Motherboard Forums


Reply
Thread Tools Display Modes

Can't slow down, any ideas?

 
 
Pete
Guest
Posts: n/a
 
      06-16-2006, 06:04 AM
Hi All:

I have an opty 150 that seems to be quite stable (24 h prime and completed a
5 day calculation in MATLAB) at 256x11@1.55V on a DFI LP NF4 Ultra-D
motherboard. The RAM/FSB is 1:1 so nothing tricky should be happening. I
have 4 GB of dual channel OCZ gold xtc PC 4000 RAM with 2T, 3-3-4-8 timings.
I have the RAM drive strength maxed out. OS is XP64. When I run memtest at
these settings no errors. If I keep everything else constant and drop the
multiplier down to 10, I get all kinds of memory errors in memtest. Any
ideas of how going slower would decrease the stability, it seems to go
against the central dogma of overclocking.

TIA,

Pete


 
Reply With Quote
 
 
 
 
Paul
Guest
Posts: n/a
 
      06-16-2006, 08:24 AM
In article <ERrkg.69350$(E-Mail Removed)>, "Pete"
<(E-Mail Removed)> wrote:

> Hi All:
>
> I have an opty 150 that seems to be quite stable (24 h prime and completed a
> 5 day calculation in MATLAB) at 256x11@1.55V on a DFI LP NF4 Ultra-D
> motherboard. The RAM/FSB is 1:1 so nothing tricky should be happening. I
> have 4 GB of dual channel OCZ gold xtc PC 4000 RAM with 2T, 3-3-4-8 timings.
> I have the RAM drive strength maxed out. OS is XP64. When I run memtest at
> these settings no errors. If I keep everything else constant and drop the
> multiplier down to 10, I get all kinds of memory errors in memtest. Any
> ideas of how going slower would decrease the stability, it seems to go
> against the central dogma of overclocking.
>
> TIA,
>
> Pete


Are you verifying with CPUZ, that when you change the multiplier,
nothing else is changing ? Try verifying the settings in each case.
(I.E. Never trust a BIOS further than you can throw it :-))

DFI is known to be tricky. I read one account about DFI BIOS, that
suggested changing one setting, then save and test. Apparently, if
you try to rush things, by just copying a script of settings into
the BIOS setup screens, you'll never get the same results twice.
With the DFI BIOS, you have to "sneak up" on your overclock.

Paul
 
Reply With Quote
 
 
 
 
Michael Brown
Guest
Posts: n/a
 
      06-16-2006, 10:35 AM
Pete wrote:
> Hi All:
>
> I have an opty 150 that seems to be quite stable (24 h prime and
> completed a 5 day calculation in MATLAB) at 256x11@1.55V on a DFI LP
> NF4 Ultra-D motherboard. The RAM/FSB is 1:1 so nothing tricky should
> be happening. I have 4 GB of dual channel OCZ gold xtc PC 4000 RAM
> with 2T, 3-3-4-8 timings. I have the RAM drive strength maxed out. OS
> is XP64. When I run memtest at these settings no errors. If I keep
> everything else constant and drop the multiplier down to 10, I get
> all kinds of memory errors in memtest. Any ideas of how going slower
> would decrease the stability, it seems to go against the central
> dogma of overclocking.


Reduce the drive strength to a point where it's stable again. More drive
strength does not mean more stable, and it's possible that with the reduced
power usage of the chip that the DRAM lines are being driven too hard and
having errors from reflections or ground bounce.

--
Michael Brown
Add michael@ to emboss.co.nz ---+--- My inbox is always open


 
Reply With Quote
 
user
Guest
Posts: n/a
 
      06-16-2006, 10:38 PM
Pete wrote:
> Hi All:
>
> I have an opty 150 that seems to be quite stable (24 h prime and completed a
> 5 day calculation in MATLAB) at 256x11@1.55V on a DFI LP NF4 Ultra-D
> motherboard. The RAM/FSB is 1:1 so nothing tricky should be happening. I
> have 4 GB of dual channel OCZ gold xtc PC 4000 RAM with 2T, 3-3-4-8 timings.
> I have the RAM drive strength maxed out. OS is XP64. When I run memtest at
> these settings no errors. If I keep everything else constant and drop the
> multiplier down to 10, I get all kinds of memory errors in memtest. Any
> ideas of how going slower would decrease the stability, it seems to go
> against the central dogma of overclocking.
>
> TIA,
>
> Pete
>
>

