The DFI NF4 stable settings thread

[Electric Bill]

Thanks - I figured it was something like that.

Bill

 

[Electric Bill]

Nice one Bill! Welcome to the Forums!

That Opty of yours is screaming fast! Now, do you have to increase the Drive strength to 6? Could you try 8 and 1?

I tried running 10x295 with the settings I posted and crashed out of 3CMark01 after about 1 min. Then I tried the same thing with DRAM Drive Strength 1, 4, 5, 6, and 8. Drive Strength 5 seemed to run the longest before crashing. I also tried to go up a bit on Vcore with no luck. Then I saw a link to a thread at DFI-Streeet that recommended Max Async Latency - 09.0 and Read Preamble Time - 07.5 so I tried it and still crashed 3DMark. Next I raised the Vcore .1v and saw my temps jump up 4°C immediately but still crashing. So I put the setting back to where they were.
I figure that 2.9GHz is a nice oc and that's where I'll stay.

Bill

 

[harlam357]

Finally got somewhere. ...will continue to update, have new PDP XBLK4800 coming soon, so this config will undoubtedly change.

My own personal quest for A64/DFI speed, help me out... - http://www.ocforums.com/showthread.php?t=425758

Last Updated:December 4, 2005
Motherboard: Lanparty Ultra-D
Bios: 623-3
CPU: 170 Opteron CCBWE 0530TPMW
VC: eVGA 7800GT
OS: Windows XP Home w/SP2
Ram: Mushkin XP3200 (991355) (TCC5) - Orange Dimm slots
Chipset Drivers: Revision 6.70 (w/out IDE or Firewall Drivers)

DRAM Timings:

• Ram Multi 200 1:1
  CPC ENABLED
  Tcl 2.5
  Trcd 04
  Tras 08
  Trp 04
  Trc 08
  Trfc 16
  Trrd 02
  Twr 02
  Twtr 01
  Trwt 02
  Tref 4078 (166Mhz @ 1.95us)
  Twcl AUTO (01)
  Bank Inter ENABLED
  Dynamic Skew Control AUTO (Off)
  Skew Value 0
  DRAM Drive Strength 07
  DRAM Data Drive Strength 02
  DRAM Response Time FAST
  Max Async 8ns
  Read Preamble 5.5ns
  Idle Cycle 256
  Dynamic Counter DISABLED
  R/W Que Bypass 16X
  Bypass MAx 7X
  32 Bit Gran DISABLED(4 bursts)

Voltages etc:

• HTT 270
  LDT X3
  Cpu Multi X10
  PCI-E 100
  Vcore Startup 1.400
  Vcore 1.400
  Vcore Special AUTO
  LDT 1.3v
  Chipset 1.6v
  Vdimm 2.8v

 

[Electric Bill]

harlam - have you tried the 180MHz divider?
My RAM doesn't like to go above 265 for very long.

Bill

 

[harlam357]

Noticed that you're running a divider on your rig... I'm not holding back from using a divider to run faster on the cpu, I really don't want to go above 2.7Ghz right now. Seems like a nice spot, especially since I'm only barely overvolting my $400 cpu. Hoping for 300x9 1:1 with the new PDP TCCD- already shipped!

I understand that I could try to run a divider... question, would I see a marked improvement in memory bandwidth if I ran 300x9 on a 9/10(180Mhz) divider instead of 270x10 @ 1:1? Would a 9/10 divider not yield 270Mhz on the ram @ 300HTT? Over 270 these Mushkins just won't handle it. Obviously I haven't tried it. Since it's pretty much impossible to saturate the HTT bus, I didn't think that I'd see any different between 270x3HTT vs. 300x3HTT, but maybe that's my error.

 

[Electric Bill]

I understand that I could try to run a divider... question, would I see a marked improvement in memory bandwidth if I ran 300x9 on a 9/10(180Mhz) divider instead of 270x10 @ 1:1? Would a 9/10 divider not yield 270Mhz on the ram @ 300HTT?

