DFI RD600 overclocking tutorial and an demostration of DDRII 1200 achievement

Recently, there are two new chipset released in market.
They are Nvidia’s 650/680i and ATI’s RD600.
The performance of both the board is real good.

DFI LANParty UT ICFX3200 T2R/G adopts ATI RD600 chipset

http://img67.imageshack.us/img67/4228/drd6001rx2.jpg

http://img149.imageshack.us/img149/9185/drd6002tg6.jpg


Most of the enthusiasts realized that DFI LANParty series has a very good OC
performance.
Moreover, BIOS settings is also fabulous…

Frankly, I’m personally not drive the board well from very beginning. But after
I’ve found the knack to this board, I was getting better and better each
time I OC.
I’d like to share some tricks about BIOS of this board to you guys. We can
start from FSB500 and DDRII 1200 as a basis. Coz, DRAM settings of this
mobo is totally separate from CPU.
So, if you got a very good memory, you might capable to create a new
history record.

Let’s get to the point here:
About voltage setting, the most important thing that we have to do is to rise
up FSB.
Therefore, you have to take these voltage options: CPU VTT, NB CORE, NB
PLL, and NB Core.

http://img67.imageshack.us/img67/1126/rd600b1lp0.jpg


CPU’s FSB 500/ DRAM frequency: 601/ PCIE 100.

http://img208.imageshack.us/img208/7817/rd600b2vz6.jpg


The DRAM adjusting options:
DDRII 1200 CL4-4-4-9 2T, the others of below options are the “performance optimized” options.

http://img412.imageshack.us/img412/6757/rd600b3td3.jpg

http://img57.imageshack.us/img57/6333/rd600b4cp9.jpg


The detail of my test bed
CPU:INTEL Core 2 Extreme X6800
MB: Nvidia nForce 680i SLI
DRAM:DFI LANParty UT ICFX3200 T2R/G
VGA:ELSA GLADIAC 790GS PH2 256DT
HD:Seagate 7200.7 160GB
POWER:Corsair HX620W Modular Power Supply
Cooler:TT Big Typhoon VX

FSB500/ DDRII1200 is the starting point.

The first memory modual:
Mushkin Tuner XP Redline
The default setting: DDRII 1000 CL4 5-4-11

http://img57.imageshack.us/img57/9329/drd6003pi8.jpg


DDRII 1200 CL4 4-4-9 2T 2.54V(BIOS:2.57V)
Super PI 32M finished

http://img412.imageshack.us/img412/5...m1200pigd4.png


The second memory modual
Corsair TWIN2X2048-6400C3
Default: DDRII 800 CL3 4-3-9

http://img57.imageshack.us/img57/3519/drd6004mz6.jpg


DDRII 1200 CL4 4-4-9 2T 2.49V(BIOS:2.52V)
Super PI 32M finished

http://img412.imageshack.us/img412/1...c1200piwu7.png


The 3rd memory modual:
Patriot PDC21GB8000+XBLK
Default: DDRII 667 CL3 3-3/DDRII 1000 CL4 4-4

http://img208.imageshack.us/img208/7827/drd6005aq8.jpg


DDRII 1200 CL4 4-4-9 2T 2.52V(BIOS:2.54V)
Super PI 32M finished

http://img412.imageshack.us/img412/9...p1200pidn7.png


Actually, after I’ve set the 1st & 2nd memory to run in DDRII 1200(1GB x 2)

Then, the 3rd one I tried whether the voltage can be a bit lower or not (by
512MB x 2)
Finally, it ends up with a “about the same” result. It seems this model is not
better than the 1st and 2nd one.


The 4th memory modual
CORSAIR Dominator TWIN2X2048-8888C4DF
Default: DDRII 1111 CL4 4-4-12

http://img57.imageshack.us/img57/9058/drd6006uc2.jpg


CPU is in FSB 511status
DDRII 1200 CL4 4-4-9 2T 2.35V(BIOS:2.37V)
Ends up in Super PI 32M

http://img208.imageshack.us/img208/3...v1200piph3.png


I tried to found out how much of voltage I should add to NB when I want the
FSB is higher than 500.

