For the first time in a very long time I have switched over my main enthusiast rig to an AMD platform, moving away from my very familiar Intel platforms in the last decade. AMD has almost come out of nowhere with Ryzen and completely changed the game in the last few years, all culminating in the Zen 2 CPU and X570 chipset.
I have switched over to AMD's new Ryzen 7 3700X processor and ASRock X570 Taichi motherboard, with a super-crazy high-speed PCIe 4.0 NVMe M.2 SSD that pumps up to 5GB/sec and 16GB of G.SKILL Trident Z Royal DDR4-3200 memory.
I personally play on either my 32-inch monitor with a native 2560x1440 at 144Hz or the super-high-end 4K 144Hz display, and I consistently use the GeForce RTX 2080 Ti as my personal graphics card as it is the best. I have stuck with Intel CPUs because they have, at least until now, been superior for these types of gaming workloads.
So for me to enjoy my gaming sessions of Overwatch and Apex Legends at high FPS, I have to have a great platform to maintain that performance and the new Ryzen 7 3700X and X570 combo really kicks ass. I have moved from a mix of the Intel Core i7-8700K and Core i9-9900K processors to the new Ryzen 7 3700X and in Overwatch and Apex Legends, we have near equal performance.
The benchmarks will show a little otherwise, but for the most part I have been experiencing performance that I can't really tell between the 8700K and 9900K with the new 3700X. AMD has stepped it up to the point of enthusiasts and gamers alike are able to confidently buy a new Ryzen 3000 series CPUs and have a kick ass time.
Fan Speed on NVLink/SLI
Now that I've got a new high-end AMD Ryzen 7 3700X rig set up, I thought I would do some new testing with:
- NVIDIA GeForce RTX 2080 Ti in NVLink with STOCK fan speed
- NVIDIA GeForce RTX 2080 Ti in NVLink with 100% fan speed
- NVIDIA GeForce GTX 1080 Ti in SLI with STOCK fan speed
- NVIDIA GeForce GTX 1080 Ti in SLI with 100% fan speed
These tests show some interesting results, with up to 10% more performance when the fans are cranked up -- keeping those hot GPUs cooler allowing GPU Boost to do its thing on Pascal, and more optimal, on Turing.