X48 Rampage Formula Preview.

[Alien Grey]

well try to do this manualy: http://rapidshare.com/files/209947592/BAR_Save.zip here is my C:\BAR_Save (its default location BAR_Edit creates to load at startup). Then make shortcut to startup folder to BAR_Edit.exe from BAR_Save. In there you will find BAR_Edit.ini and just change addresses & values you want to change.

I think that I've found the problem. I didn't install Bar_Edit to the default location and it puts BAR_Save in the Bar_ Edit folder and not in the ROOT.

I moved BAR_Save to the ROOT and now it loads the settings at startup.

Thanks for your help.

[Fritz_the_germ]

Hallo Alien Grey!
My question about a Dram voltage problem is answered in another thread here in XS forum.
But by the way, I found out that the lowest voltage I can set in the Rampage bios is 1,80v (=1,89v real). I think that´s another issue to the Asus bios programmers to make a lower dram voltage possible, ne c´est pas?

So far I´m having a little trouble with memsetV4. The programm caused freezes to my system -by my own fault.
When I put in my new sticks and changed bios settings I forgot to change to memset4 settings and it caused freezes by calling up windows. The bios setting were PL 6 and the memset4 settings were PL 7. That might have been the fault.
Yesterday I was loading down memset4 again and installed it. But every time I call up windows memset4 has to get activaded manually. Before, I just applied it and it worked by itself without doing anything after starting windows.
What is wrong?

[Alien Grey]

Hallo Alien Grey!
My question about a Dram voltage problem is answered in another thread here in XS forum.
But by the way, I found out that the lowest voltage I can set in the Rampage bios is 1,80v (=1,89v real). I think that´s another issue to the Asus bios programmers to make a lower dram voltage possible, ne c´est pas?

I posted an answer in your G.Skill thread but I post it here to.
My Rampage Formula doesn't over volt that much. It over volts it between 0.00V and 0.02V.
The NB is under volt with 0.02V. FSB Termination Voltage is under volt with 0.1V.

I'm going to get my G.Skill F2-9600 memory back from RMA tomorrow.

So far I´m having a little trouble with memsetV4. The programm caused freezes to my system -by my own fault.
When I put in my new sticks and changed bios settings I forgot to change to memset4 settings and it caused freezes by calling up windows. The bios setting were PL 6 and the memset4 settings were PL 7. That might have been the fault.
Yesterday I was loading down memset4 again and installed it. But every time I call up windows memset4 has to get activaded manually. Before, I just applied it and it worked by itself without doing anything after starting windows.
What is wrong?

I have no idea. Did you delete your previous saved settings?

[Fritz_the_germ]

Yes! I deleted the whole folder with everything in it. It´s the same case when I hit the xp partition. I have to apply the memset settings manually.
Forgot to mention that the "PL" and "all precharge to activate" are greyed out (?). That means those settings aren´t active.

When you will get your G.Skills back tomorrow, do you prefer a 1:1 divider or do you set 4:3? What is going to bring out the best performance?

[Alien Grey]

Yes! I deleted the whole folder with everything in it. It´s the same case when I hit the xp partition. I have to apply the memset settings manually.
Forgot to mention that the "PL" and "all precharge to activate" are greyed out (?). That means those settings aren´t active.

The reason I'm asking this for MemSet is that if you didn't delete the saved settings they might still be stored somewhere and screw up your new saved settings. I don't know if that is the problem but it's possible.

When you will get your G.Skills back tomorrow, do you prefer a 1:1 divider or do you set 4:3? What is going to bring out the best performance?

Why would you want to use 1:1 with 1200MHz memory?

There is only one reason to use 1:1 and that is with FSB 600MHz.
You could change the CAS to 4T and run 1:1 but I don't see the point of doing this with this memory. With FSB 450MHz and DRAM 900MHz you can't use tRD 6.