On paper perhaps not.
We're both running 2x512MB TCC5 at about the same timings.
Your Sandra Mem Bandwidth is 7115/7051.
When I ran my Venice at 10x280 without a divider my bandwidth was 7305/7247.
Now at 10x290 with a 180MHz divider its 7459/7384.
I never considered using a divider before since my OCZ ran for 2 months at 280MHz without any problem. Then the mem controller started to fail and I had to drop the cl to 3.0 to even run at 270MHz. When I installed the Opty I decided to try the 180 divider and was pleasantly suprised at the results.
Just wondered if you'd tried it?

Bill

 

[harlam357]

^^^^^
I see that... you're getting much better bandwidth with the HTT bus clocked higher and your ram clocked slower. Some might argue that this is the result of raw clock speed... and it may be. I'm still figuring out the quirks of an A64 system. This falls inline with the statements that I read regarding the HTT bus, just clock it as high as you can, within standard tolerance, and run your ram on a divider... most of these folks obviously know what their talking about. But the wider HTT bus must also improve the efficiency of the ram in general. We'll find out what happens on my rig tonight. I'll try the 300x9 config holding my ram @ 270Mhz. and see if I get any additional bandwidth. This way the cpu & ram will stay static and the only thing to change will be HTT.

 

[harlam357]

Did not bother taking a screen shot... bandwidth did not change @ 300x9 vs. 270x10 w/memory held to 270Mhz. Seems as though your increased bandwidth is due to your higher cpu clock Bill.

 

[Electric Bill]

harlam - now that you know that the divider doesn't hurt bandwidth why not run at 10x280 on a 180 divider so its not the RAM that's holding you back?

Bill

 

[harlam357]

I'm gonna wait until I get my Patriot PC4800's before I really start tweaking again. Must FOLD! I've got folks gunnin' for me.

I'm still hoping for 300x9 1:1 or maybe a little better with the new PDP's. My 170 is already @ 43-45C under load. I don't want to run it much hotter than that. 2 cores @ 2.7Ghz is fine w/me.

 

[jaydoggg]

Hi guys, first time posting here. I came here to view some memory configurations to try. I'm running the 6/23-3 bios currently. Have tried the 7/04-3 but had problems. Anyways, you guys have some killer overclocks and I'm hoping to join ya but i'm stuck at 220 mhz HTT. I'm trying to decide whether my LanParty Ultra D is just defective. My PSU is a blue storm 500w (Rgone and angry-games @dfi-street.com both recommended) and my memory was just at a friends house of mine running stable @284 HTT - 299 max with windows crashing after 2 hours. So, it must be my inexperience or the board is just bad. Right now i'm at the point where anything over 220 HTT I get memtest errors on 4,5, & 8. I have alot of work ahead but if any of you guys are willing to help or give a few suggestions i'll be very happy to try them. This thread appears to be for good overclocks only and not for helping a newbie so I politely NOT post here again but Please direct me to a thread where I can get some help...

Thanks in advance,
-Jaydoggg

 

[Electric Bill]

Try these:

DFI NF4 (Ultra or SLI) VX, TCCD/TCC5, BH-5 & General INFO - scroll down the page for the one that matches your needs.
TCCD DDR600+ DFI nF4 Ultra/SLI guide
Overclocking with G.Skill TCCD - screen shots are on the next page.

