Summarizing CPU and GPU Design Trends with Product Data
Not on the map. Why?
Abstract (arXiv)
Moore's Law and Dennard Scaling have guided the semiconductor industry for the past few decades. Recently, both laws have faced validity challenges as transistor sizes approach the practical limits of physics. We are interested in testing the validity of these laws and reflect on the reasons responsible. In this work, we collect data of more than 4000 publicly-available CPU and GPU products. We find that transistor scaling remains critical in keeping the laws valid. However, architectural solutions have become increasingly important and will play a larger role in the future. We observe that GPUs consistently deliver higher performance than CPUs. GPU performance continues to rise because of increases in GPU frequency, improvements in the thermal design power (TDP), and growth in die size. But we also see the ratio of GPU to CPU performance moving closer to parity, thanks to new SIMD extensions on CPUs and increased CPU core counts.
Key facts. Cited by 10 papers in the map (data through 7 Oct 2026). First posted on arXiv 26 Nov 2019.
Builds on
No reference list yet.
Built on by
Citing papers per year: 2020: 1; 2021: 4; 2022: 0; 2023: 2; 2024: 2; 2025: 0; 2026: 1.
10 papers.
- 2020GShard: Scaling Giant Models with Conditional Computation and Automatic Shardingcited by 1,601
- 2021Deep physical neural networks enabled by a backpropagation algorithm for arbitrary physical systemscited by 125
- 2023Reversible Vision Transformerscited by 33
- 2021NetKet 3: Machine Learning Toolbox for Many-Body Quantum Systemscited by 27
- 2024NoMAD-Attention: Efficient LLM Inference on CPUs Through Multiply-add-free Attentioncited by 12
- 2024Beyond Efficiency: Scaling AI Sustainablycited by 8
- 2021A Survey of Large-Scale Deep Learning Serving System Optimization: Challenges and Opportunitiescited by 5
- 2021EnergyVis: Interactively Tracking and Exploring Energy Consumption for ML Modelscited by 4
- 2023ISimDL: Importance Sampling-Driven Acceleration of Fault Injection Simulations for Evaluating the Robustness of Deep Learningcited by 1
- 2026Democratizing AI: A Comparative Study in Deep Learning Efficiency and Future Trends in Computational Processingcited by 0
- 2026Democratizing AI: A Comparative Study in Deep Learning Efficiency and Future Trends in Computational Processing21 Mar 2026
- 2024Beyond Efficiency: Scaling AI Sustainably8 Jun 2024
- 2024NoMAD-Attention: Efficient LLM Inference on CPUs Through Multiply-add-free Attention2 Mar 2024
- 2023ISimDL: Importance Sampling-Driven Acceleration of Fault Injection Simulations for Evaluating the Robustness of Deep Learning14 Mar 2023
- 2023Reversible Vision Transformers9 Feb 2023
- 2021NetKet 3: Machine Learning Toolbox for Many-Body Quantum Systems20 Dec 2021
- 2021A Survey of Large-Scale Deep Learning Serving System Optimization: Challenges and Opportunities28 Nov 2021
- 2021Deep physical neural networks enabled by a backpropagation algorithm for arbitrary physical systems27 Apr 2021
- 2021EnergyVis: Interactively Tracking and Exploring Energy Consumption for ML Models30 Mar 2021
- 2020GShard: Scaling Giant Models with Conditional Computation and Automatic Sharding30 Jun 2020