Can't stabilize E8500 OC with DFI LP LT X48

[Rpg]

Hello!

I recently built a new rig with an E8500 and a DFI LP LT X48 board. I did some quick OCing and made my way to 4Ghz. The problem is that I can't get it Prime stable. It never crashed once during the OC process and it would boot up no matter what the clock is (into Windows), but it would error out in Prime either in 2 minutes, 15-20 minutes, or an hour and 15 minutes with a rounding error.

I've tried bumping up the CPU voltage, NB voltage, and VTT (read somewhere that it's the same as FSB voltage on my old Gigabyte DS3L board). I've OCed my last rig which was a quad, but this is the first time I've had a C2D. I've tried the OC with 1.4v on the NB, 1.25v on VTT, and 1.36v on the CPU and it would error out in Prime still. It's always a rounding off error. My old quad rig would crash in Prime if it wasn't stable so I don't know what's the problem

Once, it would was Prime stable for more than an hour and 15 minutes. Prime ran for 2 hours and 40 minutes before I closed Prime to use my computer. The next day, it would error out at practically the same settings in one of the three lengths of time.

I'll post my BIOS settings in a bit, but my RAM is at 5-5-5-15 at 842Mhz with 2.1v, the CPU is at 421x9.5, and almost everything is on auto except for the CPU features. Multi Core CPU, Thermal Monitoring, Execute Disable Bit, and Virtualization Technology are enable. The C1E and E1ST are disabled.

In short, can anyone help me stabilize my OC? Let me know if I need to type up any more info. Do I need to mess with GTL/REF voltages? Everything runs fine and there are no problems, but I don't know if it's fully stable.

--Edit--

Here's the BIOS. My board overvolts the CPU (probably because of the CPU VID Add Special) so it's 1.36v idle and 1.344v load in CPUZ.

PHP Code:

[b]DFI LT X38-T2R LanParty[/b]
[b]CPU Feature[/b]
- Thermal Management Control: Enabled
- PPM(EIST) Mode: Disabled
- Limit CPUID MaxVal: Disabled
- CIE Function: Disabled
- Execute Disable Bit: Enabled
- Virtualization Technology: Enabled
- Core Multi-Processing: Enabled

Exist Setup Shutdown: Mode 2
Shutdown after AC Loss: Disabled
CLOCK VC0 divider: AUTO
CPU Clock Ratio Unlock: Enabled
CPU Clock Ratio: 9.5x
Target CPU Clock: 3999 Mhz
CPU Clock: 421
Boot Up Clock: Auto
DRAM Speed: 400/800
Target DRAM Speed: DDR2 842
PCIE Clock: 100
PCIE Slot Config: 1X 1X

CPU Spread Spectrum: Disabled
PCIE Spread Spectrum: Disabled
SATA Spread Spectrum: Disabled

[b]Voltage Settings[/b]
CPU VID Control: 1.3125v
CPU VID Special Add: Auto
DRAM Voltage Control: 2.120v
SB Core/CPU PLL Voltage: 1.510v
NB Core Voltage: 1.330v
CPU VTT Voltage: 1.100v
Vcore Droop Control: Enabled
Clockgen Voltage Control: 3.45v
GTL+ Buffers Strength: Strong
Host Slew Rate: Weak
GTL REF Voltage Control: Disable
x CPU GTL1/3 REF Volt: 110
x CPU GTL 0/2 REF Volt: 110
x North Bridge GTL REF Volt: 110

[b]DRAM Timing[/b]
- Enhance Data transmitting: AUTO
- Enhance Addressing: AUTO
- T2 Dispatch: Disabled

[b]Clock Setting Fine Delay[/b]
Ch1 Clock Crossing Setting: AUTO
- DIMM 1 Clock fine delay: Current
- DIMM 2 Clock fine delay: Current
- Ch 1 Command fine delay: Current
- Ch 1 Control fine delay: Current


Ch2 Clock Crossing Setting: AUTO
- DIMM 3 Clock fine delay: Current
- DIMM 4 Clock fine delay: Current
- Ch 2 Command fine delay: Current
- Ch 2 Control fine delay: Current

Ch1Ch2 CommonClock Setting: Auto

Ch1 RDCAS GNT-Chip Delay: Auto
Ch1 WRCAS GNT-Chip Delay: Auto
Ch1 Command to CS Delay: Auto

Ch2 RDCAS GNT-Chip Delay: Auto
Ch2 WRCAS GNT-Chip Delay: Auto
Ch2 Command to CS Delay: Auto 