[Fritz_the_germ]

I asked you for the fsb divider because some people say it´s senseless to run a higher memory bus by using an Intel chipset. The Intel system wouldn´t let more datas into the "working zone" than the frequency of the fsb is. Only the maximum fsb should be the right approach, so they say.
I really don´t know what those poeple exactly mean by saying that. But I can see that the "read" performance is higher when the memory bus is faster than the fbs.
For my part would it be foolish to buy two expensive memory sticks for running them with a 1:1 divider
The best pick to run a high fsb would be a fsb up to 500 MHz, memory divider 6:5, tCL 5, TRD=14.0 ns results into a tRD=7.
Because [tRD-(tCL/N)]>[X/N]; X> 466 fsb MHz=2.75
Given fsb=500 MHz, N=6:5, tCL= 5. Solve: [7-(5/{6/5})]>[2.75(6/5)]=2.83>2.29
Answere: 2.83 is greater than 2.29
The reason why I will run a 4:3 divider (fsb 450 MHz/ 600 MHz memory) is because I have a pretty good performance by doing so and I won´t stress my system with a high voltage.

[RCG_Bex]

Lovin this BIOS - just gave my CPU a good kick!

440*10 - not stable, prime "stops responding" after 1st loop, seems my rig just needs alot of NB.....

However 1). it needs alot more V's than I remember and 2). Not sure what to change to make it stable.... Going to go fiddle some more

~Bex

[mikeyakame]

Zucker,

Do these L2 Read/Write/Copy/Latency values look alright to you? These are on a Q9550 E0 i grabbed on the weekend. PL8 @ 415 x 8.5 on my DFI X48 board.

I'm a bit uncertain of the Write and Latency mostly. I don't know whether the latency should be that bad as I've never noticed it being that low on a 45nm chip, but maybe it was for all i know!

Anyway just let me know what you think when you get a chance to have a look. If in your experience they look alright I won't worry about it, want to determine whether its because of larger 6mb cache and heavier cost to access, or some kind of cache coherency issue my chip has!

[Brama]

I have a Xeon E3110 (twin of E8400) that has very similar results on L2 cache.

[RCG_Bex]

Squeezing everything I can out this board - Currently at 455FSB, going to try more tonight.





Heh, and to think they say the Maximus is limited to ~440 max

~Bex

[pby5cat]

The information from Zucker2k, mikeyakame, A_Grey, and others has been a great help in reducing the blue screen occurrences on my friends system at 4Ghz. Yet the blue screens refuse to go away, I need some input.

Goal: 4GHz using the max multiplier of 8.5 and run ram at 1:1.
system specs:
Bios listed below

Xeon X3360 (Q9550) C1 water cooled with a stock fuzion v2 along with EK northbridge wb, laing D5 @ speed 5, and PA120.3 in one loop. Prime temps are 58C, LinX are 67C, northbridge temp is about 30C.

second loop is EK mosfet wb, 2 EK FC-R600 HD 2900XT crossfire, laing D5 @ speed 5, and PA120.3. The south bridge is heatsinked with a low profile copper sink, it stays in the high 30C.

Video cards are a mismatched crossfire of HD2900XT 1GB and HD2900XT 512mb, both water cooled. Blue screens happen with or without crossfire on.

Ram is 2X2GB GSkill F2-8800CL5GBPI (DDR2 1100) in the blue slots, Vista 64bit Ultimate SP1 all patched up. Jmicron controller is disabled, all sata AHCI installed. Power supply is a Silverstone OP-1000 1Kw. The motherboard is a X48 Rampage Formula.

So the problem is windows blue screens saying "A clock interrupt was not received on a secondary processor within an allocated time" or "General hardware fault". This has been going on for months, and thanks to this thred, it usually happens once or twice a day now. The information about AGTL+ bus signaling, GTL ref, bios 803, cpu clock skew, and Zuckers BIOS, decreased blue screens alot.

Memtest can run all day without errors, and processor has primed for 6 hours no errors, no blue screens. The blue screens very rarely occure under heavy load from gaming or priming, only once that I can remember. They occur during idle desktop, internet browsing, and word documents. Never had any issues during windows load screen.

The CMOS was cleared befor and after flashing the bios, and profiles where re-entered and saved by hand.

Load line calibration is enabled, and tested fine. CPUZ and Everest show no Vcore fluctuations between idle and load. The Vcore stays at 1.360v all the time no matter the load. The board was vdroop pencil moded and sealed by tape during inital building.

