X48 Rampage Formula Preview.

[papatsonis]

disagree with the last part for the 8gb... i run 4x2gb gskill pi 8800 (8gb kit) like "for ages" , either 450fsb-1200 pl6, or 467(right now)-1167-pl7-1.35vMCH-16120ref totally STABLE (but of course i have troubles to go beyond ~470 in terms of actual stability, and without so much time to fine tune and research.. im happy at 467.). 8gb problems might be likely not optimal socket for each module.. you have 24 possible placements for 4 modules.. and you need to narrow down the ONE producing the least errors in memtest, while testing in the "thin line of line" of stability/instability to be able to see if the errors increase or disappear

[Alien Grey]

disagree with the last part for the 8gb... i run 4x2gb gskill pi 8800 (8gb kit) like "for ages" , either 450fsb-1200 pl6, or 467(right now)-1167-pl7-1.35vMCH-16120ref totally STABLE (but of course i have troubles to go beyond ~470 in terms of actual stability, and without so much time to fine tune and research.. im happy at 467.). 8gb problems might be likely not optimal socket for each module.. you have 24 possible placements for 4 modules.. and you need to narrow down the ONE producing the least errors in memtest, while testing in the "thin line of line" of stability/instability to be able to see if the errors increase or disappear

I didn't say that it's impossible. What I'm trying to say is that with changing the ODT timings it's going to be a lot easier to find true stability. And I mean true stability because I'm tired of trying to tweak the board to have it stable running to find out later that it isn't stable anymore. I've done the research and I've seen the difference and I know for sure that the ODT's are causing these instabilities.

[leo27]

Could you please tell me why you need such options in BIOS？

To have a stable running board for once and for all.

Is there something wrong with your settings for the PC at present?

Running 4 X 2GB at DDR +1066MHz is almost impossible. A few degrees higher temperature and the board starts randomly failing. A MCH Read Delay of 8T is to tight for Quads in almost every situation.

And most important of all Intel only suggested to use these timings as default but I don't think that they've said that you can't relax them to improve stability.

That's pretty much my reply A-Grey

[greg.m]

ASUS answer:

Dear Valued ASUS Customer!

We cannot provide any support for overclocking, it is done at your own risk and without support.

Best Regards / Med Vänliga Hälsningar,
xxxx
ASUS TMSS Support



And here is the key phrase:

"I need the clock skews to find a stable overclock with my Q9650"

I think that i got this answer from asus nordic and this is why it is so deterrent

[Alien Grey]

I'm not surprised that they gave you such an answer. They don't have a clue what you're talking about. I think that we're smart enough to know that overclocking/tweaking is at our own risk but to do that the least I expect is that the BIOS has all the features available.

This makes me wonder if anyone at ASUS ever tried it because you don't have to be an engineer to see the difference in stability with relaxing the ODT. With doing that the thin line of stability suddenly becomes a highway.

[Scougar]

Been lurking on this thread for some time now.

Got to say this guys, this is a real PR failure for Asus support. Simply makes Asus look bad in comparison to such companies as Gigabyte :-(

Anyway, I have this motherboard, and just got a Q6600 (Known to do 3.6Ghz@1.3v [9*400 VID 1.215v])... so will be trawling the thread for suggestions (170 pages is a lot of reading so I'll be a while lol) :-) Unsure of my BIOS at present, could someone point me to the latest MOD BIOS to try out please?

Keep up the fight guy's, I really don't understand what can be that hard about adding a few options to a BIOS for Asus other and being stubborn about who foots the bill to pay the guy(s) who do it (I'm a software developer btw, so understand the issues faced by programmers :-) ) Having a community presence from Asus would definately be welcomed!

EDIT: If SOLAR can edit the BIOS, surely it isn't game over? Even if it's seperate BIOS's that have different hard set ODT values... it isn't quite 'game over' like a few pages back, just means flashing a different BIOS to try and see what value you need for higher clocks?

EDIT2: SOLAR's link no longer seems to work for the 1002 BIOS :-( Does anyone have a copy or a link please? (or do I need 1002B mod bios for my Q6600 and rampage formula? (non-extreme).

[Alien Grey]

Been lurking on this thread for some time now.

Got to say this guys, this is a real PR failure for Asus support. Simply makes Asus look bad in comparison to such companies as Gigabyte :-(

Anyway, I have this motherboard, and just got a Q6600 (Known to do 3.6Ghz@1.3v [9*400 VID 1.215v])... so will be trawling the thread for suggestions (170 pages is a lot of reading so I'll be a while lol) :-) Unsure of my BIOS at present, could someone point me to the latest MOD BIOS to try out please?