Bill

 

[infinitevalence]

6-23, 3800+, Redline (UTT)
-----------------------------------------------------------------------

Last Updated:December 10th 2005
Motherboard: Lanparty Ultra-D
Bios:6-32
CPU: 3800+ Manchester
VC: Sapphire X800GTO² (16p 540c 600m)
OS: Windows XP pro 32bit
Ram: Mushkin Redline UTT Slots #2,#4

DRAM Timings:

• Ram Multi 200 1:1
  CPC ENABLED
  Tcl 2
  Trcd 3
  Tras 5
  Trp 02
  Trc 7
  Trfc 16
  Trrd 3
  Twr 2
  Twtr 1
  Trwt 03
  Tref 1563 (200MHz 15.6us)
  Twcl 01
  Bank Inter EN
  Dynamic Skew Control Auto (ON)
  Skew Value 100
  DRAM Drive Strength 08
  DRAM Data Drive Strength 03
  Max Async 7ns
  Read Preamble 4.5ns
  Idle Cycle 16
  Dynamic Counter DIS
  R/W Que Bypass 16X
  Bypass MAx 7X
  32 Bit Gran DIS

Voltages etc:

• HTT 260
  LDT X4
  Cpu Multi X10
  PCE-I 101
  Vcore Startup 1.425
  Vcore 1.500
  Vcore Special 110%
  LDT V 1.3
  Chipset V 1.7
  Vdimm 3.5

 

[jaydoggg]

Try these:

DFI NF4 (Ultra or SLI) VX, TCCD/TCC5, BH-5 & General INFO - scroll down the page for the one that matches your needs.
TCCD DDR600+ DFI nF4 Ultra/SLI guide
Overclocking with G.Skill TCCD - screen shots are on the next page.

Bill

THanks for the links bill. I had some luck with them but I have problems getting windowz to load at higher HTT than stock. THis has been a problem all along. I get NTLDR cant load or NTKRNL cant load OS. Or i get NT.sys corrupt. This happened before when i was getting 220 HTT stable then I cant load windows. So, i got a stable clock of 220 HTT and decided to install windows again. Couldnt do that cuz of BSOD's. Had to disable cache in the bios. This caused windows to install in 9 hrs. I am really regretting this damn Lanparty Ultra D board. I have never had any problems with asus or abit. Does anyone think the board can be bad? I'm running at 250 HTT 2.5-3-3-8 2.9 vdimm 4078 tref and pass memtest 10 times complete. Then, No windows. i'm running winxp pro X64 if that helps.
thanks

 

[Electric Bill]

jaydoggg - I suggest you start over again.
This thread should help.

Bill

 

[Electric Bill]

Got some new OCZ Platinum Elite PC4800 (DDR600) day before yesterday and a new OCZ Powerstream 520 PSU.
Took it up to 3.03GHz with the RAM/CPU at 1:1. Here's the result.

Then to make it official I ran Prime over 12 hrs at 3.00GHZ 1:1 .

It does run a bit warm so I'll probably keep it at 2.9 or 2.95GHz for everyday use.
NOTE: I changed to Bigfoot's 714-2BTA.

Beta BIOS 714-2BTA
Genie BIOS Setting
_______________________
>DRAM Configuration Press Enter
FSB Bus Frequency 300
LDT/FSB Frequency Ratio x3.0
LDT Bus Transfer Width d16u16
CPU/FSB Frequency Ratio x10.0
PCI eXpress Frequency 101MHz
K8 Cool'n'Quiet Support Disable
xCool'n'Quiet MAX FID Auto
CPU VID StartUp Value 1.425V

CPU Core Voltage 1.48V
LDT Bus Voltage 1.30V
Chip Set Voltage 1.70V
DRAM 2.5 Voltage 2.74V
CPU VID Control 1.375V
CPU VID Special Control Above VID * 110%
LDT Voltage Control 1.30V
Chip Set Voltage Control 1.70V
DRAM Voltage Control 2.70V
DRAM+ .03V If it's not 3.2V Disable

PCI eXpress Configurator X2 X1 X16 X1
MAC Lan Disabled
MAC Media Interface Pin Strap
Machine MAC (NV) Address Disabled
xMAC (NV) Address Input Press Enter
MAC Lan Boot ROM Disabled
Internal Phy SATA 1/2 Enabled
Internal Phy SATA 3/4 Disabled
VIA 1394 Control Disabled
MARVELL GigaLAN Control Disabled

