p4gd1 torture chamber

Printable View

Show 100 post(s) from this thread on one page

03-02-2006, 02:10 PM
e.lectronick

Thanks, Johnny... Reading it this thread afap

:D

Quote:

Originally Posted by Johnny Bravo

Hey thats why we post, to learn :toast:

The reason everyone seems to be referring to 167MHz is because setting the FSB equal or higher to this enables the PCI-E/PCI locks so when you do increase the FSB ( with say clockgen) the PCI-E/PCI frequencies stay the same. Please note however that there is a "limit" to how high you can increase the FSB. THE general rule of thumb is that your PCI-E frequency * 2.2 = the maximum theoretical FSB you can attain. Please refer to the very first post for more details in this matter.

Thanks, Johnny, for the answer on that. I'm reading (and re-reading this entire thread along with anything else which came up after a search for "P4GD1" and "Mod"). Initial readings weren't very revealing, but more and more is becoming clear as I read everything over again.

I wanted to mention that I am the one who purchased Winterwind's ebay auction for "Motherboard Asus P4GD1 + CT-479 - Dothan ready combo!". I've got a lot of mods to understand before it arrives in the mail and I try to fire it up. I don't wanna fry anything. I'll probably post some more questions over the next few days.

Thanks again. and Cheers (or is that the same thing?:banana: )

-Erik.
03-03-2006, 07:06 AM
Speedy Gonzales

Everywhere I look for information on the P4GD1 OC-ing, I keep on stumbling on the PCI-E problems. Is it possible to achieve higher FSB frequencies than ~250MHz without that the PCI-E will run out of specifications, maybe by using a turbo-pll, just like back in the RDRAM days?
03-03-2006, 07:10 AM
Speedy Gonzales

http://www.overclockers.com/tips745/
03-03-2006, 01:27 PM
Nanok

If there is anybody who want to sell there P4GD1 Mainboard plz Pm me.
and a moddet P4GD1 even better :p:
Sure there is a lot of your guy`s opdating your system to the new dualcore
03-03-2006, 11:45 PM
hipro5

Quote:

Originally Posted by Speedy Gonzales

Everywhere I look for information on the P4GD1 OC-ing, I keep on stumbling on the PCI-E problems. Is it possible to achieve higher FSB frequencies than ~250MHz without that the PCI-E will run out of specifications, maybe by using a turbo-pll, just like back in the RDRAM days?

These things were done ONLY to get more fsb than motherbord could give us......Now that the mobos give up to 400MHz fsb we don't need them any more.....;)

You can pick the 110MHz PCI-E within bios and all the onboard hardware of it will work OK........
03-07-2006, 11:22 AM
Speedy Gonzales

Quote:

Originally Posted by hipro5: These things were done ONLY to get more fsb than motherbord could give us......Now that the mobos give up to 400MHz fsb we don't need them any more.....;)

Ermmm... Well, in case I'm totally wrong on this then I'd be happy to be aware of it. But from what I've read so far, turbo-pll's were made and used for several different kinds of reasons or combinations of these different reasons. And not only to get more FSB than certain motherboards could give us.

Sometimes they were used because of the fact that many turbo-pll allowed to indeed provide the motherboards with higher FSB's than the motherboards and their BIOSses settings allowed at default. Yet also sometimes they were used because along with that higher max FSB's, *some* (of the more sophisticated) turbo-pll's also allowed the PCI/AGP settings to remain within specs while raising the FSB on mobos that didn't have a PCI/AGP lock. Kind of like an artificial PCI/AGP lock. Something that the more simple do-it-yourself turbo-pll as described on http://www.overclockers.com/tips745/ (indeed) doesn't mention but certain turbo-pll's *did* or *do* support. Jordan, a company who makes/used to make turbo-pll's, for example affirmed that to me in a reply email when I inquired at it in about half a year ago. And for example, also according to wwilliam54 at http://www.xtremesystems.org/forums/...ead.php?t=5177 :

Quote:

Originally posted by wwilliam54: It can also be used to lock the PCI/AGP by sending a constant 33/66mhz to them instead of relying off a divider of the FSB