NB voltage has been up to 1.73v, SB 1.5v to 1.55v, PL 10 and 9, ai clock twister, all GTL's that fit have been tested, with no improvement and returned to as stated below.

Anyone have any input on settings or experience with thies blue screens?
If the GTL ref are kept within 0.8 to 1.0, is the 1.4 volt max safe FSB termination voltage still valid for long term daily use?

Extreme Tweaker
Ai Overclock Tuner : [Manual]
CPU Ratio Setting : [8.5]
FSB Strap to North Bridge : [333]
FSB Frequency : [471]
PCI-E Frequency: [101]
DRAM Frequency: [DDR2-943]
DRAM Command Rate : [2N]
Dram CLK Skew on Channel A : [AUTO]
Dram CLK Skew on Channel B : [AUTO]
DRAM Timing Control: Manual
CAS# Latency : 5
RAS# to CAS# Delay : 5
RAS# Precharge : 5
RAS# ActivateTime : 15
RAS# to RAS# Delay : Auto 3
Row Refresh Cycle Time : auto (52)
Write Recovery Time : Auto 6
Read to Precharge Time : Auto 3

Read to Write Delay (S/D) : Auto 8
Write to Read Delay (S) : Auto 3
Write to Read Delay (D) : Auto 5
Read to Read Delay (S) : Auto 4
Read to Read Delay (D) : Auto 6
Write to Write Delay (S) : Auto 4
Write to Write Delay (D) : Auto 6
DRAM Static Read Control : [Enabled]
Ai Clock Twister : [Lighter]
Ai Transaction Booster : [Manual]
Common Performance Level : [08]

All Pull-ins = [Disabled]

CPU Voltage : [1.37500v] 1.360v real
CPU PLL Voltage : [1.60v] 1.648v real
North Bridge Voltage : [1.65v] 1.64v real
DRAM Voltage : [1.90v] 1.904v real
FSB Termination Voltage : [1.52v] 1.424v real
South Bridge Voltage : [1.100v] 1.104v real
SB 1.5v Voltage : [1.50v] 1.504v real
Loadline Calibration : [Enabled]
CPU GTL Reference : [0.62x]
North Bridge GTL Reference : [0.67x]
DDR2 Channel A REF Voltage : [AUTO]
DDR2 Channel B REF Voltage : [AUTO]
DDR2 Controller REF Voltage : [AUTO]
NB LED Selection : [NB Volt]
SB LED Selection : [SB Volt]
CPU LED Selection : [CPU Volt]
Voltiminder LED : [Enabled]
CPU Spread Spectrum : Disabled
PCIE Spread Spectrum : Disabled
CPU Clock Skew : [Delay 200ps]
NB Clock Skew : [Normal]

Advanced -> CPU Configuration

CPU Ratio Setting : [8.5]
C1E Suppport : Disabled
CPU TM Function : Disabled
Vanderpool Technology : Disabled
Execute Disable Bit : Disabled
Max CPUID Value Limit : Disabled

[mikeyakame]

I have a Xeon E3110 (twin of E8400) that has very similar results on L2 cache.

Ahh lovely. It's weird that the 65nm chips which are older and slower perform better on L2 cache. Guess thats just how it is then. I've found on this 45nm cpu with Everest bandwidth test, everything has to be idle or the L2 cache values become screwed up!

At least now I've managed to get the L2 bandwidths and latency fairly consistent boot to boot, took some fine CPU0/3 GTL adjustments to get it there. After playing with this new DFI X48 board for the past month I've realized how critical having fine GTL adjustment actually is. Though I do admit having 0.005v Vtt adjustment and GTL1/2 GTL0/3 GTLMCH being mapped out from values of 0-160 against set real Vtt values to obtain anywhere from 0.53x to 0.82x or so makes fine tuning GTLs a long slow process, and only changing one GTL at a time by +-1 which of 1 = ~0.007 - 0.009v. Worst part is doing extrapolation calculations. Vtt GTL tables are at 1.06v, 1.21v, 1.31v, 1.41v, 1.51v, 1.55v. So anywhere in between those Vtt real voltages to work out the value of GTL for ratio you want it requires a scientific calculator and printouts of the tables handy at all times!

