Xeon paired with Noctua cooler achieves lower base clock during testing without throttling.
Xeon paired with Noctua cooler achieves lower base clock during testing without throttling.
Hello. I recently replaced my processor from an i5-4460 to a Xeon E3-1230v3. The switch seemed smooth overall, except for the issue where the Intel cooler caused the CPU to hit 90°C or more, sometimes slowing down when it exceeded that temperature. I bought a used Noctua NH-U14S and it worked wonders, bringing temperatures down to around 60°C maximum. The challenge is with benchmarking: using CPU-Z's Stress and Bench tools, the CPU consistently stays at 3.5Ghz during idle and load (base clock 3.3, turbo 3.7). Single Thread Stress also caps at 3.5Ghz. However, when I ran Cinebench R23 or Blender Benchmark, the CPU drops to 3.4Ghz for a few seconds before settling back to 3.3Ghz, then briefly jumps to 3.5Ghz afterward. Temperatures remain stable at 40°C idle and peak at 60°C under load. I’m puzzled by this behavior. I disabled Intel SpeedStep and it didn’t help. My power plan is set to High Performance, but switching to Balanced doesn’t change the issue. The motherboard is a GIGABYTE B85M-D3H v2. The CPU uses Corsair Vengeance 2x4GB DDR3 at single channel (a pin on my board stopped me from using other slots). The graphics card is an MSI Gaming X GTX 1650 Super with 4GB, and the PSU is a budget 500W unit. The cooler I installed performed well. Any advice would be appreciated. Edit: Prime95 stress test mirrors this pattern—CPU drops to 3.2Ghz during testing and then returns to 3.6/7 afterward. Motherboard remains unchanged.
It's normal for the CPU to show stress at high frequencies like 3.5Ghz, especially when using tools such as CPU-Z. This pattern appears consistently in Cinebench and Blender Benchmark but not always in other tests. Consider running additional benchmarks like 4K Video Rendering or Disk Benchmark to get a fuller picture.