Jump to content


- - - - -

OCZ 4000VX on NF2 Lan Party B


  • Please log in to reply
211 replies to this topic

#25 Mott13

Mott13

    Geek

  • Members
  • PipPipPip
  • 674 posts

Posted 18 February 2006 - 07:20 PM

Darien,No where in this thread did I see how much vcore your giving to the cpu.If your passing memtest at those speeds but failing prime will sometimes mean you need a lil bump in vcore.That is at least what I found reaching 250 with my double sided BH-5...

#26 medianOCer

medianOCer

    Geek

  • Members
  • PipPipPip
  • 714 posts

Posted 18 February 2006 - 08:06 PM

It sounds like with the VX you're running much tighter timings than with TCC5 RAM
that you've used. Also, with tCR(command rate) at 1 instead of 2 you're going to
get more bandwidth and tranfer rates going on there as well. So even if you're not
hitting as high of an FSB with the VX, as opposed to the TCC5, you're probably
stressing your board, CPU and etcetera more with the VX than with the TCC5.

You could look to trying something like this with your RAM timings:
11-2-2-2-2.0, 13-16 (looser), 3-3-4-3-2-2-3
or 11-2-2-2-2.0, 9-12 (tighter), 3-3-4-3-2-2-3

Those would be with modifying the tRC and the tRFC values. You could also try
varying those values somewhere in between the two ranges I posted above.
Such as, 10-13, 11-14 or 12-15 and see if that does anything for you as well.

Another area you could look at, would be altering the DIMM Driving Strengths
(DDS) and Driving Slew Rates (DDSR). I've found with mine that running them
at 3-10, 3 for the Driving Strength on each DIMM, and 10 for the Driving Slew
Rates has helped with a bit of stability and smoother running. That may or may
not be the case with your RAM. You'd just have to try and see if it would help or
not. My RAM is single sided for each stick, so the default/auto settings for those
come up as 2 on the DDS and 11 for the DDSR, just to let you know those relative
to the changes I have made.

If you RAM is double sided, then the default/auto DDS and DDSRs are probably
something like this for each DIMM slot, 4 for the DDS and 11 for the DDSR. You
could see if 5 for DDS and 10 for the DDSR would make a difference. Maybe try
the 5 value for the DDS first, then if that doesn't seem to help you could add the
change for the DDSR as well, and try setting it at 10. You could try setting that
for each DIMM slot (1 through 3), even if you have one slot that's empty.

Someone else asked what you're running for CPU voltages. Depending on what
you've been using and/or how effective your cooling is currently and also, those
would possibly be areas to consider too.

Have you looked at sinking the onboard mosfets for your CPU power supply, and
also the same for your RAM power supply mosfets and a chip or two in that area?
Maybe you've already done that. Thought I'd throw that in as well, and in case
you have yet to consider doing such.

I also find that it's easier when I report some settings I've been trying, if I
report much more detailed info on settings that there can tend to be less of a
need for additional posts related to that info. Here's an example of what I'm
currently running and is relative to the components and other info in my sig:
E-A-E, 9-2-2-2-2.0, 9-12, 2-2-4-3-2-2-3, E-E-F-E, 3-10-3-10-3-10, E-E-256MB
233MHz-x10-67MHz-1:1, 1.7v-1.6v-1.7v-3.1v
APIC on, Virus Warning disabled, CPU Thermal-Throttling disabled
Using the ACPI (ACPI) PC hal setting in Windows.

APIC being set to disabled might be better for you, as you're trying for a higher
FSB than I'm running consistently. Nice to see that you tried it enabled, though,
and just to see if helped or hindered. Always worth a shot. Thanks for the
feedback on what you found too. ;)

Trying a single different setting or variation at time can also make it easier to
find out what might make a difference with a stable OC, or at least make it much
easier to figure out if you would need to look to another area/setting(s) in order
to obtain stability and etc.

Sorry if I've thrown too much out on the table. Maybe I didn't..? I'm hardly trying
to make anything confusing, there's just quite a few areas of configurability with
these boards, and others as well. Just attempting to cover a few areas at a time,
and hoping that maybe something will help in getting you where you're wanting
to go. :)

Back later...

#27 kitfit1

kitfit1

    Only Clocking When Awake

  • Members
  • PipPipPipPipPip
  • 3721 posts
  • Location:UK

Posted 19 February 2006 - 12:23 AM

update 243FSB:
failed prime 95 in 58 min (which seriously pissed me off :mad: ) that was odd maybe a fluke... passes SnM long test no errors... just finished S&M... trying prime 95 again... (GRRRR!! AHHH!!! WTF! fails prime in 2 minutes!!!)


What changes did you make to pass S+M long memory test? as you were'nt passing it before.

Ryzen 2700x @ 4.2Ghz

Asus CH6

32Gb B-die @ 3333 mhz

Asus 1080ti


#28 Guest_Darien_*

Guest_Darien_*
  • Guests

Posted 19 February 2006 - 01:11 AM

Darien,No where in this thread did I see how much vcore your giving to the cpu.If your passing memtest at those speeds but failing prime will sometimes mean you need a lil bump in vcore.That is at least what I found reaching 250 with my double sided BH-5...

