Testing Zen 2 for gaming results in improved performance and stability.
Testing Zen 2 for gaming results in improved performance and stability.
Many opinions have emerged lately about the optimal CPU choice for gaming in specific settings. I’m considering running some tests myself. The setup would include: CPU model 3700X, 2x8GB RAM, a 3600 GPU. Expecting performance to be challenging. My most unused unit is likely a Vega 56. Should I run at 1080p high to ease the GPU load and shift more work to the CPU, or should I go even lower? For CPU scaling experiments, I plan to set the 3700X to 4.0 GHz and adjust core configurations. Options include 4+4 SMT (standard), 4+4 without SMT (may skip this), 3+3 SMT (similar specs), 3+3 without SMT (around 3500), 4+0 SMT (around 3300X), 2+2 SMT (around 3100). I want to fix the clock speeds to 4 GHz to minimize variations from turbo, cooling, etc. My goal is to see how configurations scale without relying on specific CPU models. For RAM scaling, with the 3600 kit I can try single or dual channels at different frequencies. Lastly: which games? I need either the game itself or a free demo/benchmark. The titles should be current and relevant. I have three choices: SOTTR benchmark (recent), Shadowbringers for FFXIV, or GTAV—though I’m unsure if it’s stable enough. I’m open to suggestions. You might not know your own library, so focus on games with demos or benchmarks. Demos should support a benchmark mode. I expect similar tests already exist, so I can skip this if a solid one exists.
I’d suggest 1080 and 1440 for casual use. Any other resolution isn’t really practical for everyday needs. I’d skip the exact numbers you mentioned and go with something like 9700k. For a stock cooler might work, but voltages could differ. It’s better to leave the boost clock as it is—consumers don’t care about that. A cooler that’s widely available would be fine, whether for games or casual use. Maybe consider XFR or PBO settings if you already have a compatible CPU like Rise of the Tomb Raider. For games like GTA V or Overwatch, a less demanding setup could be better. I’m not a gamer, so I don’t know much about games. This should help clarify things.
I ran some tests after updating my system from a 1800x to a 3700x resolution at default settings for the CPUs. It turned out there weren’t significant improvements for gaming.
I'm performing these tests myself using a 3700X processor, 1080Ti, 32GB RAM, and a 1440p display. Activating Precision Boost Overdrive should give you a solid mix of single and multi-threaded performance. I believe it’s the most effective method for real-world results rather than just theoretical scores found on gaming sites. Based on my experience, the boost clocks remain consistent across runs, likely because of the 360mm Corsair AiO cooler. You’re probably not testing with the original cooler, which might affect outcomes.
The test structure seems better for demonstrating performance across varying numbers of cores rather than directly contrasting different CPU models.
You're looking to see how various factors such as core count, RAM speed, and frequency influence benchmark results. It helps to understand what works best depending on these settings. If latency changes by just a little—like 1-2 ms—then it's fine. But if the difference is more significant, aim for a balanced setting like 16 cores and 16 GB RAM.
You're looking to understand if upgrading from 6 to 8 cores offers noticeable benefits for gaming performance. It's about evaluating whether the extra cores make a meaningful difference in speed and overall experience.
Yes, skip the core clocks. Looking for something newer with solid thread scaling for testing. GTAV remains a contender, but I’m considering alternatives that might scale better. SOTTR works well so far, while FFXIV is an older title with lower scaling.