CPU Performance

To start off our testing I went with the rendering-focused Blender benchmark, specifically the latest version, Blender Benchmark 4.2. To get a good look at the overall performance I have all three of the results stacked together and in this test, the Ryzen 7 9800X3D sits ahead of the Core Ultra 5 245K but below both the 14900K and Core Ultra 9 285K by a significant margin.

graph1

Continuing with the video encoding theme I also have Handbrake which is an open-source transcoder. For this test I am taking a 4k video down to 1080p 30 FPS, the results are the average FPS of that task. This is right with last generation’s Ryzen 9 7950X and 7950X3D with it averaging 147.5 FPS when encoding the test video. This was a big improvement from the 7800X3D though which averaged 108 in this same test. Intel is still way out in front here however with the 285K out in front and the last few flagships are all still ahead of the Ryzen CPUs tested. I should note however that we don’t currently have the Ryzen 9 9950X or 9900X to see where they would fit.

graph2

For the always popular Cinebench, I am testing with the older Cinebench R23 as well as the newer 2024 edition as well. I always like Cinebench because we have multi-core and single-core performance which gives us a good look at the performance of the whole CPU and IPC performance. In R32 the 9800X3D was right with the Intel 14600K and 13600K in the multi-core tests but did improve on the single-core performance compared to the 7950X and 7950X3D. In Cinebench 2024 that single-core performance was even better scoring 133 putting it on par with the 14900K. The multi-core result is still down with the 14600K and 13600K however.

graph3

graph4

Next up we have ray tracing-focused rendering benchmarks. Here I tested the CPUs in POV-Ray and V-Ray Benchmark 5. POV-Ray was also tested with a single core and across all of the cores. The Ryzen 7 9800X3D scored a 766.53 for the single core result in POV Ray, this wasn’t enough to catch up to the highest-end Intel CPUs but was a big step forward compared to the 7800X3D which scored 685.21. Its multi-score result was 7718, again just behind the 13600K and 14600K. In V-Ray Benchmark 5 the Ryzen 7 9800X3D was finally not running with the XX600K’s, this time it came in behind the Ryzen 9 7900 with its 18557. That was 4110 higher than the 7800X3D however.

graph5

graph6

I also tested using the CPUz’s built-in benchmark which does run on a single thread and with all threads. In the multi-thread test, the Ryzen 7 9800X3D scored an 819.1 for its single-threaded performance and 8630.6 for multi-threaded performance. That’s a 29% improvement over the 7800X3D in the multi-threaded test and 17.5% for single-threaded. The single-threaded result put the 9800X3D up ahead of the Core Ultra 5 245K but still behind the Ryzen 7 9700X. Multi-threaded however, had the 9800X3D up over the 9700X but still behind older Intel mid-range CPUs like the 13600K and 14600K.

graph7

graph8

Next, we have wPrime which is a classic overclocking benchmark that calculates pi out to 1024 million digits and is timed. This is a multi-thread heavy test which has the older high-core count CPUs still all over the top of the charts and you can see that with the top CPUs sitting at 8 or more cores. The Ryzen 7 9800X3D scored a 54.122 which was 8 seconds slower than the 285K but was 13 seconds faster than the 7800X3D.

graph9

7 Zip is another open-source program, this time for compressing and decompressing all of your files. Here I have run the benchmark and we have three results. The combined MIPS is a combination of compressing and decompressing performance. Then I have it broken down between the two. The combined results have the Ryzen 7 9800X3D with a score of 138441 MIPS and sitting behind the 16 core 7950X and an impressive 59% improvement over the 7800X3D. Both the decompressing and compressing results were simple with the decompressing numbers for the Ryzen 7 9800X3D sitting a little ahead of the compressing result for the 9800X3D.

graph10

graph11

Jetstream 2.1 is a compilation benchmark that takes a long list of HTML5 and Java in-browser tests and runs them all three times and puts together an overall score. I love this benchmark because let's be honest, most people are using their browser more than any other game or program. Jetstream gives some interesting results sometimes though but it tends to prefer high IPC or single-core performance and there are some situations where it prefers lower core-count CPUs. In this case, improved clock speeds of the 9800X3D helped it a lot, putting it in second behind the Ryzen 7 9700X.