The Rampage in comparison is so coarse and brute force, if GTL values don't work raise/drop Vtt 2 or 3 steps and pray that you can get the right combination! But it works much better until you reach higher FSB when it cripples potential!

Thanks mate for replying

[ChaosMinionX]

Zucker,

Do these L2 Read/Write/Copy/Latency values look alright to you? These are on a Q9550 E0 i grabbed on the weekend. PL8 @ 415 x 8.5 on my DFI X48 board.

I'm a bit uncertain of the Write and Latency mostly. I don't know whether the latency should be that bad as I've never noticed it being that low on a 45nm chip, but maybe it was for all i know!

Anyway just let me know what you think when you get a chance to have a look. If in your experience they look alright I won't worry about it, want to determine whether its because of larger 6mb cache and heavier cost to access, or some kind of cache coherency issue my chip has!

Your L2 looks similar to mine.... my Q6600's L2 ran way faster.

The information from Zucker2k, mikeyakame, A_Grey, and others has been a great help in reducing the blue screen occurrences on my friends system at 4Ghz. Yet the blue screens refuse to go away, I need some input.

Goal: 4GHz using the max multiplier of 8.5 and run ram at 1:1.
system specs:
Bios listed below

Xeon X3360 (Q9550) C1 water cooled with a stock fuzion v2 along with EK northbridge wb, laing D5 @ speed 5, and PA120.3 in one loop. Prime temps are 58C, LinX are 67C, northbridge temp is about 30C.

second loop is EK mosfet wb, 2 EK FC-R600 HD 2900XT crossfire, laing D5 @ speed 5, and PA120.3. The south bridge is heatsinked with a low profile copper sink, it stays in the high 30C.

Video cards are a mismatched crossfire of HD2900XT 1GB and HD2900XT 512mb, both water cooled. Blue screens happen with or without crossfire on.

Ram is 2X2GB GSkill F2-8800CL5GBPI (DDR2 1100) in the blue slots, Vista 64bit Ultimate SP1 all patched up. Jmicron controller is disabled, all sata AHCI installed. Power supply is a Silverstone OP-1000 1Kw. The motherboard is a X48 Rampage Formula.

So the problem is windows blue screens saying "A clock interrupt was not received on a secondary processor within an allocated time" or "General hardware fault". This has been going on for months, and thanks to this thred, it usually happens once or twice a day now. The information about AGTL+ bus signaling, GTL ref, bios 803, cpu clock skew, and Zuckers BIOS, decreased blue screens alot.

Memtest can run all day without errors, and processor has primed for 6 hours no errors, no blue screens. The blue screens very rarely occure under heavy load from gaming or priming, only once that I can remember. They occur during idle desktop, internet browsing, and word documents. Never had any issues during windows load screen.

The CMOS was cleared befor and after flashing the bios, and profiles where re-entered and saved by hand.

Load line calibration is enabled, and tested fine. CPUZ and Everest show no Vcore fluctuations between idle and load. The Vcore stays at 1.360v all the time no matter the load. The board was vdroop pencil moded and sealed by tape during inital building.

NB voltage has been up to 1.73v, SB 1.5v to 1.55v, PL 10 and 9, ai clock twister, all GTL's that fit have been tested, with no improvement and returned to as stated below.

Anyone have any input on settings or experience with thies blue screens?
If the GTL ref are kept within 0.8 to 1.0, is the 1.4 volt max safe FSB termination voltage still valid for long term daily use?

I also get that BSOD from time to time, however I bumped my core voltage up a bit and have yet to see that error again.

[mikeyakame]

The information from Zucker2k, mikeyakame, A_Grey, and others has been a great help in reducing the blue screen occurrences on my friends system at 4Ghz. Yet the blue screens refuse to go away, I need some input.

