Line Adjustment Calibration and additional details
Line Adjustment Calibration and additional details
Hi everyone, can someone clarify whether the LLC setting influences only the Vcore or if it also affects other parameters such as FSB term, CPU PLL, RAM, and MB?
I remember when I had my old E8400 about a few years ago. It was stable at 4.06 GHz for roughly four years. Now I own an E5450 (E0) in my P5Q Deluxe. The first one I tried to get was stable at 3.84 GHz for just two weeks, which made me really frustrated—I ended up buying another E5450 E0 for 24 euros.
At that time, I thought it was a bad chip. But now I have the second one, which runs at 3.96 GHz and is only nearly stable with the FSB and Vcore. I always use LLC in my BIOS, but if I change any other voltage settings—like FSB term, CPU PLL, MB, SB—it causes random crashes in Prime95 LARGE FFT. These crashes happen randomly, sometimes within minutes and sometimes after about six hours. The latest memtest also reports random failures during memory tests.
So, does LLC affect RAM as well? If I set the maximum rated voltage in BIOS to 2.1V for G.Skill DDR2 1066MHz RAM (which runs at 3.96 GHz, FSB 440, Strap 400), will it actually work with 2.1V under LLC or does it cause a voltage drop to the RAM?
Also, does MemTest86 only test the RAM or also the CPU cache? I’m wondering if a failure in MemTest indicates a RAM issue or if it could be due to a CPU cache problem, which would then show up as a RAM failure in the test results?
I’ve had good ITB tests and stable temperatures, but the crashes are random.
Additionally, regarding Prime95 tests: I understand they run one after another. If the voltage settings aren’t correct, it will crash at that exact moment during the test. The same applies if I run a different test—after a few hours or even two hours, it might fail. But I’m not sure if it’s random like this.
I’m currently running 2x2 GB RAM modules. These are quite challenging to overclock. I remember that with my E8400, it worked fine for four days at 3 GHz, and I didn’t push it beyond that. I was satisfied because I found the minimum voltages it could handle above 4 GHz without trying to go higher.
With these two E5450 units, it’s a different story—I can’t crack them! They’re not capable of much more than what most people expect.
My current setup is:
- FSB: 440
- Strap: 400
- Vcore: 1.2875 (LLC-on)
- BIOS: multi-x.9.0
- Everything else: Auto
Please let me know your thoughts!
In those earlier times, Loadline calibration mainly adjusted the voltage shown, increasing it by a fixed amount just like manually changing it. It gave users a sense of satisfaction because they perceived a lower voltage setting. For voltages tied to a proportion of the vCore, such as GTL Ref, they would naturally increase too.
Additionally, C2 Quads consistently proved more challenging to overclock compared to Duos.
Currently, VRMs are faster, which seems to improve performance, but now CPU voltages are extremely low—often exceeding 100A—making it even more difficult to manage overshoot and undershoot.
In those earlier times, Loadline calibration mainly adjusted the voltage shown, increasing it by a fixed amount just like manually changing it. It gave users a sense of satisfaction because they perceived a lower voltage setting. For voltages set at a fraction of the vCore, such as in GTL Ref, they would also rise noticeably.
Additionally, C2 Quads consistently proved more challenging to overclock compared to Duos. Nowadays, with faster VRMs, performance seems better, but CPU voltages are extremely low—often exceeding 100A—which complicates managing overshoot and undershoot.