Other reasons why turbo-pll's were made and used was because some of them added the frequencies to be set variable in steps of about 1MHz even when a motherboard itsself didn't have that feature at default. And/or the fact that *some* turbo-pll's were possible to link to a nice controller unit with an LCD screen that displays the frequencies on the front of the PC (controller was built into an empty 5.25" bay on the front of a PC). Example:
http://www.innovatek.de/contentServ/...hp?StoryID=162 (not for sale anymore nowadays, according to an official reply email from Innovatek of a few weeks ago when I inquired at if it's still for sale or not)

By the way hipro8... I'm still a noob at this forum, but I already found out yesterday when I was looking for info on DDR Boosters, that you have made that very nice DDR Maximizer. :) I'm very impressed, very nice job! When relating this to the turbo-pll and so on; if there's people that very possibly do have the needed skills for making a turbo-pll like the one that I'm looking for, I guess you might be one of them? So are you interested and can I hereby lure you into exploring the possibilities and thinking along with me on this? Maybe even into attempting to make one? :D

Quote:

Originally Posted by hipro5: You can pick the 110MHz PCI-E within bios and all the onboard hardware of it will work OK........

I didn't try it out yet, for I still didn't buy a PCI-E graphic card yet (and have the brand new P4GD1 still laying unused in a box). But are you sure that if I pick the 110MHz PCI-E setting, I can fix it to that 110MHz and then upper the FSB to the 300(+)MHz that I want in order to run my Prescott ES and OCZ PC4800 Platinum EL Platinum sticks at 1:1? Hummm...
03-08-2006, 10:31 PM
hipro5

Quote:

Originally Posted by Speedy Gonzales

Ermmm... Well, in case I'm totally wrong on this then I'd be happy to be aware of it. But from what I've read so far, turbo-pll's were made and used for several different kinds of reasons or combinations of these different reasons. And not only to get more FSB than certain motherboards could give us.

Sometimes they were used because of the fact that many turbo-pll allowed to indeed provide the motherboards with higher FSB's than the motherboards and their BIOSses settings allowed at default. Yet also sometimes they were used because along with that higher max FSB's, *some* (of the more sophisticated) turbo-pll's also allowed the PCI/AGP settings to remain within specs while raising the FSB on mobos that didn't have a PCI/AGP lock. Kind of like an artificial PCI/AGP lock. Something that the more simple do-it-yourself turbo-pll as described on http://www.overclockers.com/tips745/ (indeed) doesn't mention but certain turbo-pll's *did* or *do* support. Jordan, a company who makes/used to make turbo-pll's, for example affirmed that to me in a reply email when I inquired at it in about half a year ago. And for example, also according to wwilliam54 at http://www.xtremesystems.org/forums/...ead.php?t=5177 :

Other reasons why turbo-pll's were made and used was because some of them added the frequencies to be set variable in steps of about 1MHz even when a motherboard itsself didn't have that feature at default. And/or the fact that *some* turbo-pll's were possible to link to a nice controller unit with an LCD screen that displays the frequencies on the front of the PC (controller was built into an empty 5.25" bay on the front of a PC). Example:
http://www.innovatek.de/contentServ/...hp?StoryID=162 (not for sale anymore nowadays, according to an official reply email from Innovatek of a few weeks ago when I inquired at if it's still for sale or not)

By the way hipro8... I'm still a noob at this forum, but I already found out yesterday when I was looking for info on DDR Boosters, that you have made that very nice DDR Maximizer. :) I'm very impressed, very nice job! When relating this to the turbo-pll and so on; if there's people that very possibly do have the needed skills for making a turbo-pll like the one that I'm looking for, I guess you might be one of them? So are you interested and can I hereby lure you into exploring the possibilities and thinking along with me on this? Maybe even into attempting to make one? :D

I didn't try it out yet, for I still didn't buy a PCI-E graphic card yet (and have the brand new P4GD1 still laying unused in a box). But are you sure that if I pick the 110MHz PCI-E setting, I can fix it to that 110MHz and then upper the FSB to the 300(+)MHz that I want in order to run my Prescott ES and OCZ PC4800 Platinum EL Platinum sticks at 1:1? Hummm...

If you pick 110MHz PCI-E within bios and then play with clockgen within windows, the PCI-E is rockstable at 110MHz nomatter what fsb you pick.......One problem though as for the fsb and the rams, is that you HAVE to go higher on the PCI-E so as to play higher the fsb........