Goal: 4GHz using the max multiplier of 8.5 and run ram at 1:1.
system specs:
Bios listed below

Xeon X3360 (Q9550) C1 water cooled with a stock fuzion v2 along with EK northbridge wb, laing D5 @ speed 5, and PA120.3 in one loop. Prime temps are 58C, LinX are 67C, northbridge temp is about 30C.

second loop is EK mosfet wb, 2 EK FC-R600 HD 2900XT crossfire, laing D5 @ speed 5, and PA120.3. The south bridge is heatsinked with a low profile copper sink, it stays in the high 30C.

Video cards are a mismatched crossfire of HD2900XT 1GB and HD2900XT 512mb, both water cooled. Blue screens happen with or without crossfire on.

Ram is 2X2GB GSkill F2-8800CL5GBPI (DDR2 1100) in the blue slots, Vista 64bit Ultimate SP1 all patched up. Jmicron controller is disabled, all sata AHCI installed. Power supply is a Silverstone OP-1000 1Kw. The motherboard is a X48 Rampage Formula.

So the problem is windows blue screens saying "A clock interrupt was not received on a secondary processor within an allocated time" or "General hardware fault". This has been going on for months, and thanks to this thred, it usually happens once or twice a day now. The information about AGTL+ bus signaling, GTL ref, bios 803, cpu clock skew, and Zuckers BIOS, decreased blue screens alot.

Memtest can run all day without errors, and processor has primed for 6 hours no errors, no blue screens. The blue screens very rarely occure under heavy load from gaming or priming, only once that I can remember. They occur during idle desktop, internet browsing, and word documents. Never had any issues during windows load screen.

The CMOS was cleared befor and after flashing the bios, and profiles where re-entered and saved by hand.

Load line calibration is enabled, and tested fine. CPUZ and Everest show no Vcore fluctuations between idle and load. The Vcore stays at 1.360v all the time no matter the load. The board was vdroop pencil moded and sealed by tape during inital building.

NB voltage has been up to 1.73v, SB 1.5v to 1.55v, PL 10 and 9, ai clock twister, all GTL's that fit have been tested, with no improvement and returned to as stated below.

Anyone have any input on settings or experience with thies blue screens?
If the GTL ref are kept within 0.8 to 1.0, is the 1.4 volt max safe FSB termination voltage still valid for long term daily use?

Well for starters, AI Clock Twister = Lighter can make your system unstable, hell mine ran worse than with Moderate, to the point it was no better than Stronger.

Moderate for your memory is recommended.

I can probably help with blue screens but I need to know exactly what the blue screen says.

I need all the MCE (machine check exception) data from it. Reason, MCE hex value, memory address locations and what ever else you can give me you might think is valuable.

I've caused myself that many BSOD's that I'm explicitely familiar with most of the common ones, and if not I can at least help you figure out what it means and go from there.

General hardware fault usually means your period stability went too far out, or some other critical minimum value could not be met. In my experience I only see this when there is way too much clock jitter at high FSB or CPU/NB clock skews need adjustment. If you are seeing this usually means you've over stepped your chips FSB limit, and need to knock 1-3mhz off it, unless of course you can resolve it.

Just noticed how low your CPU GTL is, unfortunately you can't run as low as 0.62x on all GTL0-3. I have a strong feeling this is the cause of your problem. Raise Vtt by 0.02v and increase CPU GTL to 0.63 or 0.65x and see if it helps.

I'll explain why. CPU has 4 GTL reference pins, 2 pairs, 0/3 and 1/2. 0/3 are data bus GTLs, 1/2 are address bus GTLs.

GTLREF1/2 are the GTL+ reference voltage points where the CPUs logic receiver samples on the address bus, address strobes are setup on rising clock edges and are shutdown on the falling clock edge. We sample early before ringback has been stabilized we can cross-talk with the other clock mid way during its address strobe, or misinterpret logic swings and open a strobe during wrong clock ending up trying to access protected memory space or addressing.

GTLREF0/3 are the points the logic receiver samples on the data bus, reads are edge aligned on the following clocks rising edge, writes are centre aligned to the following clocks falling edge. Low clock we invert data, high clock we don't invert data. If we read logical high as low, we assume data is inverted next clock and read it back to front. If we read logical low as high, we assume data is not inverted as read it as it is, which is back to front!

