DFI ICFX3200 T2R/G FSB Scaling & Voltage Settings

[The Nemesis]

I've done some extensive testing with the RD600 and have compiled the data into a quick chart to help others in setting up quick overclocks on their board. Memory was left async, DDR2-800, 2T for most of the tests. All testing done on open test bed (HSPC Top Deck Tech Station). North bridge cooled by Microcool NorthPole XE Whisper Chipset Cooler, cpu cooled by Tuniq Tower with stock fan. Your results may vary, but these were the exact settings used to reach each final result.

CPU: E6400 L628B194
Memory: G. SKill 6400 HZ's
Video: Evga 7900GT's
PSU: OCZ GameXtreme 700w
Case: HSPC Top Deck Tech Station

**** Use these voltages & settings at you own risk****


It has been said by other more knowledgable enthusiasts that 1.5v on the nb is the max safe 24/7 voltage.



Bios Settings

[HTML]
Genie Bios

>Performance options

Memory Clock Frequency: DDR-400
Control CPU Frequency: Manual
x CPU Frequency (Mhz): 4XX
NB FSB Strapping: Auto or 333
PCI-E Overclock: 105
Software Reset Clockgen: Enabled
Memory Overclocking: Disabled
Overclock Frequency(MHz):

>NorthBridge ASIC CFG
>NorthBridge Selections
Memory Channels: Auto
Memory Power Mangement: Disabled
IGD Clock Gate: Disable
NB Channel Mapping: Auto
Simple Addr Translationisable
Chip-Select Swap: Disable
Burst Length: 8
Bank0 Map GFX: 4
Bank1 Map GFX: 5
Bank2 Map GFX: 6
Channel Mapping NB: 5
Channel Mapping GFX: 1
Bank0 Map NB: 5
Bank1 Map NB: 11
Bank2 Map NB: 7

>Memory Timing Menu

CS Output Timing: Delay 1/4 clock
CK Output Timing: AUTO
Dual channel Split Mode: AUTO
ADDR/CMD Driving Pre-AMP: Disabled
ADDR/CMD Driving Strength: 05
Clocking Driving Pre-AMP: Always
Clocking Driving Strength: 05
Address/Command Timing: Auto
Memory Timing Select: Manual
CAS Timing (TCAS): 4
Read RAS-CAS Delay (TRCDR): 4
Write RAS-CAS Delay (TRCDW): 4
RAS Precharge Time (TRP): 4
Cycle Time (TRAS): 12
Row Cycle Time (TRC): 22
Refresh to Active (TRFC): 42
Bank to Bank CMD (TRRD): 3
Write Recovery (TWR): 6
Read-Write Command (TRTW): 3
Write-Write CS (TWTW_CS): 2
Write-Read Command (TWTR): 3
Write-Read CS (TWTR_CS): 2
Read-Read CS (TRTR_CS): 2
Read Data Transfers: 5

Memory Auto-Calibration
Read Data Transfer: Enable
Read Strobe Delay: Disable
Read Strobe Arrival: Enable

>FSB IO Strength

FSB IO Driving COntrol: Default
x 1X NMOS Strength
x 1X PMOS Strength
x 2X NMOS Strength
x 2X PMOS Strength
x 4X NMOS Strength
x 4X PMOS Strength



CPU Vid Control: 1.**v
CPU VID Special ADD: Auto
DDR2 Ram 1.8 Voltage: 2.25v
CPU VTT 1.2V Voltage: 1.**v
NB Core 1.2V Voltage: 1.**v
NB PLL 1.8V Voltage: *.**v
NB PLL 1.2V Voltage: 1.**v
NB PCI-E 1.2V Voltage: 1.**v
NB Core 1.2V convert from: *.**v
SB Core 1.2V Voltage: 1.**v
Clockgen Voltage: *.**v
GTL REF Voltage control: Enabled
CPU Core 1/2 GTL REF Volt: 48
CPU Core 3/4 GTL REF Volt: 48
North Bridge GTL REF Volt: 128
CPU PLL Setting Overide: Enabled
PLL Override loop count: AUTO
Conroe P-State Multiplier: x8
C1E Function: Disabled


[/HTML]

445FSB 4 hrs orthos ---465FSB 4 hrs Orthos----470FSB 4hrs orthos----480FSB 2 hrs orthos---478FSB 12hr Orthos---488FSB 2 hrs Orthos

[Kain XS]

Stock Cooling on NB?

Adding some fans?


Nice post.

[p0tter]

Nice post and seems helpful.

[The Nemesis]

Stock Cooling on NB?

Adding some fans?