Status I, when NB only has a fan-less solution:

http://img57.imageshack.us/img57/8766/rd600bhtd2.jpg


Super PI 32M (the OS of this test is moved from 975/ 650i with reinstallation, so, the PI result will be slower than the normal one)

http://img412.imageshack.us/img412/1...0v1200hpy8.png


Status II: If we put a air cooling to enhance heat reduction,

http://img412.imageshack.us/img412/6242/rd600blam3.jpg


Super PI 32M

http://img57.imageshack.us/img57/747...0v1200lmo3.png


NB has a fan: CPU VTT 1.44V/NB CORE 1.82V/NB PLL 2.31V/NB Core 1.89V
NB without a fan: CPU VTT 1.44V/NB CORE 1.76V/NB PLL 2.21V/NB Core 1.81V

I found that the heat of NB seems high is caused by the “temperature
monitoring” program.
When I put my finger on the top of NB’s heat sink, actually, it’s only “warmly”
not hot. I can say 975/965 is a bit hotter than RD600’s NB. So, the actual
temperature of it is 10 to 15C lower than what system shows.

BTW, the highest FSB under BIOS seems around 511Mhz,
If you use a AMD tweaking software under OS, it can push a little to 520Mhz,
furthermore, if you decrease the clockGen voltage inside the BIOS, then, you
might push FSB higher than 520Mhz.

The lines above is my personally experience of DFI RD600
Hope all these can help other enthusiast who has the same board.
Compare to other Intel base platform mobos, RD 600 is truly has many
advantages.
I hope the price of it can be a little bit lower, let us has lot more choices!!

some war booty I grab from PC mall recently

http://img208.imageshack.us/img208/3208/drd6007lk0.jpg

http://img443.imageshack.us/img443/8265/drd6008ki7.jpg

very nice

lol all those wharehouse goodies :D

Nice guide, I'll have to try some of this out and see how it works for me.

I really appreciate many members taking time to put out some good info on this board as it's a fun board for me with tons of potential.

Thanks for the work windwithme.

Not crapping this thread but AMD specifically told me running over 1.5V on the Nb will kill it in a matter of days thru electron migration.

I advise 1.5V max as does DFI on their forum, with jumps to 1.7 or so for benching but NOT 24/7.

I revised my OC guide on this advise and would advise caution running the NB voltages shown here.

Thanks for the guide. Very usefull, similar Settings I have, so I could try.

Tony, about running those high voltages on NB creating electromigration is right. But, are you talking with stock heatsink. Could this issue be fixed with a good Water Block? at least to say 1.7v stable?

Quote:

---

Originally Posted by Kain XS

Thanks for the guide. Very usefull, similar Settings I have, so I could try.

Tony, about running those high voltages on NB creating electromigration is right. But, are you talking with stock heatsink. Could this issue be fixed with a good Water Block? at least to say 1.7v stable?

---

I think tony mentioned DFI had run test samples with 1.6-1.7v with no damage however I dont know the duration of the test. AMD as the chip designer would know more and I would trust them. Simple fact is that people have killed their chip with 2.0 core voltage in around 7 hours I believe. Bingo13 can attest to that.

Electron migration is more seperate than close to heat I believe. The higher voltage draws more amps and that's what causes it I think. Heat is another source of damage as well. maybe if you phased it :p: it'd do better though.

you will experience deterioration of oc potential with prolonged exposure to high vnb, I have benched 3d mark 01 with the vnb at full at ddr1300, but the chipset does not hold out for long in terms of it's oc ability, this was with water cooling so the nb was sub 50 degs. For quick races you can bump (as benchers invariably will) but not for 24/7, I'm with Tony on this one, and both of us have the board, I actually use mine 24/7 rather than a quick drop of screenshots, funky crystals won't keep the nb alive..

gj i still cant get mine over about 590 on the memory

Quote:

---

Originally Posted by Kain XS

Thanks for the guide. Very usefull, similar Settings I have, so I could try.

Tony, about running those high voltages on NB creating electromigration is right. But, are you talking with stock heatsink. Could this issue be fixed with a good Water Block? at least to say 1.7v stable?

