CPU Benchmark Performance: AI and Inferencing

As technology progresses at a breakneck pace, so too do the demands of modern applications and workloads. With artificial intelligence (AI) and machine learning (ML) becoming increasingly intertwined with our daily computational tasks, it's paramount that our reviews evolve in tandem. Recognizing this, we have AI and inferencing benchmarks in our CPU test suite for 2024. 

Traditionally, CPU benchmarks have focused on various tasks, from arithmetic calculations to multimedia processing. However, with AI algorithms now driving features within some applications, from voice recognition to real-time data analysis, it's crucial to understand how modern processors handle these specific workloads. This is where our newly incorporated benchmarks come into play.

As chip makers such as AMD with Ryzen AI and Intel with their Meteor Lake mobile platform feature AI-driven hardware within the silicon, it seems in 2024, and we're going to see many applications using AI-based technologies coming to market.

We are using DDR5-5200 memory as per the JEDEC specifications on the Ryzen 7 8700G and Ryzen 5 8600G, as well as DDR4-3200 on the Ryzen 7 5700G and Ryzen 5 5600G. The same methodology is also used for the AMD Ryzen 7000 series and Intel's 14th, 13th, and 12th Gen processors. Below are the settings we have used for each platform:

  • DDR5-5200 CL44 - Ryzen 8000G
  • DDR4-3200 CL22 - Ryzen 5000G
  • DDR5-5600B CL46 - Intel 14th & 13th Gen
  • DDR5-5200 CL44 - Ryzen 7000
  • DDR5-4800 (B) CL40 - Intel 12th Gen

(6-2) DeepSpeech 0.6: Acceleration CPU

(6-3) TensorFlow 2.12: VGG-16, Batch Size 16 (CPU)

(6-3b) TensorFlow 2.12: VGG-16, Batch Size 64 (CPU)

(6-3d) TensorFlow 2.12: GoogLeNet, Batch Size 16 (CPU)

(6-3e) TensorFlow 2.12: GoogLeNet, Batch Size 64 (CPU)

(6-3f) TensorFlow 2.12: GoogLeNet, Batch Size 256 (CPU)

(6-4) UL Procyon Windows AI Inference: MobileNet V3 (float32)

(6-4b) UL Procyon Windows AI Inference: ResNet 50 (float32)

(6-4c) UL Procyon Windows AI Inference: Inception V4 (float32)

(6-1) ONNX Runtime 1.14: CaffeNet 12-int8 (CPU Only)

(6-1b) ONNX Runtime 1.14: CaffeNet 12-int8 (CPU Only)

A major focal point of AMD's Ryzen 8000G series is the inclusion of the Xilinx-based Ryzen AI NPU. While AI benchmarks and those measuring capabilities using large language models (LLMs) are thin off the ground, none of our benchmarks utilize the NPU itself. Much of the Ryzen AI NPU is based and, as such, is focused on enabling software features such as those generative AI capabilities within Microsoft Studio Effects and software such as Adobe and Davinci.

In ONNX Runtime using the utilized INT8 model, we can see that the Ryzen 7 8700G and Ryzen 5 8600G don't offer world-beating AI performance, but we intend to investigate this more deeply.

Using the latest firmware, which removes the STAPM limitations, we can see that the Ryzen 5 8600G shows the most gains, especially in DeepSpeech 0.6, where we saw a 12% bump in performance. The Ryzen 7 8700G also posted some very impressive gains in the UL Procyon Windows AI Inferencing benchmark, with a 34% jump in performance in our charts, but this could be a case where it underperformed in the MobileNet V3 test in the first place.

CPU Benchmark Performance: Science And Simulation iGPU Gaming Performance: 720p And Lower
Comments Locked

111 Comments

View All Comments

  • zodiacfml - Monday, January 29, 2024 - link

    thanks but it would have been nicer for me with an i3-12100 in the charts. The 13100f or 12300f tests from old reviews not comparable.
  • TheinsanegamerN - Tuesday, January 30, 2024 - link

    Take 13100f and extrapolate. Not that hard.
  • PeachNCream - Tuesday, January 30, 2024 - link

    Not a bad CPU overall, though it does absolutely devour electrical energy. Competition is far worse, but that shouldn't justify a CPU alone consuming more power than would be required to provide illumination to an entire home - it's worth eleven 800 lumen lights! In the evenings or night, I usually have four or fewer bulbs active for half the power consumption or less than this CPU at sub-maximum workloads WITHOUT the rest of the supporting components a PC requires to provide useful functions. Perspective makes it obvious that's quite terrible when we live on a world that is overpopulated, polluted, and hanging on the precipice of being unable to sustain enough food production to feed us and we all know what happens when humans are hungry and forced to compete for limited resources.
  • TheinsanegamerN - Tuesday, January 30, 2024 - link

    If you're that scared of power use, buy a celeron mini PC and be quiet.
  • PeachNCream - Tuesday, January 30, 2024 - link

    Some of us have to care because it's obvious a lot of us don't and have our heads buried in the sand.
  • TheinsanegamerN - Wednesday, January 31, 2024 - link

    Why are you using a PC if you care? why are you not in a commune growing organic crops by hand if you care so much?

    Nobody cares about your virtue signaling.
  • PeachNCream - Wednesday, January 31, 2024 - link

    Clearly you feel threatened enough by your own lifestyle choices to care by going on the attack and suggesting some extreme alternative like this commune nonsense as if suggesting it eliminates any slight adjustment to your own actions that could offer a reduction in the guilt you're coping with by lashing out.
  • erotomania - Wednesday, January 31, 2024 - link

    How exactly is a 65W processor with graphics "gobbling power"? If you mean inefficient, i suppose we could discuss, with facts. But these are modern Ryzen cores, with some mobile genetics - I don't think inefficient applies.

    In the past I had some Richland APUs (with FX cores) that were definitely inefficient but still idled as low an anything else. I have a 5600G system that idles so low my UPS can't detect it, event though when not idle the system is OC'ed. I would not characterize either as gobbling power.
  • PeachNCream - Thursday, February 1, 2024 - link

    I've already somewhat pointed out why there the consumption is a considerable factor. ~87W at full load as indicated on AT's measurements is enough to provide illumination for an entire home. That isn't a comment on efficiency (or work accomplished for power expended) and I didn't indicate that in my initial post. It's an observation about the power cost implications and impacts of a PC when a single component consumes that much energy and it still, as a standalone device, is an incomplete representation of overall power consumption of a PC built around it.

    And, it's fair to point out that it is NOT an Intel CPU with far higher consumption. I also mentioned that as well in the same post. AMD's CPUs demonstrate a better work-to-wattage ratio so please realize that I'm aware that among all desktop CPUs, this particular chip is far from the worst possible option.
  • maxijazz - Saturday, February 3, 2024 - link

    Maybe people feel threatened because proud woke people (aka communists-fascists) want to enforce their lifestyle on others? By lobbied out new laws, by propaganda, by censorship.
    Otherwise nobody would feel threatened.
    Live your life and let others live theirs.
    If you like force others for good of society, world or universe, go to North Korea. They have communism as state's religion.

Log in

Don't have an account? Sign up now