Nice post.

oops I edited the pic to make is smaller and removed the configuration. Added the info original post.

[rick_fx]

For 515fsb, I used 2.03V NB Core (stock cooling), the NB pll's were maxed, SB core at 1.34V and Clockgen VTT at 3.40V

Past 515fsb Orthos stable is not possible for me, even though my e6700 and e6850 are capable of higher fsb on p5b-dlx.

http://i12.tinypic.com/34esrqf.jpg

Sorry for pic only being at 4minutes
I stopped at 1h 20mins, that was stable enough for me

[The Nemesis]

For 515fsb, I used 2.03V NB Core (stock cooling), the NB pll's were maxed, SB core at 1.34V and Clockgen VTT at 3.40V

Past 515fsb Orthos stable is not possible for me, even though my e6700 and e6850 are capable of higher fsb on p5b-dlx.

http://i12.tinypic.com/34esrqf.jpg

Sorry for pic only being at 4minutes
I stopped at 1h 20mins, that was stable enough for me

Nice, I can run orthos @ 500FSB with a lower multiplier (7), but it took over 2v on the NB as well. I decided not to post any results over 500FB as the voltages were a bit extreme & I'd hate for someone to fry their board using my data

[BlueAqua]

Nice guide, very informative. Thanks for taking the time to put this together, makes things just a bit easier for this motherboard. I think I'll try some of your combinations to see how they work for me.

[nealh]

hi..I saw you set pcie overclock to 105...why?

I did this ony my SLI-DR mobo with an ATI card(x800xl as it was recommended by AG)....but you have an nvidia card

on my RD600 with x1900xt..with my 24hr prime95 stable OC I crash 3dmark05 in the cpu tests..I can run it one out of 5 times....

go figure this one..I was using free version of 3mark05...crashes
I used 3dmark06...no crashes, using ATI Overdrive x1900xt CF set to 675/792..I get 6874

I am not sure but it seems on my setup..ATI Tray tools will not play well with this mobo

[The Nemesis]

hi..I saw you set pcie overclock to 105...why?

I did this ony my SLI-DR mobo with an ATI card(x800xl as it was recommended by AG)....but you have an nvidia card

on my RD600 with x1900xt..with my 24hr prime95 stable OC I crash 3dmark05 in the cpu tests..I can run it one out of 5 times....

I run the 7900GT's @ 700 core 850 memory which yield performance similar to 7900GTX's. They wouldn't run 3DMark @ those clocks without 105-110mhz on the PCI-e bus back on my other lan party boards so I just set it to that when I first configured this board. I'm not sure if it helps or hurts overclocking this board, the gpu's just don't run above 690mhz without it.

[nealh]

go figure 3D Mark06 is running fine..but forget 3Dmark05...crashes....

[goatboy]

Thanks for the post N. A lot of work and very informative.

[Cornelious0_0]

Thanks for the information Nemesis...very thorough and informative...just what I need with my E4300/ICFX3200 on the way.

[Mul]

Very informative. I have a DFI LanParty ICFX3200 which I bought for a rather nice price, winging it's way to me next week but couldn't find much about how well the board clocked at given voltages, so this was exactly what I was looking for!

I'll be using a Core 2 Duo E6300 so unfortunately, I'll need high FSB's without a doubt. I'm going to be planting a HR05 Non SLI Northbridge Cooler on the board with a 70mm fan ripped from an AMD stock cooler as I hear the RD600 runs silly hot.

I'll be using a 29B CPU so it should clock well. Looks like it's all down to how high a FSB the board will pull. Has anyone actually killed their RD600 in a short period of time by using NB Volts upwards of 1.6V?

Cheers,

Mul.

[Cornelious0_0]

I'm pretty sure with after market cooling 1.6v should be alright, it just sounds like anyone I've seen running it close to 2.0v (to break the 511FSB "barrier") is asking for trouble though.

I'll probably end up referencing this guide, but I also just found a post by RyderOCZ on OCZ's forums with memory timing setup using the ICFX3200, the same 2GB OCZ PC2-8500 DDR2 I'll be using, and an E6600 with the same video cards and everything. Since the E4300 runs 266FSB just fine I'll pretty much be "set" to get the board/memory running perfectly and with a slight boost on the CPU until I have time to really test it.

[nealh]

ok beware...1.6v on NB can shorten its life ..this from Tony who has stated DFI says over 1.5v and the NB can die and have shorten life span

[Cornelious0_0]

ok beware...1.6v on NB can shorten its life ..this from Tony who has stated DFI says over 1.5v and the NB can die and have shorten life span

Ah yes...I do remember him saying that, sorry.