We have 2 pins for each pair, we sample twice on a swing, mid and end, If mid we sample high and end we sample high it must be high. If mid we sample low and end we sample low it must be low. If mid we sample low or high, and end we sample low or high we end up causing a general hardware exception because we have conflicting data and we can't proceed so we must log a MACHINE CHECK EXCEPTION stating we have a hardware fault, and what caused it or when it happened. If we get the swings back to front we either corrupt data by reading it as 0h = Fh or Fh = 0h, or we corrupt the destination address we are setting up a strobe too.

Either way we end up with a BSOD hopefully, because a BSOD is better than corrupting data on our system in a bad way.

That's my understanding.

[Alien Grey]

The information from Zucker2k, mikeyakame, A_Grey, and others has been a great help in reducing the blue screen occurrences on my friends system at 4Ghz. Yet the blue screens refuse to go away, I need some input.

Goal: 4GHz using the max multiplier of 8.5 and run ram at 1:1.
system specs:
Bios listed below

Xeon X3360 (Q9550) C1 water cooled with a stock fuzion v2 along with EK northbridge wb, laing D5 @ speed 5, and PA120.3 in one loop. Prime temps are 58C, LinX are 67C, northbridge temp is about 30C.

second loop is EK mosfet wb, 2 EK FC-R600 HD 2900XT crossfire, laing D5 @ speed 5, and PA120.3. The south bridge is heatsinked with a low profile copper sink, it stays in the high 30C.

Video cards are a mismatched crossfire of HD2900XT 1GB and HD2900XT 512mb, both water cooled. Blue screens happen with or without crossfire on.

Ram is 2X2GB GSkill F2-8800CL5GBPI (DDR2 1100) in the blue slots, Vista 64bit Ultimate SP1 all patched up. Jmicron controller is disabled, all sata AHCI installed. Power supply is a Silverstone OP-1000 1Kw. The motherboard is a X48 Rampage Formula.

So the problem is windows blue screens saying "A clock interrupt was not received on a secondary processor within an allocated time" or "General hardware fault". This has been going on for months, and thanks to this thred, it usually happens once or twice a day now. The information about AGTL+ bus signaling, GTL ref, bios 803, cpu clock skew, and Zuckers BIOS, decreased blue screens alot.

Memtest can run all day without errors, and processor has primed for 6 hours no errors, no blue screens. The blue screens very rarely occure under heavy load from gaming or priming, only once that I can remember. They occur during idle desktop, internet browsing, and word documents. Never had any issues during windows load screen.

The CMOS was cleared befor and after flashing the bios, and profiles where re-entered and saved by hand.

Load line calibration is enabled, and tested fine. CPUZ and Everest show no Vcore fluctuations between idle and load. The Vcore stays at 1.360v all the time no matter the load. The board was vdroop pencil moded and sealed by tape during inital building.

NB voltage has been up to 1.73v, SB 1.5v to 1.55v, PL 10 and 9, ai clock twister, all GTL's that fit have been tested, with no improvement and returned to as stated below.

Anyone have any input on settings or experience with thies blue screens?
If the GTL ref are kept within 0.8 to 1.0, is the 1.4 volt max safe FSB termination voltage still valid for long term daily use?

Is it a real ASUS Rampage Formula board?

Whjen it's a real ASUS Rampage Formula board try this.

Use Ai Clock Twister Moderate. DRAM at 1130MHz (DRAM Voltage around 1.94V in BIOS - Mikeyakame should know what Voltage for 1130MHz) tRD 7.

Lower NB Voltage. I don't think you need 1.65V for FSB 471.
I'm using 1.47V (1.45V real) at the moment in the BIOS for FSB 475MHz.

Do you need higher CPU PLL than 1.50V in the BIOS? I'm still using 150V for 3.8GHz 8 X 475MHz.

FSB Termination Voltage is a tricky one. With +1.4V real it's probably to high. I'm using 1.36V (1.26V real) for 3.8GHz 8 X 475MHz.

Try with CPU GTL Voltage reference 0.63X.