I'm currently pushing 1.775vcore... which is too much frankly... it was doing 250x10 @ 1.70 before with the Tcc5... but then It started to error so I pushed it up to 1.750vcore.... i pushed it up even further to 1.775 vcore to make sure there is plenty juice there to r/o cpu issues...
so 1.775vcore should be way more than enough for 243x10... good thought though, I will try higher for crap and giggles... but I don't think thats it...

#29 Guest_Darien_*

Guest_Darien_*
  • Guests

Posted 19 February 2006 - 01:13 AM

What changes did you make to pass S+M long memory test? as you were'nt passing it before.

dropped from 250 to 243 :rolleyes: nothing changed.... I thought I would drop down to 243 for a prime burn in like you said... when i found out it could prime 243 for 11 minutes I said no way and tired the S&M long test... which it passed no problem at 243.... but not 250

#30 Guest_Darien_*

Guest_Darien_*
  • Guests

Posted 19 February 2006 - 01:18 AM

medianOcer:
thx for the input, I give that stuff a try I'll post all of my detailed settings and maybe even a pic of my mobo... it has puney sinks on the fets... things could and will be cooler once I get to it...

due to all the issues I'm having with this board, I've got another DFI LP B on the way and I'm gonna try it out and see how it does, mainly with the SATA issues.. but it also occured to me... it may not be that my particular VX ram is picky but possilby my memory controller... so maybe the other board will handle it better...

I'll take the better of the boards and sell the other... the board I keep will be modded and sinked and pampered and spoiled like an only child...

but for now I'm not overly concerned with heat as I don't believe that's the issue...

I'll post more later, had a long long day and just got home bedtime... can't play

#31 Mott13

Mott13

    Geek

  • Members
  • PipPipPip
  • 674 posts

Posted 19 February 2006 - 08:59 AM

I'm currently pushing 1.775vcore... which is too much frankly... it was doing 250x10 @ 1.70 before with the Tcc5... but then It started to error so I pushed it up to 1.750vcore.... i pushed it up even further to 1.775 vcore to make sure there is plenty juice there to r/o cpu issues...
so 1.775vcore should be way more than enough for 243x10... good thought though, I will try higher for crap and giggles... but I don't think thats it...

think again many will pump 1.85-1.95 to reach 2500-2700 speeds,personally I used 1.8 24/7 give it a try you wont kill that chip...

#32 Guest_Darien_*

Guest_Darien_*
  • Guests

Posted 19 February 2006 - 09:59 AM

think again many will pump 1.85-1.95 to reach 2500-2700 speeds,personally I used 1.8 24/7 give it a try you wont kill that chip...

I appreciate your help dude... but I've had 4 different xp-m chips, and I'm aware of what you're saying... but 'this' chip was doing 250x10 with my tcc5 at 1.70vcore and 255x10 1.75vcore....
I was using 1.775vcore to do 243x10 with the vx and it was failing... so I tend not to think it was a vcore issue... but willing to try i bumped it up to 1.85v for 240x10.... blew out in less than 1 minute in prime95...

yet still able to bench 3dmark01 no problem at 250x10

as an update I primed 200 x 10 overnight and 8 hours later it was still going... so I bumped it up to 230 x 10 and it's done about 5-10 minutes so far... so hopefully that will work and we'll go up from there... hopefully...

geez... maybe I should just ask King to make me that toaster sig 'as is'... could be a while before I get my OC how I want it...

#33 Mott13

Mott13

    Geek

  • Members
  • PipPipPip
  • 674 posts

Posted 19 February 2006 - 10:59 AM

ok at least you tried right:)

#34 Guest_Darien_*

Guest_Darien_*
  • Guests

Posted 19 February 2006 - 12:24 PM

230x10 failed prime95 50 minutes...

I can't believe there is a burn in that will take it from 200 prime stable to 250 prime stable when everything else if completly fine at 250 and S&M at 240....

there must be a drive strength or slew setting or something that needs to be changed and then BAM it will be solid at 240 - 250....

because it will do memtest great above 250 for 8 hours with no errors... it has to be some setting somewhere to make it happy with windows... will post detailed config settings and a picture in few hours.
http://www.dfi-stree...ntid=3780&stc=1

#35 Guest_Darien_*

Guest_Darien_*
  • Guests

Posted 19 February 2006 - 02:42 PM

DFI bios settings

http://www.dfi-stree...ntid=3781&stc=1
http://www.dfi-stree...ntid=3782&stc=1
http://www.dfi-stree...ntid=3783&stc=1
http://www.dfi-stree...ntid=3784&stc=1
http://www.dfi-stree...ntid=3785&stc=1

#36 medianOCer

medianOCer

    Geek

  • Members
  • PipPipPip
  • 714 posts

Posted 19 February 2006 - 05:38 PM

Thanks for the images.

Have you tried Sync Mode Memory Bypass set to Enabled?

Looks like you've got the mosfets for the CPU and DIMM sinked, nice.

There's a small chip directly underneath where the Video Card overhangs,
and about an inch away from the AGP latch, it's about 5mm and square in
shape. That chip gets hotter than the RAM and CPU mosfets.
Have you sinked it?