Keep up the fight guy's, I really don't understand what can be that hard about adding a few options to a BIOS for Asus other and being stubborn about who foots the bill to pay the guy(s) who do it (I'm a software developer btw, so understand the issues faced by programmers :-) ) Having a community presence from Asus would definately be welcomed!

EDIT: If SOLAR can edit the BIOS, surely it isn't game over? Even if it's seperate BIOS's that have different hard set ODT values... it isn't quite 'game over' like a few pages back, just means flashing a different BIOS to try and see what value you need for higher clocks?

EDIT2: SOLAR's link no longer seems to work for the 1002 BIOS :-( Does anyone have a copy or a link please? (or do I need 1002B mod bios for my Q6600 and rampage formula? (non-extreme).

It's still "Game Over" but I didn't say it would be for us.

Look at page 169. SoLoR moved his stuff to an other download place.

[Scougar]

It's still "Game Over" but I didn't say it would be for us.

Look at page 169. SoLoR moved his stuff to an other download place.

That's embarassing.. I actually went 'back' through the pages and eventually settled on searching for 'download' so got the old link. Can't believe I missed a link only one page back d'oh. Thanks

Honestly, think you've inspired people, your commitment for the board is awesome :-) I wish the same could be said for Asus. I'll post up with my results as I try them. (I've only managed a stable 355FSB so far on my old e4300, so hoping the board isn't stopped at this FSB).

Update: (Currently BIOS 0902) First few mins, managed 3.6Ghz at 1.33/1.35 volts in windows (1.4v asked for in BIOS, although monitoring in bios shows 1.33v). Northbridge temp reached 60C whilst in a game however. NB voltage at 1.36, maybe too much? At idle it's 48C at present and dropping as just been gaming. (bear in mind 3600Mhz 400 FSB). Think my northbridge might not be hooked up properly? CPU is idling at 29/30C, SB at 43C, MB @ 33C.

Think I am gonna have to take my motherboard out 'again', and try and pop the northbridge and see what the deal is, so I can get the temperatures down.

EDIT: Also noticed I can select my Q6600 as a QX chip in the BIOS as well.. any advantage to this?

Update: Tried prime 95 64 bit and core 4 fails the test on small tft or whatever its called within a few seconds. Will try and get it stable later today. Also put a 50mm fan on the northbridge to help temporarily, but will have to make the permanent fix soon.

Update: Tried a little more vcore, so I get 1.37v in windows, but the droop seems to be quite significant. About 0.06v droop down to 1.30-1.32v. Is this bad? I'm not using load line calibration or setting any termination settings yet. Core 4 lasted about 1 minute this time before failing. I had also reducd the northbridge voltage by 0.02v to help temps, and apart from the already failing prime 95 test, it seemed better. CPU was getting about 55C on the short test.

Update: Couldn't get stable at 3.6Ghz (400 * 9) so reducing to 3.4Ghz and so far after 23 minutes it's passing prime 95 just fine with small FFT's. Mini update: 35 mins stable so far on prime 95 :-) MUCH better. Core 4 is also way ahead of the others strangely... (Well one test ahead). According to Coretemp, temps were 73/73/69/69. Might have to reseat cooler. (and use less paste lol).

[leo27]

It looks like I've finally fixed my stability issue, it was caused by three things:

1. A bad flash to BIOS 1002 which corrupted the CMOS.
2. Prime95 26.4 eventually fails at FATAL ERROR: Rounding was 0.5, expected less than 0.4. Version 25.11 ran for 8 hours so i recommend everyone to not use the 26.x builds yet.
3. I found that for some reason vDRAM drops a little under load from 1.872 to 1.85 this is why the memory was unstable at tight timings but is of course perfectly stable at default timings and frequency.
4. Using Bar_EDIT to change the MCH ODT Timing to 11 from 8 increased the stability of the system and is now P95, LinX and HCI Memtest stable.

[neo_rtr]

Was getting some issues with SC2 and after some time i though to try some other settings

Ai Overclock Tuner [Manual]
CPU Ratio Setting [9.0]
FSB Strap to North Bridge auto
FSB Frequency [440MHz]
PCIE Frequency [105MHz]
DRAM Frequency [1103MHz]
DRAM Command Rate [2N]
DRAM CLK Skew on Channel A [Normal]
DRAM CLK Skew on Channel B [Normal]
DRAM TimingControl [Manual]
Refresh Period auto
DRAM Static Read Control [Disabled]
Ai Clock Twister [Lighter]
Ai Transaction Booster [Manual]
Common Performance Level [07]
Pull-In of CH A/B all disabled
CPU Voltage [1.3625V]
CPU PLL Voltage [1.52V]
North Bridge Voltage [1.59V]
DRAM Voltage [2.02V]
FSB Termination Voltage [1.39V]
South Bridge Voltage [1.05V]
SB 1.5V Voltage [1.50V]
Loadline Calibration [Enabled]
CPU GTL Voltage Reference [0.65X]
NB GTL Voltage Reference [0.67X]
DRAM Controller Voltage REF [Auto]
DRAM Channel A/B Voltage REF [Auto]
CPU Spread Spectrum [Disabled]
PCIE Spread Spectrum [Disabled]
CPU Clock Skew [Delay 200ps]
NB Clock Skew [Delay 100ps]

