The conversation highlights a significant variation despite similar components.
The conversation highlights a significant variation despite similar components.
I put together 2 identical units:
CPU: Currently testing [Q9550], but I've used many [QX9650][QX9770] [Q6600][E5200]
Memory: Asus P5Q-SE2 (link provided)
RAM: DDR2 800MHz
Case: Chieftech 2008
Heatsink: Artic cooling (budget)
PSU: Bequiet
PC1: Works flawlessly at 3900 MHz with the Q9550 (459x8.5) and remains stable above 4000 with Extreme processors
PC2: Unstable even at 3400 MHz with the Q9550 (400x8.5)
Same motherboard, I've tried all options, but the performance gap is incredible! Which P45 model fails to reach 430 MHz?
I believe the issue lies with the MB: a P45 motherboard which tends to cause problems at 420 FSB speeds.
In winter 2012/2013 I constructed two PCs for my son and myself using a 3770k. I was somewhat fortunate to receive a solid 3770k, achieving 4.5 Ghz at roughly 1.18V (for everyday operation 4.3 Ghz at 1.14V). My son's 3770k required 1.14V for 4.0 Ghz. As it increased to 4.2 (around 1.18v), it became quite hot, so I kept it at 4.0 for daily use. Both units had the same memory and were cooler. As others have mentioned: silicone lottery. No two pieces are exactly alike.
I believe the same applies to memory and even motherboards. VRM FETs face the same challenges as other components, with capacitors and inductors being influenced by similar factors as 'electrolyte' and 'ferrite'. Therefore, at the extreme end, some samples might offer more consistent power output than others. This is why manufacturers include design margins in their systems, which overclocking seeks to take advantage of. Some designs naturally provide better performance.