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 memory on the Core i9-14900K, Core i7-14700K, Core i5-14600K, and Intel's 13th Gen at the relative JEDEC settings. The same methodology is also used for the AMD Ryzen 7000 series and Intel's 12th Gen (Alder Lake) processors. Below are the settings we have used for each platform:

  • DDR5-5600B CL46 - Intel 14th & 13th Gen
  • DDR5-5200 CL44 - Ryzen 7000
  • DDR5-4800 (B) CL40 - Intel 12th Gen

Note: As we are running with a completely refreshed CPU test suite, this means we are currently re-testing other processors for our data sets. These will be added to the below graphs as soon as we have more results, and these will also be added to our Bench database. Thanks for your understanding.

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

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

(6-1c) ONNX Runtime 1.14: Super-Res-10 (CPU Only)

(6-1d) ONNX Runtime 1.14: Super-Res-10 (CPU Only)

(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)

Digesting performance in the latest addition to our CPU test suite for 2024, it's clear that the extra L3 cache on the Ryzen 7000X3D processors has a clear benefit in ONNX when using the INT8 model. Even in ONNX with Super-Res-10, the AMD Ryzen CPUs seem to do a better job of inferencing than Intel. It does appear that more L3 cache can benefit performance in this area, and even in TensorFlow with VGG-16, the AMD Ryzen 9 7950X seems to perform the best. Outside of that, the Core i9-14900K is next best, just marginally ahead of the Core i9-13900K. 

The Core i7-14700K performs respectively, and while the Core i5-14600K can certainly handle AI and inferencing, it doesn't offer quite the grunt of the Core i9 and Core i7 series chips.

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

57 Comments

View All Comments

  • bananaforscale - Wednesday, October 18, 2023 - link

    Why?
  • nandnandnand - Wednesday, October 18, 2023 - link

    AMD has not been generous with multi-threading at the entry level, e.g. 7600/7800X.

    I don't agree that they have to start selling 32-cores soon. They have a few ways to address multi-threading with Zen 5 or Zen 6. The easiest one for Zen 5 would be to make a 24-core with an 8+16 configuration. But that probably won't stop them from releasing an entry level 6-core.

    If they boost the core counts of the normal/fast chiplets in the future, then core counts will rise across the board. For example, a 12-core chiplet would probably get disabled to 8-10 cores for the entry level, instead of 6. A 16-core chiplet could get disabled to 10-12. That is not happening with Zen 5 as far as we know.
  • SanX - Wednesday, October 18, 2023 - link

    Why Intel and AMD not make enthusiastic 100-200 core 6-7 GHz processors with TDP 1 kW for those who don't care how much they consume because already have 10-20 kW solar panel systems on their roof ? Let others be jealous. Green energy proponents will be dancing in joy.
  • bananaforscale - Wednesday, October 18, 2023 - link

    Shut up, troll.
  • cmdrdredd - Wednesday, October 18, 2023 - link

    No, there IS a place for ALL the performance at ANY cost. Some people are like that really. Not everyone lives and dies based on power consumption like the comments have you believing. I don't care, I can pay the bill and it's really peanuts difference between a 14900k and a 7950x in terms of the difference it would make on the power bill. If I am looking for max performance for my usage that isn't even a consideration, only performance matters. By that measure the 14900k is better sometimes and sometimes not. So it would come down to very specific use cases. What I'm saying is, everyone here moans about power usage but ignores the fact that not everyone pinches pennies on the power bill or needs to worry about it and just wants the max performance. That's why 4090s exist and sell well.
  • ItsAdam - Wednesday, October 18, 2023 - link

    I own a 4090, still wouldn't buy a 14900k, was gonna buy then they decided to refresh.

    I dislike the heat element. This much possible power draw is broken,how can you cool that and maintain peak speeds. It's pointless.

    The only ones who could use these are people on propper custom loops or ln2 overclocking.

    They are literally pointless CPUs and the new gen of intel CPUs was just a total waste of time, development and retail. Just think how much waste they've made for this new gen that's the same, just the packaging and everything what a waste.

    Pointless. Like your post.
  • SanX - Thursday, October 19, 2023 - link

    I also dont buy this gen Intel processors, so what?Cheap solar energy and efficient heat pumps will slowly change mentality. Now power supply in PC 1kW, soon they will be 2 kW and rarely who will care. People don't care about power efficiency. 99.9% don't even know the price per kilowatt-hour. Kids playing games dont even want to hear about it. If you care then find the solutions how to use that possible 0.25-1 megawatt of solar power falling on your property. Power efficiency is just the salespeople buzzword
  • The Von Matrices - Wednesday, October 18, 2023 - link

    They already do make super-high TDP processors. They're called EYPC and Xeon.
  • SanX - Thursday, October 19, 2023 - link

    Server and supercomputer high core count chips heavily rely on power efficiency, hence they have clocks 2-3 GHz. For example the 64 core EPIC processor essentially equivalent to the future 24-32 core consumer processor at 6 GHZ at 3-5x cost. If someone worry too much about power consumption just buy couple used solar panels for 100 bucks and you will cut the cost by half or 100% with smart backup
  • PeachNCream - Wednesday, October 18, 2023 - link

    Those peak power numbers are disgustingly bad. I wouldn't want to pay the utility bill for that so I'm glad I paired a keyboard to my phone and make do with less than 8w peak total system power consumption. Oddly enough, I don't feel as though I'm missing anything without some obnoxious box filled with a CPU like one of these and some obesity-level graphics card.

Log in

Don't have an account? Sign up now