Run Mem Test 86+ Disabled

DRAM Configuration
_______________________
DRAM Frequency Set (MHz) 200 (DRAM/FSB:9/10)
Command Per Clock (CPC) Enabled
CAS Latency Control (Tcl) 2.5
RAS# to CAS# delay (Trcd) 04 Bus Clocks
Min RAS# active time (Tras) 08 Bus Clocks
Row precharge time (Trp) 03 Bus Clocks
Row cycle time (Trc) 08 Bus Clocks
Row refresh cyc time (Trfc) 16 Bus Clocks
Row to Row delay (Trrd) 03 Bus Clocks
Write recovery time (Twr) 03 Bus Clocks
Write to Read delay (Twtr) 02 Bus Clocks
Read to Write delay (Trwt) 03 Bus Clocks
Refresh period (Tref) 3120 Cycles
DRAM Bank Interleave Enabled

DQS Skew Control Increase Skew
DQS Skew Value 0
DRAM Drive Strength weak 3
DRAM Data Drive Strength Level 3
Max Async Latency Auto
DRAM Response Time Fast
Read Preamble Time Auto
Idle Cycle Limit 256 Cycles
Dynamic Counter Disable
R/W Queue Bypass 16x
Bypass Max 04x
32 Byte Granularity Disable (4 Bursts)

 

[RotKT]

thanks

 

[harlam357]

New results w/Patriot TCCD.

Last Updated:December 16, 2005
Motherboard: Lanparty Ultra-D
Bios: 623-3
CPU: 170 Opteron CCBWE 0530TPMW
VC: eVGA 7800GT
OS: Windows XP Home w/SP2
Ram: Patriot XBLK-PC4800 (TCCD) - Orange Dimm slots
Chipset Drivers: Revision 6.70 (w/out IDE or Firewall Drivers)

DRAM Timings:

• Ram Multi 200 1:1
  CPC ENABLED
  Tcl 2.5
  Trcd 03
  Tras 06
  Trp 03
  Trc 12
  Trfc 14
  Trrd 03
  Twr 03
  Twtr 02
  Trwt 03
  Tref 3072 (200Mhz @ 3.9us)
  Twcl AUTO (01)
  Bank Inter ENABLED
  Dynamic Skew Control INCREASE
  Skew Value 255
  DRAM Drive Strength 07
  DRAM Data Drive Strength 02
  DRAM Response Time NORMAL
  Max Async 8ns
  Read Preamble 5.5ns (set to 5.0 in bios )
  Idle Cycle 128
  Dynamic Counter DISABLED
  R/W Que Bypass 16X
  Bypass MAx 7X
  32 Bit Gran DISABLED(4 bursts)

Voltages etc:

• HTT 300
  LDT X3
  Cpu Multi X9
  PCI-E 100
  Vcore Startup 1.400
  Vcore 1.400
  Vcore Special AUTO
  LDT 1.3v
  Chipset 1.6v
  Vdimm 2.7v

 

[DragonFire]

Here are my current settings, anyone see anything I can change to improve things? The memory chips that my sticks use are by Aeneon, who the hell is that? I have been unable to find any useful info about these chips, so I have no idea if I should treat them like TCCD or UTT....

Motherboard: Lanparty SLI-D
Bios: 7.04-2BTA
CPU: Opteron 148
VC: XFX 7800GT
OS: Windows XP Pro w/SP2
Ram: 2GB OCZ Platinum EL (1GBx2)
Chipset Drivers: Revision 6.70 (w/out IDE,Lan,Sound Drivers)

DRAM Timings:

• Ram Multi 166
  CPC ENABLED
  Tcl 2.5
  Trcd 03
  Tras 06
  Trp 03
  Trc 14
  Trfc 16
  Trrd 03
  Twr 03
  Twtr 02
  Trtw 03
  Tref Auto (166Mhz 3.9us in A64 tweaker)
  Twcl 01
  Bank Inter EN
  Dynamic Skew Control Auto (Off)
  Skew Value 0
  DRAM Drive Strength Normal3
  DRAM Data Drive Strength Level 3
  Max Async 8ns
  Read Preamble 6ns
  Idle Cycle 64
  Dynamic Counter EN
  R/W Que Bypass 8X
  Bypass MAx 4X
  32 Bit Gran DIS

Voltages etc:

• HTT 300
  LDT X3
  Cpu Multi X10
  PCE-I 100
  Vcore Startup 1.375
  Vcore 1.460
  Vcore Special 110%
  LDT V 1.4
  Chipset V 1.6
  Vdimm 3.0

 

[Woody29]

Well this what I run with at 2.7ghz. prime stable 10hrs stopped not failed.
Dram was the most important for me under 2.9v no go, only 1 1/2hrs prime but 2.9v all is fine and dandy. Even though my ram passes 275mhz memtest at 2.7v strange.

Motherboard: LanParty NF4 UT SLi-D
BIOS: 632-2
Corsair Twinx 4400 DDR55o C25PT 2x512mb

3700+ San Diego CABGE 0515 2.7GHZ
FSB ............: 270
LTD/FSB ........: Auto
CPU/FSB ........: X 10
PCI ............: 100MHZ

CPU Vid Controls .......: 1.375v
CPU Vid Special control : 110%
LTD Voltage ............: 1.2 V
Chipset Voltage control : 1.5V
Dram VoltaGe ...........: 2.9V

DRAM Frequency: 200MHz
CPC : Enabled
tCL : 2.5
tRCD: 4
tRAS: 8
tRp : 4
tRC : 11
tRFC: 16
tRRD: 3
tWR : 3
tWTR: 2
tRWT: 3
tREF: 3072
tWCL: Auto
Dram Bank Interleave: Enable

DQS Skew Control .........: Increase
DQS Skew Value ...........: 0
DRAM Drive Strength ......: 7
Dram Data Drive Strength..: 2
Max Async Latency ........: 7
Dram Response Time .......: Normal
Read Preamble Time .......: 5
Idle Cycle Limit .........: 256
Dynamic Counter...........: Disabled
R/W Queue Bypass .........: 16x
Bypass Max................: 7x
32 Byte Granulation ......: Disable

Here is my settings for 2.75ghz, pretty much the same mostly voltage setting changes. This is only 5hrs prime passed, not failed stopped by myself.
For this speed It needed 3V Dram otherwise I only get around 1hr prime stability. Not to sure if this is to high voltage for TCCD. but it seems to like it.


3700+ San Diego CABGE 0515 2.75GHZ
FSB ............: 275
LTD/FSB ........: Auto
CPU/FSB ........: X 10
PCI ............: 100MHZ

CPU Vid Controls .......: 1.45v
CPU Vid Special control : 110%
LTD Voltage ............: 1.3 V
Chipset Voltage control : 1.6V
Dram Voltage ...........: 3.0V

DRAM Frequency: 200MHz
CPC : Enabled
tCL : 2.5
tRCD: 4
tRAS: 8
tRp : 4
tRC : 11
tRFC: 16
tRRD: 3
tWR : 3
tWTR: 2
tRWT: 3
tREF: 3072
tWCL: Auto
Dram Bank Interleave: Enable

DQS Skew Control .........: Increase
DQS Skew Value ...........: 0
DRAM Drive Strength ......: 7
Dram Data Drive Strength..: 3
Max Async Latency ........: 7
Dram Response Time .......: Normal
Read Preamble Time .......: 5
Idle Cycle Limit .........: 256
Dynamic Counter...........: Disabled
R/W Queue Bypass .........: 16x
Bypass Max................: 7x
32 Byte Granulation ......: Disable