Choose quicker memory performance or more precise clock speeds.
Choose quicker memory performance or more precise clock speeds.
So last year in August 2019 I assembled a budget Ryzen setup. It featured a B450 Tomahawk processor, a Ryzen 2700 chipset, DDR4 3000 MHz RAM with two 8GB sticks. It wasn’t the most powerful configuration, but it was within my budget. Later, I upgraded to a 5700 Sapphire Pulse. I transferred my old Intel components and installed my Seasonic Focus Gold 750W PSU along with a Master ML360R 360mm AIO cooler. I’ve attached photos of the RAM and timings details. Below are screenshots from Ryzen DrmCal showing what the timings look like. You’ll notice if you keep RAM at 3000 MHz, the timings are quite tight. If you try to overclock RAM to 3200 MHz, the timings become much looser. Some tech communities suggest going for faster RAM speeds with looser timings, while others prefer lower speeds and tighter settings. My question is: should I stick with 3000 MHz and tight timings, or push RAM up to 3200 MHz with looser specs? Or should I attempt to overclock RAM to 3400 MHz and check if the system boots properly? Thanks for your time.
The IMC excels with quicker memory, that’s clear. (It’s strict about board and RAM options though.) The real challenge lies in the memory setup. 16-18-18 3000mhz feels extremely sluggish. E-Die performs better if it operates at a much higher frequency. When testing overclocks, use AIDA64 to check latency and cache performance at default settings. After adjusting the RAM speed, compare the results of both overclock attempts. It’s well known that higher memory speeds reduce latency and boost read/write speeds, though this also depends on CPU speed—keep that consistent across tests. I’d doubt a 3400mhz frequency would succeed without significant boosts in voltage and SOC, but only you can confirm that by trying it. By the way, the calculator isn’t the complete picture—it offers a baseline; my own experience shows the numbers don’t always match reality. A few sub-timings are overly constrained, making optimization harder.