any more idea how to move the FSB. I was changing it
with "System Performance" set to "Normal", in raw 33 MHz steps, as
well as in "Manual" mode, one by one - no difference at all. Maybe I
will flash the BIOS and see if it helps, I'm getting really fed up and
****ed.

Michael Brown wrote:

With the CPGA chips, the copper traces on the top are fakes. You can

remove
them without damaging the chip at all (yes, this was verified on the

Duron
in a rather unfortunate accident ...). The real traces are actually

under
the ceramic, so you need a diamond drill to cut through them.


I also considered changing multi by opening/closing the L6 bridges,
but that really put me off. Where do you got this info from?

Oh, and the last thing, Gene Puhl wrote:

The Duron is "double pumped", an FSB of 100mhz is really 200mhz to

the
system.


So, what about DDRs? CPU-Z (or WCPUid) reports that *FSB* is 200 and
*system bus*, or *system clock* (WCPUid) is 100 - DDR runs at 200,
according to CPU-Z. Now, is this physical frequency they run at, and
then data rate is doubled, cause they are DDRs, or is it calculated
double-times-FSB, meaning they actually run at 100MHz? (BIOS says they
run at 400...) And which solution is more effective considering the
fact it's AMD?

Matt
  #8  
Old June 6th 04, 01:33 AM
Michael Brown
external usenet poster
 
Posts: n/a
Default

matiii wrote:
SHHEEESHH!!

1. I set mobo multiplier jumpers to auto, FSB jumper (CLK_SW) to off,
meaning 100MHz "fix 200MHz CPU" as the manual says. It also says "You
must set CLK_SW to 100MHz when you used FSB 200MHz CPU". So it is
done.


I think this just makes the FSB at 100MHz, and unchangable. This is not what
you want, so try setting CLK_SW to on.

[...]
Michael Brown wrote:

With the CPGA chips, the copper traces on the top are fakes. You can
remove them without damaging the chip at all (yes, this was verified
on the Duron in a rather unfortunate accident ...). The real traces
are actually under the ceramic, so you need a diamond drill to cut
through them.


I also considered changing multi by opening/closing the L6 bridges,
but that really put me off. Where do you got this info from?


http://www.beachlink.com/candjac/duronocpg2.htm
Except this page is no longer in existance ... archive.org to the rescue!
http://web.archive.org/web/200302020...ndjac/duronocp
g2.htm

The "top traces are fake" was verified by messing around with the VID
bridges on a CPGA duron and not having anything change despite completely
removing one bridge from the chip. Well, they are sorta fake ... you can
rejoin a bridge using the ones on the top, but not break one.

The diamond-tipped drill info from various sources, the most interesting
being
http://www.bunkermentality.net/keychain.html
(down the bottom)


Oh, and the last thing, Gene Puhl wrote:

The Duron is "double pumped", an FSB of 100mhz is really 200mhz to
the system.


So, what about DDRs? CPU-Z (or WCPUid) reports that *FSB* is 200 and
*system bus*, or *system clock* (WCPUid) is 100 - DDR runs at 200,
according to CPU-Z. Now, is this physical frequency they run at, and
then data rate is doubled, cause they are DDRs, or is it calculated
double-times-FSB, meaning they actually run at 100MHz?


Welcome the the world of marketing-induced FSB confusion. I'm sure Wes will
do a nice rant on this but essentially half the bus (the control part) is
running at 100MHz, and the other half (the data part) is running at 200MHz
(which is why it's called DDR). The actual clock signal fed to the CPU is
the 100MHz one. Which speed is reported varies with whichever tool you are
using. Some report the doubled speed, some report the base speed. The
"marketing" speeds are the doubled ones, whereas the BIOS ones are are quite
inconsistant. If you have 1MHz adjustments, these will be the base speeds,
and if you have ones that allow you to select 400,333,266,200 then they're
the doubled speeds.

The base speed is what is multiplied by the CPU (since this is what it is
fed), ie: with a 200MHz marketing bus, you have a 7x multiplier to get
700MHz.

[...]

--
Michael Brown
www.emboss.co.nz : OOS/RSI software and more
Add michael@ to emboss.co.nz - My inbox is always open


  #9  
Old June 6th 04, 06:40 PM
phsstpok
external usenet poster
 
Posts: n/a
Default

The problem is that the CPGA Durons (and TBirds, possibly) have the
multiplier set in TWO places: The L6, which tells the M/B what multiplier to
use, and the L1/L2/L3/L4 setup which tell the CPU what to use. Some BIOSes
check to make sure these are equal, and won't POST if they are different.