---

I dobut it. Personally, I would not run over 1.59v as FCG has had very good results with that.

The best solution to RD600 is to not buy rd600 unfortunatly :(

I like the board, but I wouldn't recommend for someone without a E6600-X6800 as a primary system with 24/7 use. With 1.5v on the NB being the max safe voltage, this puts it in the same category as the 975X boards for max 24/7 FSB overclocks. 1.5vnb will not get you even 450FSB in most cases. Combine that with the performance results it seems the only advantage of this board is the asynchronis memory clocks. I use the board for testing cpu's & memory now. Trying out the IN9 32 Max in hopes it will replace my P5W DH as my primary gaming box.

http://xtremesystems.org/forums/showthread.php?t=133706

Added to the list :) Thanks windwithme.

Quote:

---

Originally Posted by The Nemesis

I like the board, but I wouldn't recommend for someone without a E6600-X6800 as a primary system with 24/7 use. With 1.5v on the NB being the max safe voltage, this puts it in the same category as the 975X boards for max 24/7 FSB overclocks. 1.5vnb will not get you even 450FSB in most cases. Combine that with the performance results it seems the only advantage of this board is the asynchronis memory clocks. I use the board for testing cpu's & memory now. Trying out the IN9 32 Max in hopes it will replace my P5W DH as my primary gaming box.

---

Oh, make a thread about your results the IN9 ;)

I am going to try the QuadGT, as 1066 on 965 is fast and I do not need SLI :)

Quote:

---

Originally Posted by fhpchris

Oh, make a thread about your results the IN9 ;)

I am going to try the QuadGT, as 1066 on 965 is fast and I do not need SLI :)

---

I might, its become rather promising since this beta 6 bios:)
http://img250.imageshack.us/img250/2...5fsbjz6.th.jpg

gotta say I am disappointedwith the RD600 based on expectations....I cant run 1T without really boosting voltages...

way too expensive and I should have gotten the Badaxe 2....this board while I am 24hrs blend stable @3420 (9x380)...nothing special with 10th anniv crucial at DDR800 but at 2T.... I got nothing extra from this mobo except taken for extra $70....

Quote:

---

Originally Posted by nealh

gotta say I am disappointedwith the RD600 based on expectations....I cant run 1T without really boosting voltages...

way too expensive and I should have gotten the Badaxe 2....this board while I am 24hrs blend stable @3420 (9x380)...nothing special with 10th anniv crucial at DDR800 but at 2T.... I got nothing extra from this mobo except taken for extra $70....

---

Kinda sad, my RD600 cannot even pass 32m @ 430 fsb...

nd 13.4sec 1m @ 4005 mhz :(

iirc, electromigration is just an innate issue depending on the insulation material used, as well as just brute voltage. better cooling won't do much in the way of preventing it, unfortunately.

Also, I think when you summarized in the text after the pics, the voltages for fanned operation and fanless are opposite? I'm sorry if I was just stupid on this one. I need to get to bed; haven't slept in almost 40 hours

Quote:

---

Originally Posted by fhpchris

Kinda sad, my RD600 cannot even pass 32m @ 430 fsb...

nd 13.4sec 1m @ 4005 mhz :(

---

IMHO the problem with the mobo aside from performance vs 975 chipsets is the voltages needed to be boosted too much for extra speed at 1T and no one has really said what are going to be reasonable long term values

I would like to run 1T at my overclock but I need all my mobo voltages raised, I can run it but not sure if I will have a mobo degrade..I do not think they are extreme values..I keep my NB at 1.42v in bios....

Hearing 450 fsb is not going to be possible at 1.5v for everyone or less is absurd..they had all this time and I feel misled by performance...

If I thought I could sell for a reasonable price I would dump it in a heart beat...I should have listened to Travis aka AngryGanes..DFI layed an egg

I dont get me started on the Promise sata controller..it islows as a dog with just an sinlge sata drive..setup as eSATA....supposedly this is due to use small block transfer in the damn controller bios...so big files move slow..not an issue for small files..total BS!!!!

So I basically only have 4 sata ports and I want a sata ODD but here you do not want it on the same controller as your HDD