the only memory who can handle it well is Buffalo memory, this memory
can extend it self without causing any harm in the energy, OCZ it doesnt
seem s to be the right product to compare it, its only a reputation of
the name and anyone says (OCZ... OCZ) ... other will say who is this
ocz... i prefer what i tested and what i know more on the product then
the name... we test ocz they have a failure on following the rest of the
resquest on board,

you have to calculate the board energy and the speed it travel, the time
that the energy fluid on board and the immensity of juice it takes to
get to the memory and to calculate how much time the memory take to
absorbe the energy before returning it. with all this will conclude that
your power supply is one of the failling, second you need the cpu
drivers, and a healty bios (Phoenix are 1 on market). your calculation
on one field its no quite good.. i am not putting you down just
correcting your theory and to search more...


Overlocking its the same then taking speeds or cockaine, it will burn
the coding inside the chip.


Michael.g from Antiss
Amd Solution provider and Dev software partner.
 
Reply With Quote
 
Pete
Guest
Posts: n/a
 
      06-17-2006, 01:02 AM

"user" <(E-Mail Removed)> wrote in message
news:aqGkg.18781$(E-Mail Removed)...
> Pete wrote:
>> Hi All:
>>
>> I have an opty 150 that seems to be quite stable (24 h prime and
>> completed a 5 day calculation in MATLAB) at 256x11@1.55V on a DFI LP NF4
>> Ultra-D motherboard. The RAM/FSB is 1:1 so nothing tricky should be
>> happening. I have 4 GB of dual channel OCZ gold xtc PC 4000 RAM with 2T,
>> 3-3-4-8 timings. I have the RAM drive strength maxed out. OS is XP64.
>> When I run memtest at these settings no errors. If I keep everything
>> else constant and drop the multiplier down to 10, I get all kinds of
>> memory errors in memtest. Any ideas of how going slower would decrease
>> the stability, it seems to go against the central dogma of overclocking.
>>
>> TIA,
>>
>> Pete

> the only memory who can handle it well is Buffalo memory, this memory can
> extend it self without causing any harm in the energy, OCZ it doesnt seem
> s to be the right product to compare it, its only a reputation of the name
> and anyone says (OCZ... OCZ) ... other will say who is this ocz... i
> prefer what i tested and what i know more on the product then the name...
> we test ocz they have a failure on following the rest of the resquest on
> board,
>
> you have to calculate the board energy and the speed it travel, the time
> that the energy fluid on board and the immensity of juice it takes to get
> to the memory and to calculate how much time the memory take to absorbe
> the energy before returning it. with all this will conclude that your
> power supply is one of the failling, second you need the cpu drivers, and
> a healty bios (Phoenix are 1 on market). your calculation on one field its
> no quite good.. i am not putting you down just correcting your theory and
> to search more...
>
>
> Overlocking its the same then taking speeds or cockaine, it will burn the
> coding inside the chip.
>
>
> Michael.g from Antiss
> Amd Solution provider and Dev software partner.


Thanks for the info, I will look into it. I would not worry about burning
the codiing in the chip, because I usually get bored with a chip and replace
it before the chip fails. I have a drawer full of old but working chips.

Best wishes,

Pete


 
Reply With Quote
 
Pete
Guest
Posts: n/a
 
      06-17-2006, 01:08 AM

"Michael Brown" <(E-Mail Removed)> wrote in message
news:4492896a$0$25130$(E-Mail Removed) u...
> Pete wrote:
>> Hi All:
>>
>> I have an opty 150 that seems to be quite stable (24 h prime and
>> completed a 5 day calculation in MATLAB) at 256x11@1.55V on a DFI LP
>> NF4 Ultra-D motherboard. The RAM/FSB is 1:1 so nothing tricky should
>> be happening. I have 4 GB of dual channel OCZ gold xtc PC 4000 RAM
>> with 2T, 3-3-4-8 timings. I have the RAM drive strength maxed out. OS
>> is XP64. When I run memtest at these settings no errors. If I keep
>> everything else constant and drop the multiplier down to 10, I get
>> all kinds of memory errors in memtest. Any ideas of how going slower
>> would decrease the stability, it seems to go against the central
>> dogma of overclocking.

