Checking the CPU time for your Ryzen 5 5600X. Is the value normal? There are ways to improve it.
Checking the CPU time for your Ryzen 5 5600X. Is the value normal? There are ways to improve it.
Hey everyone, I recently started diving into Overclocking and was aiming to boost my Time Spy score (think "Scarface-Push it to the limit" vibes). Here’s what my PC setup looks like:
- Motherboard: MSI X570 gaming edge WiFi
- CPU: Ryzen 5 5600X
- GPU: RX 6900X
- RAM: Crucial Ballistix 3600MHz CL16
- PSU: 850W fully air-cooled
- Current score: High in Timespy – [link]
To reach this, I made these tweaks:
- Raised GPU power limits via MorePowerTool (GPU=340)
- Set TDC limits at A-GFX=370
- Overclocked with Radeon software
- Tuned RAM timing to fast timing & max clock
- Adjusted GPU voltage to 1.4V, left at 0%
- Applied a custom fan curve for better cooling
- Reduced core clocks to 4750MHz
This pushed CPU latency from 65ns to 58.8ns and boosted clock speed to 3733. Latency dropped to 1866, with all cores hitting 4750MHz.
I tried PBO but it didn’t help much compared to manual overclocking in Ryzen Master. I experimented with limits, offsets, and voltages, but still struggled to exceed 9000 CPU points.
Questions:
- Should I revert RAM clocks to 3600 for tighter timing?
- Why doesn’t my GPU want to clock above 2700 even if it’s not limited by power or heat?
- Did I misconfigure PBO settings?
- Is it normal for all-core overclocking to give better results?
Also, since I have SAM enabled, maybe that’s affecting things. Let me know your thoughts!
It's more than just PBO; it's PBO plus a per-core undervoltage adjustment using the curve optimizer. This setup helps maximize performance from your chip. You've managed a solid overall overclock, yet single-core results are still lacking.
Your performance looks solid from my perspective. It seems your setup isn't just for quick tests—18310 points is reasonable. Still, aiming for better memory and Infinity Fabric speeds (1:1) would help. With such a fast benchmark, tight timing won’t be the main issue; focus on maximizing throughput instead.
Only since you weren't using PBO alongside an undervolt was the focus. Also, I highlighted single-core performance because that's what really counts for gaming. I'm not sure how much it affects 3DMark, but since it mainly tests 3D graphics speed, it could still matter.
Are you referring to adjusting the volume by utilizing the curve optimizer?