Printable View
well i think i stick with my DFI 975x board for now. To bad that both 680i and rd600 is not that "golden card" atm that i want. But i hope one of them will be better later on. 104c on n-b is to hot for me even it i will never run my cpu in 500+ fsb, i was hoping that rd600 was a cooler chipset than it turned out to be:hm:
The RD600 is cool & very efficient @ stock. The 104' was while overvolting using stock cooling. The same overvolt with some airflow was 70'c.Quote:
Originally Posted by Ten
yes i know 2.06v on n-b and 104c stock cooled but 70c default volt, and ~70c with delta fan on it (oc)so i think it's runs to hot for me anyway.
Quote:
Originally Posted by eva2000
thx eva, that confirms there should at least be a gap betweeen the 2, that should be maintained as you raise nb core. Setting these 2 too close will give the core regs a hard time.
regards
Raja
All I want from RD600 is a 400 fsb..maybe 450
I want a stable platform with as little vdroop as possible, cooler chipset and good memory utlization
I realize many want high fsb for extreme overclocking and hope this boards meets your needs
But I need a stable platform to try and reach 3400-3600 at as lower a vcore, chipst volt, mch voltage, vdimm as possible
BTW anyone use 10th anniver. crucial ram with this board???
From what I am seeing..the higher fsb more voltages increases that are needed..thus more heat and more issues with stablity...I wont ask if it worth it..becasue I know it is to everyone here...
Just not me anymore ..I want a good overclock, cool and stable..with as little noise as possible
Maybe that's what stopped Shamino from going higher http://www.xtremesystems.org/forums/...&postcount=876 ?Quote:
Originally Posted by raju
qdemn7, can't measure it right now as Tuniq 120 heatsink and 8pin psu connector get in the way for the ruler heh
rough guide on mosfet heatsink hole to hole would be 6-6.2cm
It is possible indeed, though Shamino usually knows best.
If it were linear regulators we are dealing with (which it is not), the nb core would fail to function with the same output voltage as input, as the regs would dropout.
edit: In fact thinking about it wht would happen is that the actual vnb would decrease a margin below what is set for the pre reg control, ie 2.06 in from step down voltage would not equal 2.06 actual out from core reg's ...
It is possible that the nbcore convert from is a reference voltage, even still having a gap is sensible, mostly a ratio between the 2 based upon the used circuitry (once again that'll be ascertained by trial and error that'll come from lots of hard work)...
It could be that the overall effect is not really noticeable depending upon how that control is configured in the bios code (actual vs set outputs). Some of these settings just make us happy. Lots to tinker with anyway, I feel it'll be a long time before people get truly every last mhz from this board..
regards
Raja
Thanks, I appreciate you trying.Quote:
Originally Posted by eva2000
Hey stilt, thanks for the answer, if possible, can you say if 450fsb 1:1 is better performance or would it be better async at 1200?Quote:
Originally Posted by The Stilt
I'm only asking is 450fsb is my personal sweet spot, so im trying to figure out which would be better...
as you said 420fsb is max in sync, so maybe 420 sync compared to 570async if possible?
Thanks again! :)
Quote:
Originally Posted by Ten
:confused: On load Rd600 is the coolest high end chipset for C2D. Temps don't say anything (the diode don't always tell the whole truth). You have to look at power consumption. And this chart says the Rd600 is 30W cooler than the 975X. http://www.anandtech.com/mb/showdoc.aspx?i=2891&p=6
I'm gonna put a Hr-05 on that NB anyway.
Good luck trying, read hereQuote:
Originally Posted by Poodle
I'm still going to fit my MCW30 on there somehow when I get it, if not just to prove him wrong. :D I think I can make it fit if I use the things used to keep the stock heatsink on the board with the waterblock, we'll see.
there's at least one chipset cooler I know of that should fit as it has adjustable hook mounts, lloking the the measurments that eva provided that seems fine. If you read that thread in full you'll see the replies.
regards
Raja
I'm 100% sure that it's possible, I have read the thread in full.
Can you take the hook mounts off the stock heatsink, for those of us with subpar cheapo chipset waterblocks that come with one set of mounts?
you sure it'll fit?Quote:
Originally Posted by Poodle
Quote:
Originally Posted by Thasp
Quote:
Originally Posted by marauder16
Quite. It comes with both the "normal" and Intel style mounting. How many more mounting styles can there be??? I will probably use the SLi version as my Ultra 120 will be take so much space and use some kind of radial blower to blow over the Hr-05 from the side (e.g. from the front of the case).
thnx, I've been considering it but wasn't sure if it'll fit, but now I'm buying it for certain :DQuote:
Originally Posted by Poodle
Quote:
Originally Posted by marauder16
Well I've read happy's post again. He ain't so sure... I'm used to "make things fit". So I'll go into the workshop if needed.
Well if you use stock hooks there should be no problem, if not I'l make it fit, of course ;)Quote:
Originally Posted by Poodle
Atleast Everest reports alot higher bandwidth and lower latency for sync mode ;)Quote:
Originally Posted by Revv23
Unfortunately I don´t have a screenshot for a 485FSB / DDR-1200 Async, but with those clocks Everest read score is just over 11k and latency 65ns or so.
In sync mode the board will work with ridicilously tight timings.
