Works well with Ryzen 5600X and XMP settings.
Works well with Ryzen 5600X and XMP settings.
I recently installed a new PC: CPU Ryzen 5 5600x with Hyper 212 Black Edition Cooler Mobo, Gigabyte Aorus Elite X570 (latest BIOS). RAM G.SKILL Trident Z RGB Series 32GB (2x16GB) DDR4 3600MHz CL17 Dual Channel Memory Kit at 1.35V. I ran standard benchmarks—cinebench r23, 3DMark, Unigine Valley, Crystal Disk Mark. Monitored via HWInfo64 and everything seemed normal, with a peak temperature of 70°C during 3DMark. When I enabled XMP, performance dropped noticeably; HWInfo64 reported the all-core boost at 4.3GHz instead of the expected 4.65GHz. RAM was running at full speed. Why is my CPU clock lower under XMP? How can I fix this? Should I manually adjust the RAM to a slower frequency, like 3200MHz? Please note—I’m just starting out.
Ensure no link exists between the two. Their interaction is limited to the CPU and memory controller sharing the same package. Turn off the PC, wait a short time, restart, and if it restarts again, you might consider manually lowering the speed to 3433 or 3200 MHz.
Idle: Below 30°C, Gaming performance near 50°C, Benchmark at 60°C (70°C in 3DMark)
I adjusted XMP settings and examined HWInfo. The difference between XMP and standard mode is visible in "Frequency Limit - Global." With XMP enabled, frequency drops during Cinebench tests. I still achieved 4.6GHz with all core boosts, but performance falls when XMP is active. My 3DMark scores improved from 7497/5765 to 8301/5800 after enabling XMP. Thermals are excellent. Who can explain the contrast between Cinebench and 3DMark results? Is it related to power consumption in Cinebench?
CInebench relies on AVX operations which are often criticized as "power-hungry" due to their high energy consumption. Manufacturers adjust motherboard settings to accommodate these instructions, setting an AVX offset that typically ranges from 100 to 200 MHz below the peak turbo frequency. While 3DMark mainly evaluates graphics cards, CPU assessments use more standard instruction sets that demand less power and are not labeled as power-hungry. They still maintain full CPU utilization but operate at a faster clock speed compared to AVX code.
It seems the Cinebench r23 result is lower when using XMP settings. I saw some discussions about this, but most people haven't tested it without XMP enabled. However, the 3D Mark Time Spy score did improve, which might explain why it's different. I'm still trying to figure out what's happening.