CPU Clock Skew and NB Clock Skew should be equal or CPU Clock Skew 100ps more delayed. With a delay of 200ps for the CPU and normal for NB it's probably not good.

Try LLC Disabled. You need higher idle voltage but you won't kill the CPU like that.

LLC is still something I don't understand how it works. I can use LLC Enabled with my Pentium D 840 up to 4.2GHz without problems. Could try higher Clocks but the CPU temperature reaches 70° Celsius under water with CPU Voltage 1.48V.
It runs great with LLC Enabled with my Q9450 at stock speeds. Overclocking my Q9450 with LLC Enabled is a total disaster. I have no idea what is wrong and it gets worse with every new BIOS Update. I can't stay 1 minute stable in Windows XP X64.
I'm not really sure but I think it has something to do with the PCIE and my Graphics Card. At first video signal disappears and a few seconds later my system reboots. It has nothing to do with the CPU Voltage or any other voltage setting.
I've asked the ASUS Global Tech Support to look what could be the problem but I don't get an answer back. It should be easier for them to try to find out what is going on when I enable LLC but I don't think that they care.

[Alien Grey]

Just noticed how low your CPU GTL is, unfortunately you can't run as low as 0.62x on all GTL0-3. I have a strong feeling this is the cause of your problem. Raise Vtt by 0.02v and increase CPU GTL to 0.63 or 0.65x and see if it helps.

I don't think he needs higher FSB Termination Voltage. My opinion it's probably to high already.

The 3.8GHz on safe voltage for 24/7 is possible with the 2 X 2 Giga G.Skill F2-9600 and my Q9450. Maybe 3.84GHz, I'll try that later.

I love that memory. Q9450@3.6GHz 8 X 450MHz tRD 6 Memory at 1200MHz and 1.8V (1.82V real). I could try lower voltage but the BIOS doesn't make it possible.

[Brama]

My opinion it's probably to high already.

Just a linguistic curiosity, if some mother English guy can clarify.

Frequently I read in many messages that too, like in too big, too much, etc., is written to, like in to go, to Paris, etc.

Is it correct the contraction of too in to or is it internet English?


Thanks,

[Zucker2k]

Nt correct.

[Fritz_the_germ]

@Alien Grey,
in my last post I told you about my problems with Memset4. You asked me if I had deleted each folder of memset4.
Yesterday I found a folder called "Memsave" in the windows folder and I am wondering if memset4 has been created that memsave folder when I hit memset4 setup.
I am pretty careful by deleting somethimg in the windows folder because I don´t want to srew up my system.
But if this memsave folder prevents me to run the new installed memset4 allright, I will delete the old memsave folder at once.
So sounds my quest: Does memset4 create a "memsave" folder in the windows folder?
If you don´t know, is it allowed to contact FELIX ?

[Zucker2k]

Just a linguistic curiosity, if some mother English guy can clarify.

Frequently I read in many messages that too, like in too big, too much, etc., is written to, like in to go, to Paris, etc.

Is it correct the contraction of too in to or is it internet English?


Thanks,

No, it's not correct.

@Mike: I think you got the info you wanted me to confirm already, right?

The information from Zucker2k, mikeyakame, A_Grey, and others has been a great help in reducing the blue screen occurrences on my friends system at 4Ghz. Yet the blue screens refuse to go away, I need some input.

Goal: 4GHz using the max multiplier of 8.5 and run ram at 1:1.
system specs:
Bios listed below

The specific error you mentioned .... a clock interrupt was not received on a specific interval... or so it reads; I've found usually, more vcore or vdimm fixes that.

My suggestion for 8.5 x 471 (1:1) - 4Ghz is:

CPUPLL = Lowest setting (manual)
FSBT = 1.24 real (1.32 bios set) at least in my case
NB = 1.35-137 real (with 8GBs running 4-4-4-10, PL7)
CPUGTL= Try all (I recommend you start with x62 first)
NBGTL = x67

[Alien Grey]

@Alien Grey,
in my last post I told you about my problems with Memset4. You asked me if I had deleted each folder of memset4.
Yesterday I found a folder called "Memsave" in the windows folder and I am wondering if memset4 has been created that memsave folder when I hit memset4 setup.
I am pretty careful by deleting somethimg in the windows folder because I don´t want to srew up my system.
But if this memsave folder prevents me to run the new installed memset4 allright, I will delete the old memsave folder at once.
So sounds my quest: Does memset4 create a "memsave" folder in the windows folder?
If you don´t know, is it allowed to contact FELIX ?

