Hi everyone,
I’m seeing a consistent and reproducible difference between All-Core and Per-Core Curve Optimizer behavior on my system and I’m trying to find out if this is BIOS-specific or expected AGESA behavior.
System
CPU: Ryzen 7 5800X
Motherboard: ASUS ROG STRIX B550-F Gaming (non-WiFi)
BIOS: 3621
AGESA: ComboAM4v2 (as reported in BIOS)
Cooling: sufficient (no thermal throttling)
Settings (baseline)
Observed behavior
1) All-Core Curve Optimizer
2) Per-Core Curve Optimizer
All cores individually set to −27
OCCT stable
All-core boost under load drops to ~4400 MHz
Same behavior regardless of whether I adjust the strongest or weakest core (e.g. −30 on a single core still results in ~4400 MHz)
So even with identical numeric offsets, Per-Core CO consistently results in a much lower all-core frequency than All-Core CO.
What I’m trying to find out
Has anyone with this exact board (or similar ASUS B550/X570) tested other BIOS versions where Per-Core CO behaves identically to All-Core CO under all-core load?
In other words:
All-Core −27 ≈ Per-Core all −27 in terms of all-core frequency (~4.7 GHz), not ~4.4 GHz.
I understand that Per-Core CO is mainly for single-core optimization, but I’m specifically interested in whether this frequency drop is a BIOS/AGESA implementation issue, or if there are known versions where this behavior was different.
Any concrete BIOS version references or firsthand test results would be greatly appreciated.
Thanks!