graph12

As a new addition to our testing, I have added in a few AI-focused tests using Geekbench AI and Procyon’s Computer Vision Benchmark. The Geekbench AI test looks at single and half-precision performance with scores for each and a quantized score as well. They test using 10 different workloads to create their score. The Procyon AI Computer Vision Benchmark focuses on machine vision tests using neural network models like MobileNet V3, Inception V4, YOLO v3, DeepLab V3, Real-ESRGAN, and ResNet 50. I test those using the Windows ML setting and in cases that support it, I test built-in NPUs as well. I have also included Windows ML and TensorRT results using an RTX 4090 as well as a reference point where these CPUs compare to a high-end GPU. For Geekbench AI the Ryzen 7 9800X3D did well on the quantized score and the single precision score but struggled with the half precision score. For comparison I had the 285K as well as the 285K tested using its NPU as well and the 285K really stands out in the quantized score. In the Procyon test, I tested using a float32 model and the Ryzen 7 9800X3D came in below the 285K but ahead of the 245K. Of course, the 4090 still dwarfs the CPU results still. With the Integer and Float16 models the Ryzen 7 9800X3D struggled with the Float16 but did outperform the 285K when using Windows ML. Of course, using the built-in NPU the 285K and 245K both doubled the performance of the 9800X3D.

graph13

graph14

graph15

For Passmark Performance Test I used the latest version, Performance Test 11. I only look at the overall CPU score which takes a few different synthetic benchmarks and combines the results to put together an overall score. This is a test that does favor multi-threaded performance over IPC. In Performance Test 11 the Ryzen 7 9800X3D scored a 40043, putting it once again right with the 14600K.

graph16

Crossmark is from BAPCo which also makes SYSmark and this is a cross-platform test where you can compare performance between phones and both Windows and Mac computers. This is the start of our overall PC benchmarks and Crossmark uses a mix of real-world tests to output an overall score. The 9800X3D scored 2383 here and was just behind the i9-149900K for this test, More impressively though it was ahead of the 16 core 7950X3D and improved on the 7800X3D by 11%.

graph17

PCMark 10 is a great test because it tests things like video calls, browser performance, Excel, and Word performance to give an idea of real-world performance. It tends to like higher clock speeds but does take raw core count into account as well which you can see. Given how well the 9700X did in this test previously, it's not too big of a surprise that the 9800X3D is sitting just behind it and ahead of both the 14900K and 285K from Intel here. Its 10473 was an 11% improvement from the 7800X3D. In the PCMark Applications test the 9800X3D scored 17160, almost tied with the 14900K.

graph18

graph19

Moving the focus over towards gaming my next test is using Dolphin 5.0 Benchmark. Dolphin 5.0 is a Wii emulator and like most emulators, it doesn’t care about high core counts at all. In fact, it only runs two in total. Clock speeds are king here most of the time which is why all of the 5 GHz+ CPUs are at the top of the chart here. The Ryzen 7 9800X3D improved on the already good results we saw from both the 9700X and 9600X and topped our chart with 161-second completion of the test. This is 22% and 47 seconds faster than the 7800X3D.

graph20

Before diving into game testing I wanted to check out synthetic performance using 3DMark. I tested with the older DX11 Fire Strike test as well as the newer DX12 Time Spy. I also added the new 3DMark CPU Profile benchmark into the mix as well which does a good job of showing the full range of thread counts that you might see being used in games as well as a max threads option that does everything above 16 threads. In the Fire Strike test, the Ryzen 7 9800X3D scored 39525, a 20% improvement over the last generation 7800X3D. That wasn’t enough to put the Ryzen 7 9800X3D up near the top, however. Time Spy was similar, with a 20% improvement once again but this time the Ryzen 7 9800X3D was the top performing Ryzen CPU tested, of course without the 9900X and 9950X those would change that. But Intel still dominates here with the entire top of the chart being all Intel. Last up I ran the 3Dmark CPU profile and the results there were similar as well but with a 25% improvement over the 7800X3D this time around.

graph21

graph22

graph23

I did of course test the two new CPUs in games as well. I will dive into testing the integrated GPU in the next section. Here though they are paired up with a powerful GPU (the RTX 4090) and I compare CPU to CPU performance. The Ryzen 7 9800X3D, like the previous two generations of 3D V-Cache CPUs before it, is designed specifically to outperform in gaming. Did it do that here? Yes, yes it did. In all eight of the games tested it topped our charts. In fact, in all but one, it isn’t even close. The Ryzen 7 9800X3D dominates in gaming leaving the Intel Core Ultra 9 285K and the 7950X3D and 7800X3D in the dust.

graph24

graph25

graph26

graph27

graph28

graph29

graph30

graph31

 

Log in to comment

We have 1020 guests and one member online

supportus