CAS Latency Time (tCL): 5
RAS# to CAS# Delay (tRCD): 5
RAS# Precharge (tRP): 5
Precharge Delay (tRAS): 15
All Precharge to Act: AUTO
REF to ACT Delay (tRFC): AUTO
Performance LVL (Read Delay) (tRD): AUTO

Read delay phase adjust: Enter

Ch1 Read delay phase (4~0)
- Channel 1 Phase 0 Pull-In: Auto
- Channel 1 Phase 1 Pull-In: Auto
- Channel 1 Phase 2 Pull-In: Auto
- Channel 1 Phase 3 Pull-In: Auto
- Channel 1 Phase 4 Pull-In: Auto

Ch2 Read delay phase (4~0)
- Channel 2 Phase 0 Pull-In: Auto
- Channel 2 Phase 1 Pull-In: Auto
- Channel 2 Phase 2 Pull-In: Auto
- Channel 2 Phase 3 Pull-In: Auto
- Channel 2 Phase 4 Pull-In: Auto

MCH ODT Latency: AUTO
Write to PRE Delay (tWR): AUTO
Rank Write to Read (tWTR): AUTO
ACT to ACT Delay (tRRD): AUTO
Read to Write Delay (tRDWR): AUTO
Ranks Write to Write (tWRWR): AUTO
Ranks Read to Read (tRDRD): AUTO
Ranks Write to Read (tWRRD): AUTO
Read CAS# Precharge (tRTP): AUTO
ALL PRE to Refresh: AUTO 


Thanks, and sorry for the long read.

[affiliate13]

Originally Posted by Rpg

Hello!

I recently built a new rig with an E8500 and a DFI LP LT X48 board. I did some quick OCing and made my way to 4Ghz. The problem is that I can't get it Prime stable. It never crashed once during the OC process and it would boot up no matter what the clock is (into Windows), but it would error out in Prime either in 2 minutes, 15-20 minutes, or an hour and 15 minutes with a rounding error.

I've tried bumping up the CPU voltage, NB voltage, and VTT (read somewhere that it's the same as FSB voltage on my old Gigabyte DS3L board). I've OCed my last rig which was a quad, but this is the first time I've had a C2D. I've tried the OC with 1.4v on the NB, 1.25v on VTT, and 1.36v on the CPU and it would error out in Prime still. It's always a rounding off error. My old quad rig would crash in Prime if it wasn't stable so I don't know what's the problem

Once, it would was Prime stable for more than an hour and 15 minutes. Prime ran for 2 hours and 40 minutes before I closed Prime to use my computer. The next day, it would error out at practically the same settings in one of the three lengths of time.

I'll post my BIOS settings in a bit, but my RAM is at 5-5-5-15 at 842Mhz with 2.1v, the CPU is at 421x9.5, and almost everything is on auto except for the CPU features. Multi Core CPU, Thermal Monitoring, Execute Disable Bit, and Virtualization Technology are enable. The C1E and E1ST are disabled.

In short, can anyone help me stabilize my OC? Let me know if I need to type up any more info. Do I need to mess with GTL/REF voltages? Everything runs fine and there are no problems, but I don't know if it's fully stable.

Thanks, and sorry for the long read.

Hello, Welcome to the forum.
Please follow this link and update your specs for us to get a broader picture of your rig.
http://www.clunk.org.uk/forums/forum...r-profile.html

thanks

[Rpg]

Originally Posted by affiliate13

Hello, Welcome to the forum.
Please follow this link and update your specs for us to get a broader picture of your rig.
http://www.clunk.org.uk/forums/forum...r-profile.html

thanks

Hrm, sorry. Normally I would do that stuff ahead of time as I'm on OCN and they have something similar to that too where you put in your system/rig. Sorry for having you repeat something like that. I see people ask questions without filling in their rig info and then the first reply is always fill in the system spec.

--Edit--
Done. I even threw in the companies/brands in case it would make a difference.

[Clunk]

Welcome to the forums

Please grab the BIOS template from the review thread and list your settings...we can see whats going on

[Kelainefes]

Welcome!
Although it's unlikely the case, your memory is certified for 800 MHz operation and you are using it at 842. What is the rated voltage on those sticks?

[Rpg]

Originally Posted by Clunk

Welcome to the forums

Please grab the BIOS template from the review thread and list your settings...we can see whats going on

Sure thing. I'll edit my first post with it.

Originally Posted by Kelainefes

Welcome!
Although it's unlikely the case, your memory is certified for 800 MHz operation and you are using it at 842. What is the rated voltage on those sticks?