RDT at CL + 4 Clocks, tWR at 4 and tWTR at 2 is not a problem ;)
http://www.armada.fi/DFI_ICFX3200-T2..._ECMB_Sync.jpg
http://www.armada.fi/DFI_ICFX3200-T2...rPI1M_Sync.jpg
Here are the settings that I use for sync mode (ASM):
And the bios settings:PHP Code:Memory CAS Latency (RD_LAT): 4 Clocks
Memory Write Latency (WR_LAT): 3 Clocks
Activate to Read Delay (tRCDR): 4 clocks
Activate to Write Delay (tRCDW): 4 Clocks
Precharge Command Period (tRP) 3 Clocks
Internal Read to Precharge Command Delay (tRTP): 2 Clocks
Write Recovery Time (tWR): 4 Clocks
Activate Blank A to Active B Command Period (tRRD): 2 Clocks
Activate to Precharge Command (tRAS): 1 Clocks
Row Cycle Time Command Period (tRC): 7 Clocks
Auto-Refresh Command Period (tRFC): 20 Clocks
Read to Read Command to Different CS (tRTR_CS): 2 Clocks
Read to Write Bus Turnaround (tRTW): 3 Clocks
Internal Write to Read Command Delay (tWTR): 2 Clocks
Write to Read Command to Different CS (tWTR_CS): 2 Clocks
Write to Write Command to Different CS (tWTW_CS): 2 Clocks
CAS to CAS Command Delay (tCCD): 2 Clocks
CKE Minimum High and Low Pulse Width (tCKE): 3 Clocks
Exit Precharge Power Down to Any Valid Command (tXP): 8 Clocks
Exit Active Power Down to read Command (tXARDS): 6 Clocks
ODT Power Down Exit Latency (tAXPD): 8 Clocks
Precharge for All 8 Banks Memories (tRPALL): 1 Clocks
Bank to Bank Activate Rolling Window (tFAW): 2 Clocks
Mode Register Set Command Cycle Time (tMRD): 2 Clocks
On Die Termination (ODT): 75 ohm
Read Data Transfer (RDT): CL + 4 Clock
PHP Code:
Performance Options
Memory Clock Frequency: SYNC
Control CPU Frequency: Manual
CPU Frequency(MHz): 450
NB FSB Strapping: 333
PCIE Overclock: 100
Software Reset Clockgen: Disabled
Memory Overclocking: Disabled
-------------------
NorthBridge Selections
Memory Channels: Auto
Memory Power Management: Disabled
IGD Clock Gate: Disable
NB Channel Mapping: Manual
Simple Addr Translation: Disable
Chip-Select Swap: Disable
Burst Length: 8
Bank0 Map GFX: 4
Bank1 Map GFX: 5
Bank2 Map GFX: 6
Channel Mapping NB: 15
Channel Mapping GFX: 1
Bank0 Map NB: 5
Bank1 Map NB: 11
Bank2 Map NB: 7
--------
Memory Timing Menu
CS Output Timing: Auto
Dual Channel Split Mode: Auto
Addr/CMD Driving Pre-AMP: Enabled
Addr/CMD Driving Strenght: 06
Address/Command Timing: 2T
Memory Timing Select: Auto
Read Data Transfer: Enabled
Read Strobe Delay: Disable
Read Strobe Arrival: Enabled
--------
FSB IO Strenght
FSB IO Driving Control: Default
1X NMOS Strenght: 14h
1X PMOS Strenght: 15h
2X NMOS Strenght: 14h
2X PMOS Strenght: 15h
4X NMOS Strenght: 14h
4X PMOS Strenght: 15h
-------------------------
CPU VID Control: 1.36250V
CPU VID Special Add: Auto
DDR2 RAM 1.8V Voltage: 2.41V
CPU VTT 1.2V Voltage: 1.44V
NB Core 1.2V Voltage: 1.72V
NB PLL 1.8V Voltage: 2.31V
NB PLL 1.2V Voltage: 1.38V
NB PCI-E 1.2V Voltage: 1.34V
NB Core 1.2V Convert From: 2.31V
SB Core 1.2V Voltage: 1.34V
Clockgen Voltage: 2.92V
GTL REF Voltage Control: Disabled
CPU Core 1/2 GTL REF Voltage: 48
CPU Core 3/4 GTL REF Voltage: 48
North Bridge GTL REF Voltage: 128
CPU PLL Setting Override: Enabled
Conroe P-State Multiplier: x8
C1E Function: Disabled
MCW30 fits perfectly on the board with the standard two hook sets.Quote:
Originally Posted by Thasp
I´m using it with Tuniq Tower 120.
Any chance for a nice pic?Quote:
Originally Posted by The Stilt
Quote:
Originally Posted by The Stilt
Wow, Thanks again! :)
Have you tried for 1T max in sync? The 1:1 performance is just awesome.
:toast:
The VR-Zone review really sucked IMO. Why did they use the infinity and turn it so the hot air blows down on the chispet heatsink? Of course your gonna wind up with a high temperature on the NB. If I do the same thing on my P5B DLX it wouldn't even boot over 420 MHz FSB and the temp would be 10C higher then if I pointed the infinity in another direction. Not to mention that man GFX card would get warmer to.
There is a Swedish expression "Gör om, gör rätt" = Do it again, do it right.
Don't think it's that bad.. i have the same fan direction with Tuniq Tower 120 on my DFI ICFX3200 as wellQuote:
Originally Posted by Gambit_2K
http://i4memory.com/reviewimages/mot..._RD600_169.JPG
Only diff is i don't shove 2.3+ volts on the NB in bios :)
I added a sunon 60x25mm 23.5cfm fan over the stock passive heatsink and dropped ~10C in temps under load.. i've only tried as high as 1.76v for NB bios read 50C NB temps
Quote:
Originally Posted by eva2000
Hmm okey but can you confirm that the NB temp is 70 C @ stock? Maybe not through bios but through touch? And over 100C @ overclocking? <-- Maybe not by touch ;). When I run 470 FSb on my PB5 DLX I can still touch my heatsink without burning myself. Suddenly a snakesnest of heatpipes around the board doesn't seem like all that bad of an idea.