>
> Reduce the drive strength to a point where it's stable again. More drive
> strength does not mean more stable, and it's possible that with the
> reduced power usage of the chip that the DRAM lines are being driven too
> hard and having errors from reflections or ground bounce.
>
> --
> Michael Brown
> Add michael@ to emboss.co.nz ---+--- My inbox is always open

I hate to mess around with RAM settings too much, but will test the effect
of backing down on the drive strength.

The RAM would not boot at 250 MHz when the drive strength was in the default
or lowest setting. The system appears to be very stable with a multiplier
of 11. Its just interesting that memtest is full of errors when the
multiplier is backed down to 10. I am using the April 6, 2006 bios which
hides many of the RAM settings behind auto. I agree that some of the RAM
timings could be switching automatically, but keep in mind. The RAM/FSB is
constant and 1:1. The FSB is at 250 MHz so changing the multiplier could
change the RAM settings? Verifying in CPU-Z is a good suggestion.

I might try a another modded bios, but last time I tried the ones on the
boot CD, they all were less stable than the DFI bios.

Best wishes,

Pete


 
Reply With Quote
 
Pete
Guest
Posts: n/a
 
      06-17-2006, 01:17 AM

"Paul" <(E-Mail Removed)> wrote in message
news:nospam-1606060424190001@192.168.1.178...
> In article <ERrkg.69350$(E-Mail Removed)>, "Pete"
> <(E-Mail Removed)> wrote:
>
>> Hi All:
>>
>> I have an opty 150 that seems to be quite stable (24 h prime and
>> completed a
>> 5 day calculation in MATLAB) at 256x11@1.55V on a DFI LP NF4 Ultra-D
>> motherboard. The RAM/FSB is 1:1 so nothing tricky should be happening.
>> I
>> have 4 GB of dual channel OCZ gold xtc PC 4000 RAM with 2T, 3-3-4-8
>> timings.
>> I have the RAM drive strength maxed out. OS is XP64. When I run memtest
>> at
>> these settings no errors. If I keep everything else constant and drop
>> the
>> multiplier down to 10, I get all kinds of memory errors in memtest. Any
>> ideas of how going slower would decrease the stability, it seems to go
>> against the central dogma of overclocking.
>>
>> TIA,
>>
>> Pete

>
> Are you verifying with CPUZ, that when you change the multiplier,
> nothing else is changing ? Try verifying the settings in each case.
> (I.E. Never trust a BIOS further than you can throw it :-))
>
> DFI is known to be tricky. I read one account about DFI BIOS, that
> suggested changing one setting, then save and test. Apparently, if
> you try to rush things, by just copying a script of settings into
> the BIOS setup screens, you'll never get the same results twice.
> With the DFI BIOS, you have to "sneak up" on your overclock.
>
> Paul


Good suggestion with CPU-Z. Is the reported RAM CPU-Z speed always
accurate? In the OC guide on dfi-street, they give another algorithm using
the ceiling function for calculating the RAM speed that usually gives a
slower result than Memtest and CPU-Z.

I have been sneaking up on my overclock for sometime. However, I wanted to
backdown the multiplier for some tests when I uncovered this instability.
Because slower means more stable, I was a bit confused. I actually did the
tests several times, so I could believe in the results.

I keep hoping that I will find the right combination to hit 3.0 GHz, but the
system has been stable at 2.8 GHz with the wall at 2.85 GHz.
When I have time, I will have to test out some of the recently modded
bioses.

Thanks for the input,

Pete


 
Reply With Quote
 
 
 
Reply

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

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
DH57dd/core i3 slow boot, slow wake up cew Intel 2 05-19-2012 11:52 PM
Leopard: Slow pages, Slow e-mail. Help. nickravo@gmail.com Apple 4 10-31-2007 05:02 AM
p5n32-sli se deluxe and Vista x64 and 2gb (slow enough) then 4gb (horribly slow issues) markm75 Asus 1 08-26-2007 12:51 PM
My 6600GT is slow as crap! Or nearly as slow. Damo Nvidia 14 12-08-2004 01:54 PM
TC1100 - SP2 - SLOW SLOW SLOW! Stef Tablet PC 10 09-13-2004 09:53 AM


All times are GMT. The time now is 05:13 AM.


Welcome!
Welcome to Motherboard Point
 

Advertisment