Napkin math research: Impact of uncore and RAM rates/time on results
Napkin math research: Impact of uncore and RAM rates/time on results
Recently, conversations have centered around the impact of uncore overclocking and whether RAM boosts Intel configurations. Feeling a bit bored, I decided to test it on my setup—specifically an Intel 8700K system paired with a Vega 64. I ran three versions: one with 50 cores, 47 cores, and 50 cores respectively, using different RAM speeds (16-36 MHz). The results showed varying performance trends across the runs. While some metrics improved slightly, others didn’t show significant gains. I ended up sticking with my default settings, keeping RAM at 3200 MHz and uncore at its highest possible. It seems uncore alone makes a small difference, but pairing it with faster RAM helps more in CPU work rather than games. Overall, the data isn’t overwhelming, but it suggests that higher-frequency RAM can still be beneficial for certain tasks. I’ll keep my RAM at 3200 XMP and try tweaking the uncore further next time.
Honestly, I didn’t attempt anything. I just set the frequency to 3600 and let the motherboard handle the timing. I might look into it more later.