Was running Furmark full screen 1920x 1080 AND at the same time Prime full run smal and later on large FFTs and did NOT crash!!
Memory was 5-5-5-15-auto-55 perflevel 7

.EDIT: DID NOT CRASH

[greg.m]

Was getting some issues with SC2 and after some time i though to try some other settings

Ai Overclock Tuner [Manual]
CPU Ratio Setting [9.0]
FSB Strap to North Bridge auto
FSB Frequency [440MHz]
PCIE Frequency [105MHz]
DRAM Frequency [1103MHz]
DRAM Command Rate [2N]
DRAM CLK Skew on Channel A [Normal]
DRAM CLK Skew on Channel B [Normal]
DRAM TimingControl [Manual]
Refresh Period auto
DRAM Static Read Control [Disabled]
Ai Clock Twister [Lighter]
Ai Transaction Booster [Manual]
Common Performance Level [07]
Pull-In of CH A/B all disabled
CPU Voltage [1.3625V]
CPU PLL Voltage [1.52V]
North Bridge Voltage [1.59V]
DRAM Voltage [2.02V]
FSB Termination Voltage [1.39V]
South Bridge Voltage [1.05V]
SB 1.5V Voltage [1.50V]
Loadline Calibration [Disabled]
CPU GTL Voltage Reference [0.65X]
NB GTL Voltage Reference [0.67X]
DRAM Controller Voltage REF [Auto]
DRAM Channel A/B Voltage REF [Auto]
CPU Spread Spectrum [Disabled]
PCIE Spread Spectrum [Disabled]
CPU Clock Skew [Delay 200ps]
NB Clock Skew [Delay 100ps]

Was running Furmark full screen 1920x 1080 AND at the same time Prime full run smal and later on large FFTs and did crash!!
Memory was 5-5-5-15-auto-55 perflevel 7

try with setting CPU GTL Voltage Reference @ 0.63X or even @ 0.62X

[neo_rtr]

SC2 issues comeback. Game hangs for no reason. Will pass Prime95 tests. I just dont get this any more. One day ok _Next day not ok

[Alien Grey]

SC2 issues comeback. Game hangs for no reason. Will pass Prime95 tests. I just dont get this any more. One day ok _Next day not ok

On Die Termination and sub timings. Sub timings can be fixed. ODT is more tricky and can only be changed in Windows with BAR_Edit so it's difficult to find the right timings.


My request to ASUS is to add MCH Read ODT from 8T to 16T, DRAM Read ODT from 3T to 6T and MCH Write ODT from 6T to 12T with increments of 1T so we can find the ODT timing necessary to have a true stable running board. A fixed ODT timing doesn't seem to work stable.

I'm tired of these instabilities and I think that everyone else also but ASUS doesn't seem to care to much about it.

[neo_rtr]

I have start looking at Gigabyte.
Iam not saying anything, just saying.. Been having ASUS for as long as i can remember.

[Scougar]

I put a post on the Asus ROG forum, but I guess you guy's have already tried that. I nicked what you said Alien Grey and shoved it in there of what we need :-)

Anyway, Last night I took my PC apart again, and I forgot to get some needle nose pliers so couldn't get the push pins out, so had to resort to just tightening the screws on the northbridge. They were VERY loose. I think Northbridge temps have dropped by about 8 degrees C since I've done that. I'm not even running a fan on the northbridge either, so that's a definate improvement. (Now at 52C when running prime 95 small ftt's). I will put the fan on again to reduce them further until I figure out what's going on with my high CPU temps, and then try and get the heatpipe off :-)

Those pushpins are shocking though. They should at least have used a spring with more force. May end up doing a bolt through mod if I can find appropriate parts.

[Alien Grey]

It is what it is. We don't have to ask for better GTL handling or anything else voltage wise. Maybe technically it isn't necessary to have these ODT timings but it's for sure the solution for all the instabilities with this board and if the ASUS engineers would take the time to do a little research and don't try to ignore the problems they would see that too.

Our stability problems with the board won't go away with trying to ignore it and all we've done the past years is corrupting data.