P.S. I know all that as for the Turbo pll.......;)
03-12-2006, 11:28 AM
Speedy Gonzales

Hi. Thanks again for your reply, hipro5. :) However, unfortunately
I still don't quite understand what you mean yet. To me, some
things that you said now don't seem to be in coördinance with
other things that you said earlier on. Please help me... I know
that the matter that we're discussing is a bit more abstract than
regular, so I'm trying to understand it fully the way that you
mean it.

See, for example what I don't get is: earlier on you said about the
turbo-pll's: "These things were done ONLY to get more fsb than
motherbord could give us......Now that the mobos give up to
400MHz fsb we don't need them any more.....". Yet after I posted
examples of other and/or more reasons for why people also used
turbo-pll's, you say: "P.S. I know all that as for the Turbo
pll.......". Why did you say in the first place then, that people
ONLY used them to get higher FSB's than motherboards could
give us? That's in contradiction to each other, right?

Also, first you say: "If you pick 110MHz PCI-E within bios and
then play with clockgen within windows, the PCI-E is rockstable
at 110MHz nomatter what fsb you pick.......", yet then right after
that, you say: "One problem though as for the fsb and the rams,
is that you HAVE to go higher on the PCI-E so as to play higher
the fsb........". Now if I get it the way you mean it, then on the
one hand you say that if I pick 110MHz PCI-E within BIOS and
then use ClockGen to change the FSB within Windows, nomatter
what FSB I pick within ClockGen, the PCI-E will stay fixed at
110MHz(?), yet on the other hand you say that I have to set the
PCI-E higher (within BIOS?) in order to be able to set the FSB
(and RAM) at high frequencies. Is that what you mean? If yes,
then these basically two things that you're saying are not in
coördinance with each others... or am I missing something?

When remaining to what you say about that I have to raise the
PCI-E when I want to reach higher FSB's... Then if I would cut
the traces that lead from the onboard clockgenerator to the
PCI-E (the traces through which the onboard clockgenerator
feeds the PCI-E with a frequency) and then create a turbo-pll
that would constantly feed the PCI-E with a 100MHz frequency,
then the PCI-E will no longer be influenced by the FSB
frequencies anymore, right. Don't you think that would solve it
all? Because that's the kind of concept for a turbo-pll that I have
in mind, see.

I created the following pictures, based on the ones that are on
the http://www.overclockers.com/tips745/ page, to
show the concept of a turbo-pll that I'm considering to create:

Situation without overclock and without turbo-pll:
http://www.geocities.com/speedygonza...rclocks/01.bmp
Situation with (high FSB) overclock but without turbo-pll:
http://www.geocities.com/speedygonza...rclocks/02.bmp
Situation with overclock and with turbo-pll:
http://www.geocities.com/speedygonza...rclocks/03.bmp

So what do you think about it, hipro5 (and other people that
might be interested, of course)? By the way, I also contacted
macci about it and am awaiting a reply. Macci used turbo-pll's to
overclock up until insane high levels back in the RDRAM times,
see. :)
03-12-2006, 12:38 PM
hipro5

Quote:

Originally Posted by Speedy Gonzales

Hi. Thanks again for your reply, hipro5. :) However, unfortunately I still don't
quite understand what you mean yet. To me, some things that you said now
don't seem to be in coördinance with other things that you said earlier
on. Please help me... I know that the matter that we're discussing is a bit
more abstract than regular, so I'm trying to understand it fully the way that
you mean it.

See, for example what I don't get is: earlier on you said about the
turbo-pll's: "These things were done ONLY to get more fsb than motherbord
could give us......Now that the mobos give up to 400MHz fsb we don't need
them any more.....". Yet after I posted examples of other and/or more
reasons for why people also used turbo-pll's, you say: "P.S. I know all that as
for the Turbo pll.......". Why did you say in the first place then, that people
ONLY used them to get higher FSB's than motherboards could give us? That's
in contradiction to each other, right?

