In my post benchmarking the Zotac Gaming GeForce RTX 3080 Trinity on a X470 motherboard with Ryzen 7 2700X the results I saw were considerably lower than what most reviewers had posted. This was to be expected, with me using lower-spec hardware elsewhere.

With the RTX 3080 being so powerful and PCIe 4.0 based, many have wondered how much of a bottleneck is caused by either the CPU or lower bandwidth available on PCIe 3.0 motherboards.

So I decided to upgraded my motherboard to an X570 in preparation for the Ryzen 5000 and borrowed a Ryzen 7 3700X for testing.

I won’t claim my testing is anywhere near as rigorous as more PC focussed websites. I have no formal benchmarking policy. I mainly use the built-in benchmarks but I also use Fraps while gaming to test if things are similar.

Test Set Up

My system was:

• CPU: AMD Ryzen 7 2700X
• GPU: Zotac Gaming GeForce RTX 3080 Trinity
• Motherboard: Asus ROG STRIX X470-F GAMING
• RAM: Kingston HyperX Fury 32GB 3600MHz
• Storage: Kingston KC2500 1TB NVMe
• Monitor: Samsung C49RG90 49-inch WQHD 1440p 120Hz

The X570 motherboard I decided on is the MSI MEG X570 UNIFY Motherboard, it is unique for many motherboards nowadays in that it has no RGB at all. However, it does have 2.5GbE and Wi-Fi 6, while being considerably cheaper than the MSI MEG X570 ACE.

For the game tests, I ran them at 1440P and then 5120×1440 because that’s the native resolution of the ridiculously large Samsung C49RG90.

Everything was tested at stock though the RAM had the XMP profile set to 3600MHz.

I meant to benchmark the AMD Ryzen 7 2700X on the X570 to see how much a difference just the PCIe 4 would make, but time limitations means I have been unable to do this yet (because this is my main work/pc gaming computer). I will do at the earliest opportunity. Probably when I return the 3700X.

Benchmark Results

3DMark

I ran the usual 3DMark range of tests, no results were particularly shocking. Moving from the 2700X to the 3700X achieved a moderate increase in the results.

There was a difference with the 3700X on the X470 and X570 but not enough to justify running out to buy a new motherboard.

Assassin’s Creed

A similar scenario with the Assassins Creed benchmark, the change of CPU saw at best 5.7% improvement. Adding in the X570 did improve things further, going up 8.1% but that only represents a 3.8% difference between using the same CPU on an X470 vs X570 motherboard.

Crysis Remastered

On this test, Crysis doesn’t offer 5120×1440 so I used 3440×1440.

Things are very different from the recently launched Crysis Remastered, both the CPU change and the PCIe 4.0 equipped X570 motherboard had a significant difference in my results.

The difference between the 3700X on the X470 motherboard vs X570 was 10% on 3440×1440.

Far Cry New Dawn

In Far Cry New Dawn the difference was again significant with bother swapping the CPU and motherboard.

At 5120×1440 there is a 7.8% difference moving from the X470 to X570

Horizon New Dawn

Things swapped back again for Horizon New Dawn with basically no difference between the CPU or motherboard change.

Overall

In my very limited testing, there does seem to be a significant difference in some games when using an X570 motherboard with PCIe 4.0 compared to the X470.

This seems to contradict what everyone is reporting, in particular, I noticed Kitguru tested the differences with Horizon Zero Dawn and Far Cry and reported much smaller differences.

At lower resolution they reported 4% then 3% at 1440p, However in Far Cry, they reported no differences with and AMD CPU switching between X470 and X570, and the PCIe 3.0 Intel system offered up to a 21% advantage.

There is a good chance that it is my testing that is flawed, most of my benchmarks are significantly lower than more professional reviews, so something is holding the performance back.