[neo_rtr]

Update:I think my board likes highe volts.
As i said: Was getting some issues with SC2 and after some time i though to try some other settings
I could run full screen Furmark on multu GPUs and Prime at the same time with no issues. On the other hand was getting random hangs,BSOD in game-


Ai Overclock Tuner [Manual]
CPU Ratio Setting [9.0]
FSB Strap to North Bridge auto
FSB Frequency [440MHz]
PCIE Frequency [105MHz]
DRAM Frequency [1103MHz]
DRAM Command Rate [2N]
DRAM CLK Skew on Channel A [Normal]
DRAM CLK Skew on Channel B [Normal]
DRAM TimingControl [Manual]
Refresh Period auto
DRAM Static Read Control [Disabled]
Ai Clock Twister [Lighter]
Ai Transaction Booster [Manual]
Common Performance Level [07]
Pull-In of CH A/B all disabled
CPU Voltage [1.3625V]
CPU PLL Voltage [1.52V]
North Bridge Voltage [1.61V]
DRAM Voltage [2.02V]
FSB Termination Voltage [1.39V]
South Bridge Voltage [1.05V]
SB 1.5V Voltage [1.50V]
Loadline Calibration [Enabled]
CPU GTL Voltage Reference [0.63X]
NB GTL Voltage Reference [0.67X]
DRAM Controller Voltage REF [Auto]
DRAM Channel A/B Voltage REF [Auto]
CPU Spread Spectrum [Disabled]
PCIE Spread Spectrum [Disabled]
CPU Clock Skew [Delay 200ps]
NB Clock Skew [Delay 100ps]
Memory was 5-5-5-15-auto-55 perflevel 7

I was testing yday and only after i raised the NB to 1.61V found it to be stable in game. As i mentioned before sometimes it works and some other times it does not. So Will see more about it over the weekend.

[neo_rtr]

Nope.. So this is the thing. It was running fine for several hours of intense game play, testing and all that good jazz. The following day it crashed with a simple copy past. It could not pass any prime test, as it failed immediately after loading prime.
So back top 435.

[Spyrus]

I see in Memset and CPUInfo utilities, tRTP to be reported = 5, but i have 3 in the bios.
I need this in order to have memset to change tREF to 16383T instead of 3120T for bios 1001.
So, if it reports tRTP =5 it is actually 5 or should i force it to be 3 when changing values within memset?

[Alien Grey]

I was waiting for your test results neo_rtr. It's exactly as what I thought it would be and what I've been saying for so long. You can't fix it with higher voltage because it isn't a voltage problem.

I can't believe that you would need that high NB Voltage for FSB 440 MHz. If this is really necessary than you better look for a new board.

We need those ODT timings in the BIOS. It's the only way to have a stable running board. Giving it more voltage to try to make it stable gives us the impression that it's better but it isn't. It's making it worse because you think it's stable but if you test it again you notice that it's still going to fail. It's just a matter of time because you never know when and how many times it's going to fail because of the to tight ODT.

I've been running fine for some time at 9 X 469MHz and DDR 1250MHz and it can be really stable if I would relax the ODT with BAR_Edit but I don't want to do that. I think that it's ASUS responsibility to fix this for us and add those timings in the BIOS so that we can have a stable running board.

[Alien Grey]

I see in Memset and CPUInfo utilities, tRTP to be reported = 5, but i have 3 in the bios.
I need this in order to have memset to change tREF to 16383T instead of 3120T for bios 1001.
So, if it reports tRTP =5 it is actually 5 or should i force it to be 3 when changing values within memset?

Why would you want to do that? Why don't you use BIOS 0001 or the modded one from SoLoR?

Both these BIOS revisions have the tREF setting available.

[Spyrus]

Yeah i could use the modded bios. But can you say to me if with modded bios the reported tRTP value from Memset/CPUInfo is 5 instead of 3?

[Alien Grey]

Yeah i could use the modded bios. But can you say to me if with modded bios the reported tRTP value from Memset/CPUInfo is 5 instead of 3?

tRTP is 5 with MOD BIOS 1002.

BIOS First Information says tRTP is 3T with timing on Auto but it's actually 5T if you look with MemSet.

[Spyrus]

Yeah, i put on SoLoR's bios and it is still 5 in Windows although 3 in bios.
Also i put 16120 tREF but it is not the same with 16383T i guess.

[Alien Grey]

Yeah, i put on SoLoR's bios and it is still 5 in Windows although 3 in bios.
Also i put 16120 tREF but it is not the same with 16383T i guess.

It isn't a BIOS bug. If you want tRTP to be 3T in MemSet than you've got to change tRTPD in the BIOS to 3T. You can find tRTPD in the Third Information.

There isn't any difference in bandwidth between MemSet tREF 16383T and BIOS tREF 16120T and it's perfectly stable with 16120T.