Old times the PCI and the AGP were NOT locked when we picked higher fsb...That was a good reason for turbo PLL....We could work in higher frequencies and lower AGP/PCI than the mobo could provide with a turbo pll.....
NOW they are LOCKED +/- no matter what.......;)

Quote:

Originally Posted by Speedy Gonzales

Also, first you say: "If you pick 110MHz PCI-E within bios and then play with
clockgen within windows, the PCI-E is rockstable at 110MHz nomatter what
fsb you pick.......", yet then right after that, you say: "One problem though
as for the fsb and the rams, is that you HAVE to go higher on the PCI-E so
as to play higher the fsb........". Now if I get it the way you mean it, then on
the one hand you say that if I pick 110MHz PCI-E within BIOS and then
use ClockGen to change the FSB within Windows, nomatter what FSB I pick
within ClockGen, the PCI-E will stay fixed at 110MHz(?), yet on the other
hand you say that I have to set the PCI-E higher (within BIOS?) in order to be
able to set the FSB (and RAM) at high frequencies. Is that what you mean?
If yes, then these basically two things that you're saying are not in
coördinance with each others... or am I missing something?

http://www.thelab.gr/images/Hipro5/D...ency_1_jpg.jpg

http://www.thelab.gr/images/Hipro5/D...ency_2_jpg.jpg

http://www.thelab.gr/images/Hipro5/D...ency_1_jpg.jpg

LOCKED as hell........;)

Quote:

Originally Posted by Speedy Gonzales

When remaining to what you say about that I have to raise the PCI-E when
I want to reach higher FSB's... Then if I would cut the traces that lead from
the onboard clockgenerator to the PCI-E (the traces through which the
onboard clockgenerator feeds the PCI-E with a frequency) and then create a
turbo-pll that would constantly feed the PCI-E with a 100MHz frequency,
then the PCI-E will no longer be influenced by the FSB frequencies anymore,
right. Don't you think that would solve it all? Because that's the kind of
concept for a turbo-pll that I have in mind, see.

I created the following pictures, based on the ones that are on the
http://www.overclockers.com/tips745/ page, to show the concept of a turbo-pll
that I'm considering to create:

Situation without overclock and without turbo-pll:
http://www.geocities.com/speedygonza...rclocks/01.bmp
Situation with (high FSB) overclock but without turbo-pll:
http://www.geocities.com/speedygonza...rclocks/02.bmp
Situation with overclock and with turbo-pll:
http://www.geocities.com/speedygonza...rclocks/03.bmp

So what do you think about it, hipro5 (and other people that might be
interested, of course)? By the way, I also contacted macci about it and am
awaiting a reply. Macci used turbo-pll's to overclock up until insane high
levels back in the RDRAM times, see. :)

Don't bother......NO need for turbo pll anymore......;)
03-13-2006, 07:12 PM
Kainthelongshot

Has anybody put more than 2 gb of ram on this Mobo?

Just curious. Has anybody placed the full 4 gigs of ram this mobo can hold? If so what kind of ram did you use to accomplish this?
03-13-2006, 08:23 PM
cirthix

Quote:

Originally Posted by Kainthelongshot

Just curious. Has anybody placed the full 4 gigs of ram this mobo can hold? If so what kind of ram did you use to accomplish this?

for me, 4 sticks of pc3200 drops the divider to 333mhz and i can't change it.

3 dimms also drops the divider.
03-14-2006, 12:32 AM
BlueWonder

Quote:

Originally Posted by cirthix

for me, 4 sticks of pc3200 drops the divider to 333mhz and i can't change it.

3 dimms also drops the divider.

Is it all the same stuff?
03-14-2006, 12:15 PM
Speedy Gonzales

Dear hipro5,

I understand and discovered now, that the PCI-E is locked at
whatever frequency you set it in BIOS. But that it needs to be
set at high frequencies there, in order to be able to set the FSB
to high frequencies of 300MHZ or over. I call that just a half
PCI-E lock by the way, since it's not a so very useful one, but
that aside. Since the formula to use for calculating the PCI-E
frequency that needs to be set at a certain FSB frequencies is
FSB=2.2*PCI-E, at 300FSB the PCI-E frequency needs to be set
to 300/2.2=136MHz, as I'm sure you're aware of. But of course
that's a way too high PCI-E frequency. The SATA controller will
stop functioning and the PCI-E graphics card might very likely fail
working too. I don't want that, so I have create a solution for
being able to set the PCI-E frequency to 100MHz and fixing it to
that and still be able to let the system work at 300MHz (or over),
instead of that I have to set it to that 136MHz (or over). The
only way to achieve this is by using a turbo-pll according to the
concept that I have described and am displaying on the pictures
that I posted links too earlier on. If you still don't get what I
mean, then I will no longer try to explain it for now, please try
and take another look into it if you're interested, because I have
to focus on trying to find out how to get it to work now. I
however thank you for your replies and wish you success with all
your own projects too.

