Your Dual Core CPU is Hurting Your PC!

This video shows how a weak dual core CPU can bottleneck your entire system, especially when benchmarking or testing modern graphics cards. It also walks through upgrading both the power supply and processor to restore accurate performance results on a dedicated test bench machine:

Originally I built this test bench machine from spare parts I had laying around. I just needed something that I could easily plug various cards into to see if they worked. Since I didn’t want to wait for parts to arrive, and also because I’m cheap, I just built it with whatever I had available.

This worked perfectly for a while. I was able to slot cards into it, often older graphics cards, to either identify them or make sure they would at least power up. However, when I got to the point where I wanted to benchmark the cards or stress test them to see how they behaved under more long term use, I ran into a problem.

The ancient 80 gig mechanical hard drive that I was using really wasn’t large enough to hold both Windows 10 as well as all the various benchmarking programs I wanted to use. Also, the old hard drive was slow as dirt. So I switched it out with the best thing I could scrounge up, which was a 128 gig SSD. However, that 128 gig still wasn’t quite enough space, so not more than a couple weeks later I broke down and bought a cheap 256 gig SSD to replace it with.

Everything was going great for a while. However, I’ve now run into two more challenges. One, the generic 300 watt power supply being used is old and doesn’t have any PCI connectors on it. I was able to get around this limitation recently while testing RX 560 and RX 570 graphics cards by using a cable like this. This is a Molex to six pin PCI connector, so I was able to combine the two Molex rails on the power supply into a single six pin PCI connection. This worked well enough for running the tests, but it’s not really a long term solution.

The second challenge is the CPU. The CPU in this machine is an AMD Athlon 200GE. This 2 core 4 thread chip is very much on the low end when it comes to processors that can even fit into this AM4 motherboard. It wasn’t really apparent when testing old graphics cards at 720p or lower. However, recently while testing overclocks for the GTX 1050 at 1080p, the overclocking improvements seemed a little bit low in some of the titles being tested.

Looking at the result for the Shadow of the Tomb Raider test that was run shows the problem. The test is showing that it was GPU bound only 60 percent of the time. The low end Athlon 200GE CPU was holding back the test and producing less accurate results.

So two things we need to do with this machine. One, we’re going to switch out the generic 300 watt power supply with a 600 watt bronze Cooler Master. And two, we’re going to switch out the Athlon CPU with a Ryzen 5 1600 CPU. This six core 12 thread processor should be a major step up for this machine and allow us to more accurately test a greater range of graphics cards. Once we flip these parts out, let’s run some of the same benchmarks we originally did with the GTX 1050 and see what’s changed.

All right, look at that, looks beautiful. Now let’s see what kind of difference the CPU upgrade made. I’m going to run through five benchmarks quickly to see how performance may have been improved. All these benchmarking programs are free. I’ll include links to them in the description.

First up, since this is a CPU upgrade, we’ve got Cinebench. Original scores with the Athlon came in at 784 single core and 2096 multicore. The Ryzen scores came in at 949 single core and 6403 multicore. These are improvements of 21 percent and 205.5 percent. So needless to say, it’s a major upgrade.

Next on the list, to test more of the change to the overall system, we’ve got PCMark 10 demo. The original Athlon score came in at 3420, while the Ryzen upgrade score came in at 4450. That’s an increase of 30.1 percent, a beautiful increase to system wide performance in general.

Third up is 3DMark demo. The original Athlon score for Time Spy came in at 1886, while the Ryzen upgrade achieved a score of 2098. That’s an increase of 11.2 percent, so not bad at all.

The number four test is the Final Fantasy XIV Endwalker benchmark tool. Running at 1080p on the Standard preset, the original Athlon benchmark achieved a score of 10496 or High rating. The Ryzen upgrade was able to deliver a benchmark score of 12819 or Very High rating. That’s an increase of 22.1 percent. This is an excellent example of how a CPU of insufficient power can affect gaming performance.

Finally we have the Shadow of the Tomb Raider trial benchmark tool. Running at 1080p on the Low preset, the original Athlon benchmark achieved a frames rendered score of 61.01 and an average FPS of 39. After the Ryzen 5 upgrade, frames rendered increased to 69.73 with an average FPS of 45. That’s a frames rendered increase of 14.3 percent.

Not only is that increase impressive, but looking at the benchmark footage for Shadow of the Tomb Raider revealed even more. In the original Athlon footage, there are spots where there are significant delays in rendering some parts of the scene. With the Ryzen footage, those same portions have no delays and everything renders smoothly.

So overall, this CPU upgrade has proven to have been a much needed change to this test system. All in all, I’m very pleased with it. Now if you have any questions, please feel free to leave a comment on the video, and thanks so much for watching!

https://www.maxon.net/en/downloads/cinebench-r23-downloads
https://store.steampowered.com/app/524390/PCMark_10/
https://store.steampowered.com/app/223850/3DMark/
https://na.finalfantasyxiv.com/benchmark/
https://store.steampowered.com/app/750920/Shadow_of_the_Tomb_Raider_Definitive_Edition/