The 5950X will not exceed 3.6 GHz.
The 5950X will not exceed 3.6 GHz.
Without PBO, at full load reaching around 3.6GHz across all cores, temperatures stay near 58-62°C (as noted, cooled with NH-D15 and Noctuas thermal paste). That indicates the temps aren’t a concern. It seems there might be an issue, as a friend with a similar TUF X570 board and a 5950X (which differs from mine in CPU-Z results) is experiencing the same problem—around 3.6GHz on all cores during Cinebench testing. I’ve attached his CPU-Z data for reference.
I recall a setting in the BIOS that prevented the CPU from boosting, but after resetting defaults it’s probably not active. It might relate to virtualization or something similar. Try these steps: 1: Use this tool from Overclock.net to check for core issues. 2: Install hwinfo64 and run it in sensor mode. 3: Execute the core cycler from step 1 and observe boost clocks while it cycles through cores. Your configuration should display boost clock values in the maximum column of hwinfo64. On my machine, this is for a 5950x on an Asus X570. If you don’t see at least 4,700 MHz, your CPU may be faulty or only barely meets AMD’s quality standards.
I've returned to BIOS, performed a reset (loaded optimized defaults), and then activated DOCP for my 3200MHz CL16 RAM. After that, I ran the test as described. Using the default configurations, I completed the CB23 multi-core benchmark. A screenshot from Ryzen Master shows TDC and EDC reaching their limits, while all cores stayed at 3.6GHz. The temperature was around 60°C, visible in RyzenMaster. The final score was 23.500 points.
Thanks for your input. This seems quite unusual. The core effective clock is showing 5.3 Ghz, which doesn’t align with normal operation—typically a working CPU’s effective clock should closely match its actual speed. The Cinebench results also show a discrepancy: a reported boost of 4.95 Ghz but an effective clock between 3.6 and 3.8 Ghz. Also, after running the core cycler for six hours, the chip reached around 1.475v, which is lower than expected. I don’t see the actual SV12 TFO voltage, but it’s likely in the 1.46v range, which is even more puzzling. If you’re comfortable with a fresh Windows install, I’d suggest backing it up and creating a new installation with all drivers copied to a USB drive. After setup, install the AMD chipset drivers, restart, then run audio, graphics, and finally Cinebench. This might help identify the problem.
It appears there could be software or configuration issues—possibly related to Windows Update on an AMD CPU with Ethernet enabled. Some users report unexpected behavior with AMD chips, especially when drivers auto-install.
If you’re open to it, sharing more details about your system would help. Otherwise, this situation might require deeper investigation into BIOS settings, driver conflicts, or firmware updates.
For reference, you can check forums like Overclock.net for similar discussions: https://www.overclock.net/forums/amd-cpus.10/