They're rated at 800Mhz @ 4-4-4-12 timings at 2.0v-2.1v. I did a 7-8 hour memtest run and it ran for 12 passes (all the tests) without errors at those timings (2.1v). I'll go back and loop test 5 overnight.

[Rpg]

I looped Test 5 in Memtest86 overnight and it had no errors for ~300 passes. I'm now sure it's not the RAM that's the problem.

If I was to up the voltage higher than 1.3625v or whatever the Intel spec is, how much higher should I go?

[MustRotate]

Well like the Q9450 the E8500 has a maximum vcore of 1.45. So dont exceed that if you want your cpu to run for a long time.

[Clunk]

Originally Posted by Rpg

I looped Test 5 in Memtest86 overnight and it had no errors for ~300 passes. I'm now sure it's not the RAM that's the problem.

If I was to up the voltage higher than 1.3625v or whatever the Intel spec is, how much higher should I go?

This question comes up a lot, along with "what is a safe voltage for x".

The short answer is that once you start to go out of the Intel guidelines, there is no "safe" voltage.

Anything over the recommended voltage and you potentially risk damaging the CPU - as long as you know this and are happy to proceed, then all is good.

In reality, 65nm CPUs can often take a lot of punishment. 45nm CPUs tend to degrade quicker and there are other contributing factors that can and will damage or kill a CPU, and among those are too high VTT or CPU PLL voltage. I've had a few CPUs now that just wouldn't overclock very well with any stability after plying them with added VTT or PLL

Heat will also degrade a CPU over time, so keep it as cool as possible.

I would say that if you are just benching for benching's sake, then voltage isn't so much of an issue, but for 24/7 use, I would always recommend to keep the VTT and CPU PLL at stock, and the CPU voltage under 1.4 (this is just my personal preference).

Another thing to keep in mind if you are chasing a 4Ghz overclock (or whatever) is how much of a difference in speed you will actually feel when using your PC when compared to say, 3.6Ghz - In all honesty, it isn't a massive amount for day to day tasks, but it may require extra voltage to get there, so you need to ask yourself if the extra heat and power usage is worth the small increase in speed.

Also, have a look at some of Praz's posts, he likes to tweak the board to death to get higher clocks with lower voltage....patience of a saint that fella

[Rpg]

Originally Posted by Clunk

This question comes up a lot, along with "what is a safe voltage for x".

The short answer is that once you start to go out of the Intel guidelines, there is no "safe" voltage.

Anything over the recommended voltage and you potentially risk damaging the CPU - as long as you know this and are happy to proceed, then all is good.

In reality, 65nm CPUs can often take a lot of punishment. 45nm CPUs tend to degrade quicker and there are other contributing factors that can and will damage or kill a CPU, and among those are too high VTT or CPU PLL voltage. I've had a few CPUs now that just wouldn't overclock very well with any stability after plying them with added VTT or PLL

Heat will also degrade a CPU over time, so keep it as cool as possible.

I would say that if you are just benching for benching's sake, then voltage isn't so much of an issue, but for 24/7 use, I would always recommend to keep the VTT and CPU PLL at stock, and the CPU voltage under 1.4 (this is just my personal preference).

Another thing to keep in mind if you are chasing a 4Ghz overclock (or whatever) is how much of a difference in speed you will actually feel when using your PC when compared to say, 3.6Ghz - In all honesty, it isn't a massive amount for day to day tasks, but it may require extra voltage to get there, so you need to ask yourself if the extra heat and power usage is worth the small increase in speed.

Also, have a look at some of Praz's posts, he likes to tweak the board to death to get higher clocks with lower voltage....patience of a saint that fella

I was thinking of something around 1.37-1.39v, but now my system seems stable. It got past the 2min, 15min, hour and a half point in Prime and went up to 4 and a half hours stable before I stopped it to use the comp.

I would like to learn how to get the same clocks for less voltage though. That would be nice. Right now I'm sitting on 1.36v idle in Windows for 4Ghz. I find it a bit high for a 4Ghz from an E8500. That's just what I think.

Thanks MustRotate and Clunk! I'll just keep my voltage 1.36v right now. Hopefully I can dumb down the CPU voltage. I found that things were unstable because of my NB voltage. It might of been the thing causing me instability. I bumped down the voltage a notch and everything ran fine.

--Edit--

Should I flash my BIOS (shipping BIOS of 3/20) to the new beta one? How do you flash the BIOS with DFI boards? My old Gigabyte one had their own BIOS flashing program from Windows that worked well for me.