Overclocking Barrier - Need Advice

[LausDeo86]

I need advice on overcoming an overclocking barrier I hit.

Here are my PC Specs:
Intel Core 2 Duo E6750 @ 2.66 GHz (333 MHz default FSB)
Mugen 2 CPU cooler, small northbridge fan
Radeon HD 4890
Motherboard: Gigabyte P35-DS3L, Rev. A2, BIOS Version: the latest F9C
4 GB of Crucial RAM, 2 x 2 GB DDR2 800 (PC26400), http://www.newegg.com/Product/Product.aspx?Item=N82E16820148212

I am overclocking for the first time but have read a ton about overclocking including guides on Intel overclocking and DS3L overclocking and am sure I have done well so far.

The highest stable overclock I could manage is the following:

CPU Clock Ratio: 8x
PCI Express Frequency (Mhz): 100 Mhz
CPU Host Frequency (Mhz): 430
System Memory Multiplier: 2.00
CAS Latency Time: 5
DRAM RAS# to CAS# Delay: 5
DRAM RAS# Precharge: 5
Precharge delay (tRAS): 15
ACT to ACT Delay (tRRD): Auto
Rank Write to READ Delay: Auto
Write to Precharge Delay: Auto
Refresh to ACT Delay: Auto
Read to Precharge Delay: Auto
Static tRead Value: 11
Static tRead Phase Adjust: Auto
DDR2 Overvoltage Control: +0.2V
FSB Overvoltage Control: Normal
(G)MCH Overvoltage Control: Normal
CPU Voltage Control: 1.46875V
CPU Enhanced Halt (C1E): Disabled
C.I.A.2: Disabled
CPI EIST Function [Speedstep]: Disabled

A little extra info: With the above settings, my core temps at full load were 53 degrees C and my northbridge was at 47 degrees C so my temps are very good. If I try the same above settings except with my RAM’s default 4-4-4-12 timings, Memtest86+ has errors in minutes so I did have to loosen my timings a bit to be stable. The above settings passed the prime95 In-place large FFTs test for 14 hours with 0 errors.

Now, whenever I increase the CPU Host Frequency to 435 MHz, I can NEVER get it stable no matter what my settings. I always leave my System Memory Multiplier to 2.00 to make sure my RAM isn’t holding me back. Also, I ran Memtest86+ with 435 MHz and it passes no problem. If I have the same stable settings as described above except I have 435 MHz, it fails prime95 by quitting out of it. If I try the same thing except with a higher 1.48125 CPU V vcore, one core fails after 1 hour, 46 minutes in prime95. If I increase CPU vcore to 1.50000 V prime95 fails by quitting out of it. If I try the previous one except add +0.1V to FSB overvoltage control and +0.1V to (G)MCH Overvoltage control, one core fails at 8 hours, 40 minutes. If I try the same thing except with +0.2V to FSB overvoltage control instead, it fails prime95 by quitting out of it.

I then tried to loosen my static tRead value to see it that helps:
CPU Host Frequency (Mhz): 435
Static tRead Value: 12
FSB Overvoltage Control: +0.1V
(G)MCH Overvoltage Control: +0.1V
CPU Voltage Control: 1.48750V
And the above settings failed by quitting out of prime95 after a number of hours.

I tried to loosen my RAM timings even farther (very loose for this overclockable RAM):
CPU Host Frequency (Mhz): 435
System Memory Multiplier: 2.00
CAS Latency Time: 6
DRAM RAS# to CAS# Delay: 6
DRAM RAS# Precharge: 6
Precharge delay (tRAS): 18
Static tRead Value: 11
FSB Overvoltage Control: +0.1V
(G)MCH Overvoltage Control: +0.1V
CPU Voltage Control: 1.48125V
And the above settings failed by quitting out of prime95

I am at my wits end to figure out how to push my PC past the 430 MHz mark. Any ideas? By the way, my RAM is rated at 2.00 V and I have my DDR2 Overvoltage set to +0.2V to make sure that in BIOS it does indeed say 2.00 V for DDR2. I haven’t dared increase it beyond that because I don’t’ want to fry my RAM and because my RAM is already super overclockable at stock voltage according to what I know, check out this review of my RAM to see:
http://www.tomshardware.com/reviews/ddr2-ram-memory,2079-6.html

Any help is appreciated.

 

[jsc]

You have done everything right as far as I can see. You didn't even fall into the trap of trying to OC your RAM. You have simply reached the limits of your CPU chip.

Here's a summary of my systems:
GA-EP45-UD3P, Q9550 OC'd to 3.6 GHz/TRUE, 4 GB Crucial Ballistix,
GTX260-216, 3 WD HD's - 2.64 TB total, Corsair 750TX, Antec 900

GA-EP45-UD3L, Q6600 OC'd to 3.6 GHz/Xig Dark Knight, 4 GB Crucial Ballistix,
XFX 4870, 2 WD HD's - 820 GB total, Corsair 550TX, Antec 900

GA-EP35-DS3P, E5200 OC'd to 3.78 GHz/Xig Dark Knight, 4 GB Crucial Ballistix,
640 MB 8800GTS, 2 WD HD's - 820 GB total, Antec TP3 550, Antec 900

All good cases for cooling and better than average HSF's. But look at System #3. Not bad for a bargain basement CPU, but others here were able run past 4.0 GHz. Like you, I hit a brick wall. Mine was at 3.78 GHz (315 Mhz X 12). I cannot use the 12.5 internal multiplier and 316 MHz doesn't work. I am well under voltage and thermal limits. I even tried 318 MHz at 1.6 volts. Nope. It will run Prime95 small fft's all week at 3.78 GHz. at a vcore of 1.40 volts, but no faster.

Clearly a case of "YMMV." Your Mileage May Vary.
----------
Overclocking since 1978 - Z80 (TRS-80) from 1.77 MHz to 2.01 MHz

 

[LausDeo86]

Jsc,

Thanks for your reply. It is very helpful for me to know that my brick wall is simply because my hardware is stretched to its limits and not because I'm doing something wrong and that there isn't anything else I can do.

It helps me to accept my 3.5 Ghz brick wall overclock even though others could get a higher 4 GHz overclock with my CPU knowing that I'm not the only one who has experienced the inability to get the same limits others can. I looked it up and it seems a 3.5 GHz wall for my CPU is reasonably expected, as indicated by someone's post: "E6xxx series cpus usually are harsh to get past the FSB1600 mark and ~3.5ghz core clock - check what others have been able to get."

I'm now going to overclock my RAM to finish my overclocking up. Overclocking has been fun (albeit frustrating at times), but once you get past the learning curve, it really isn't that bad it seems.

Thanks again.