You can delete the memsave folder. It's a folder that MemSet creates to save the settings.

[Alien Grey]

Just a linguistic curiosity, if some mother English guy can clarify.

Frequently I read in many messages that too, like in too big, too much, etc., is written to, like in to go, to Paris, etc.

Is it correct the contraction of too in to or is it internet English?


Thanks,

It's Flemish English.

[ChaosMinionX]

Just a linguistic curiosity, if some mother English guy can clarify.

Frequently I read in many messages that too, like in too big, too much, etc., is written to, like in to go, to Paris, etc.

Is it correct the contraction of too in to or is it internet English?


Thanks,

Internet English

[ChaosMinionX]

The information from Zucker2k, mikeyakame, A_Grey, and others has been a great help in reducing the blue screen occurrences on my friends system at 4Ghz. Yet the blue screens refuse to go away, I need some input.

Goal: 4GHz using the max multiplier of 8.5 and run ram at 1:1.
system specs:
Bios listed below

Xeon X3360 (Q9550) C1 water cooled with a stock fuzion v2 along with EK northbridge wb, laing D5 @ speed 5, and PA120.3 in one loop. Prime temps are 58C, LinX are 67C, northbridge temp is about 30C.

second loop is EK mosfet wb, 2 EK FC-R600 HD 2900XT crossfire, laing D5 @ speed 5, and PA120.3. The south bridge is heatsinked with a low profile copper sink, it stays in the high 30C.

Video cards are a mismatched crossfire of HD2900XT 1GB and HD2900XT 512mb, both water cooled. Blue screens happen with or without crossfire on.

Ram is 2X2GB GSkill F2-8800CL5GBPI (DDR2 1100) in the blue slots, Vista 64bit Ultimate SP1 all patched up. Jmicron controller is disabled, all sata AHCI installed. Power supply is a Silverstone OP-1000 1Kw. The motherboard is a X48 Rampage Formula.

So the problem is windows blue screens saying "A clock interrupt was not received on a secondary processor within an allocated time" or "General hardware fault". This has been going on for months, and thanks to this thred, it usually happens once or twice a day now. The information about AGTL+ bus signaling, GTL ref, bios 803, cpu clock skew, and Zuckers BIOS, decreased blue screens alot.

Memtest can run all day without errors, and processor has primed for 6 hours no errors, no blue screens. The blue screens very rarely occure under heavy load from gaming or priming, only once that I can remember. They occur during idle desktop, internet browsing, and word documents. Never had any issues during windows load screen.

The CMOS was cleared befor and after flashing the bios, and profiles where re-entered and saved by hand.

Load line calibration is enabled, and tested fine. CPUZ and Everest show no Vcore fluctuations between idle and load. The Vcore stays at 1.360v all the time no matter the load. The board was vdroop pencil moded and sealed by tape during inital building.

NB voltage has been up to 1.73v, SB 1.5v to 1.55v, PL 10 and 9, ai clock twister, all GTL's that fit have been tested, with no improvement and returned to as stated below.

Anyone have any input on settings or experience with thies blue screens?
If the GTL ref are kept within 0.8 to 1.0, is the 1.4 volt max safe FSB termination voltage still valid for long term daily use?

You have a very similar setup, try my settings from here...

http://www.xtremesystems.org/forums/...postcount=1965

I have since bumped my core voltage a little bit, but nothing else has changed.

[Fritz_the_germ]

Hi folks!
I´m planung to do the "pencil mod" because my system isn´t stable when I set LLC disabled. To enable the LLC makes to system stable but I am afraid to demage the 45nm cpu for 24/7 setting.
Isn´t the sens of the LLC to prevent short time high voltages?
Does anybody has experiences having done to "pencil mod"?


How can I´ll be Xtreme Member?