Sorry, but you are slightly in error here. L1 controls unlocking. L3
and L4 determine multipliers. L6 determines Muliplier ID. These
bridges do work the same way on Thunderbird.

I only learned recently how imortant Multiplier ID for Spitfire Duron
(a 600). Although mine was factory unlocked (L1 bridges not cut) and
I had it working with BIOS multipliers (9.5 x 106 Mhz= 1007 Mhz) I ran
into problems when performing wire mods (wire-in-socket in my case).
I used the same mods that I was using for a TBred B XP1700+. Set it
for a conservative 7X x 133Mhz. No joy. Realizing I had never tested
the Duron on this motherboard (EPoX 8KTA3PRO, KT133A) I decided to
retest at 100 Mhz. Only the default 6X would work! Tried 7X, and 9X
(perhaps a few others). Still nothing. I confirmed everything I tried
with my Tbred yet the Duron refused to POST.

Going back to BIOS multipliers the Duron worked perfectly. In fact
with a much better HSF than I had while the Duron was on my Abit KT7 I
am able to overclock it to 1112 Mhz (8 x 139). This is with the same
board on which I was trying the wiremods.

I gave up.

A month or so later I was re-reading John C's website and also
OCinside.de when it dawned on me that the functionality of the L6
bridges exists on the pin grid. With Tbreds and unlocked Bartons we
need only perform wirmods on pins AN27, AL27, AN25, AL25, and, AJ27
for multipliers (0.5X...8X) but on Spitfire/Thunderbird (and I think
Palomino) we need also mod'd pins W1, W3, Y1 and Y3, Multiplier ID
pins. At least it's necessary with legacy motherboards (all I have).
Anyway, once I learned this there were no more problems with the
Spitfire.

Phsstpok
  #10  
Old June 8th 04, 11:06 AM
matiii
external usenet poster
 
Posts: n/a
Default

Michael Brown wrote in message news:

1. I set mobo multiplier jumpers to auto, FSB jumper (CLK_SW) to off,
meaning 100MHz "fix 200MHz CPU" as the manual says. It also says "You
must set CLK_SW to 100MHz when you used FSB 200MHz CPU". So it is
done.


I think this just makes the FSB at 100MHz, and unchangable. This is not what
you want, so try setting CLK_SW to on.


Here we are! I don't know if it is thanks to the new BIOS (I flashed
it) or I haven't checked CLK_SW *on* and FSB jumpers *auto*
combination yet. Anyway, now I'm able to change fsb freely So at
the moment D700 runs at 931 MHz (FSB 133) with vcore around 1.79. So
now, definitely, system runs at asynchronous mode, doesn't it? (DRAM
speed is set to *BY SPD* which results in 400 in BIOS) Maybe I should
change it to make it run in synchro? Will it be more efficient, and is
there a particular way to test it except for overall benchmarks?

Prime95 reported an error, doh... :

FATAL ERROR: Rounding was 0.4818878174, expected less than 0.4

Don't know if to bother, I haven't checked computer earlier much, so
can't tell if it's due to higher speed, or something got just screwed
at the testing time, or there's something generally wrong with he
stuff. Memory is Twinmos 256 PC3200, so it ain't bad I guess...

The multi change still doesn't work, either I can't paint , or it
is really the case with this multiplier double-check. So, as I
understood it, you can close those bridges if they're open, but to
open'em you'd have to fight the ceramic, right? Yikes, indeed

Matt

PS. BTW, any gotchas with Athlon 1GHz? (dunno the fsb yet, it's my
girl's comp, gotta check)
PSS. Any ideas for a good aircooler for P4 at a reasonable price?
 




Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Overclocking Duron 1.2GHz CroW Overclocking AMD Processors 10 February 19th 04 10:43 PM
a goog mobo for Duron 1.4ghz o/c Andre Overclocking AMD Processors 10 November 29th 03 08:47 PM
Question about mobo for Duron 1600 overclocking. M Arino Overclocking AMD Processors 7 November 26th 03 01:18 PM
Athlon 2400+ and Gigabyte Triton series mobo: Overclock? who be dat? Overclocking AMD Processors 7 November 20th 03 05:52 PM
Duron 900 and K7S5A mobo Nick Pedley Overclocking AMD Processors 1 June 27th 03 06:15 PM


All times are GMT +1. The time now is 01:23 PM.


Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Copyright ©2004-2024 HardwareBanter.
The comments are property of their posters.