Greetings,
Speedy Gonzales
03-14-2006, 12:41 PM
skirms_fr

@ hirpo , what is your ram frequency limit on the p4gd1 ?

around 272 / 274 ?
03-14-2006, 12:56 PM
hipro5

Quote:

Originally Posted by skirms_fr

@ hirpo , what is your ram frequency limit on the p4gd1 ?

around 272 / 274 ?

278MHz 3D benchable and 285MHz max post(if I remember well)......
03-14-2006, 01:26 PM
skirms_fr

ok thanks , i have a wall @ 274 vdimm 3.96v and vchip 2.0

i thinks i m at the max of the chipset :/

and do you remember your maxscreen 780 under cascade ? to have an idée where i am :p
03-14-2006, 01:32 PM
hipro5

Quote:

Originally Posted by skirms_fr

ok thanks , i have a wall @ 274 vdimm 3.96v and vchip 2.0

i thinks i m at the max of the chipset :/

and do you remember your maxscreen 780 under cascade ? to have an idée where i am :p

Feed the chip @ 2.1Volts to 2.15Volts.......Put the PCI-E frequency into Windows OVER 130MHz.......and give it a try with NO MORE than 3.7Vdimm........

780ES @ 4050MHz(?).....
03-14-2006, 02:10 PM
skirms_fr

ok i will try soon :) thanks

around 3815/3820 for mine

with cascade R170
03-23-2006, 07:38 AM
skirms_fr

PM780 benching spi 1mo @ 3.6 at 1.81 vcore , 1.9 vchip , 3.8 vdimm , 2.3 Vcca, after crashing , impossible to reboot :/

any ideas on the problem for the p4gd1 ?
03-23-2006, 12:19 PM
hipro5

Quote:

Originally Posted by skirms_fr

PM780 benching spi 1mo @ 3.6 at 1.81 vcore , 1.9 vchip , 3.8 vdimm , 2.3 Vcca, after crashing , impossible to reboot :/

any ideas on the problem for the p4gd1 ?

Try to have lower or equal to 110MHz PCI-E in bios so as to avoid that.....
03-23-2006, 03:41 PM
skirms_fr

Quote:

Originally Posted by hipro5

Try to have lower or equal to 110MHz PCI-E in bios so as to avoid that.....

the same problem after a Cmos :/

i ve cleanning the CT due to the condensation , but no go :/

i dont understand
03-23-2006, 10:00 PM
hipro5

Quote:

Originally Posted by skirms_fr

the same problem after a Cmos :/

i ve cleanning the CT due to the condensation , but no go :/

i dont understand

The only thing that goes through my mind is coz of the x17 multi.....What is your start up frequency within bios?.......Try to put it as low as it works with 1,35VCore.......
03-23-2006, 10:16 PM
Movieman

After a 9 month seach and wait for it to be delivered my P4GD1 arrived last Friday. Like a little kid I set out to get the system going. I've built maybe 200 odd systems in the last 9 years so I'm hopefully competant. I set it up, fire it up, cdrw and HD have power,cpu and added chipset fan are functional, power lights on board are light..no post. Hmm, I figure I missed a connection, check everything, nope..I try 3 video cards(2-PCI and one PCI-e) 2 different sets of ram in single channel and dual channel, 3 different power supplies..still nothing. I clear the CMOS even though I think it's a waste of time..
Nothing..Can you say frustrated? 9 months and $190.00 to get this board.
Zagreb,Croatia to Hamburg,Germany, hand carried to NYC and mailed to me here in New Hampshire..At 54 I've leaned not to let too much bother me but this does..so much effort wasted.
03-23-2006, 10:49 PM
cirthix

try it with a 533fsb or 800fsb p4
03-23-2006, 10:50 PM
Movieman

Quote:

Originally Posted by cirthix

try it with a 533fsb or 800fsb p4

I did, a 1800 celeron..no difference..

Show 100 post(s) from this thread on one page