V.A. Solovyov1, E.I. Azimov2, M.N. Yldashev3

1–3 Bauman Moscow State Technical University (Moscow, Russia)

This article reveals the optimization problem associated with excessive heat dissipation of the cores of modern processors. Energy consumption is one of the key problems of modern computing. The ability to develop energy-efficient processors is crucial, as the use of data and its processing is constantly expanding in all spheres of society. The need for energy-efficient and fast computing is due to the growing demand for high-speed processing, which is extremely necessary in processing large amounts of data. In particular, in research, production, economic tasks with the use of technical analysis, modeling, big data processing, in the work of artificial intelligence systems, high-precision medicine.

The target is to find the best solutions to this problem, the latest generations of flagship processors of leading companies such as Intel Alder Lake, AMD Ryzen 9 5950X for personal computers, Qualcomm Snapdragon 888, Apple A15 Bionic for mobile systems are analyzed and reviewed. Their architecture, main functions and technologies that improve performance and power consumption are discussed in detail.

Based on the analysis, several ways to improve the performance and energy efficiency of processors with CISC (x86) and RISC (ARM) architectures are proposed: improving the technical process, architecture, application of coprocessors and software technologies.

The main classical and innovative directions of development of modern processors were identified. Further improvement of these methods offers a possible way to further increase the efficiency of calculations, for example, more efficient use of parallel operations, the creation of new scheduling algorithms.

Technological and physical barriers, as a rule, impose restrictions on how fast ALUs can process data in a processor using classical technologies. The proposed methods and recommendations are designed to help developers overcome these barriers.

Soloviev V.A., Azimov E.I, Yldashev M.N. Modern technologies of increasing the computing power of processors. Neurocomputers. 2022.
V. 24. № 1. Р. 53-64. DOI: https://doi.org/10.18127/j19998554-202201-05 (In Russian).

1. Rupp K. 48 Years of Microprocessor Trend Data [Elektronnyj resurs] – Rezhim dostupa: https://github.com/karlrupp/ microprocessor-trend-data
2. Alahmad M., Abd El-Aleem F. Heat Transfer Challenges in Semiconductors Processing and the Applications of Heat Pipes for Efficient Heat Removal. Journal of King Saud University – Engineering Sciences. 2003. V. 15. Iss. 1. P. 141–154. DOI 10.1016/S1018-3639(18)30767-0
3. Krajewski M. Hands-On High Performance Programming with Qt 5. Pact. 2019.
4. Tanenbaum E., Ostin T. Arkhitektura kompyutera. 6-e izd. SPb.: Piter. 2013. 816 s. (in Russian).
5. Reinders J., Jeffers J. High Performance Parallelism Pearls. 2016.
6. Presentation of «Zen 3» Microarchitecture Details [Elektronnyj resurs] – Rezhim dostupa: https://www.slideshare.net/AMD/amd-where-gaming-begins-239086719
7. Intel Architecture Day 2021 Presentation [Elektronnyj resurs] – Rezhim dostupa: https://download.intel.com/newsroom/ 2021/client-computing/intel-architecture-day-2021-presentation.pdf
8. Charles J., Jassi P., Ananth N.S., Sadat A., Fedorova A. Evaluation of the Intel® Core™ i7 Turbo Boost feature. 2009 IEEE International Symposium on Workload Characterization (IISWC). 2009. P. 188–197. DOI: 10.1109/IISWC.2009.5306782
9. Saini S., Jin H., Hood R.T., Barker D., Mehrotra P., Biswas R. The impact of hyper-threading on processor resource utilization in production applications. 18th International Conference on High Performance Computing. 2011. P. 1–10.
10. Wolf D.F., Romero R.A.F., Marques E. Using embedded processors in hardware models of artificial neural networks. 2001.
11. Gladkikh A.A., Komakhin M.O., Simankov A.V., Uzenkov D.A. Ispolzovaniye rekurrentnykh neyronnykh setey dlya veroyatnostnoy klassifikatsii arkhitektury mikroprotsessora ispolnyayemykh faylov. Neyrokompyutery: razrabotka. prime-neniye. 2021. T. 23. № 1.
    S. 43–49. (in Russian).
12. Crupi G., Schreurs D., Raskin J.-P., Caddemi A. A comprehensive review on microwave FinFET modeling for progressing beyond the state of art. Solid State Electronics. 2013. V. 80. P. 81–95. DOI: 10.1016/j.sse.2012.10.015
13. Parkale Y.V., Deokate B.H., Divekar S.N., Kadbe P.K. ARMCortex-M3: An improvement overview. 2011.
14. Gladkikh A.A., Vlasov A.I., Uzenkov D.A., Fatkhutdinov T.M. Dinamicheskaya modifikatsiya vnutrennego programmnogo obes-pecheniya vstraivayemykh ustroystv dlya resheniya zadach obratnoy razrabotki. Problemy razrabotki perspektivnykh mikro- i nanoelektronnykh sistem (MES). 2021. № 2. S. 63–69. (in Russian).
15. Shakhnov V.A., Vlasov A.I., Polyakov Yu.A., Kuznetsov A.S. Neyrokompyutery: arkhitektura i skhemotekhnika. M.: Izd-vo Ma-shinostroyeniye. 2000. 64 s. (in Russian).
16. Vlasov A.I. Apparatnaya realizatsiya neyrovychislitelnykh upravlyayushchikh sistem. Pribory i sistemy. Upravleniye. kon-trol. diagnostika. 1999. № 2. S. 61–65. (in Russian).
17. Svidetelstvo o registratsii programmy dlya EVM RU № 2018614711. Ekspertnaya sistema po klassifikatsii defektov topologii kristallov integralnykh mikroskhem. A.Yu. Viryasova. A.I. Vlasov. A.A. Gladkikh. E.V. Rezchikova (in Russian).
18. Aristov R.S., Vlasov A.I., Viryasova A.Yu., Gladkikh A.A., Makarchuk V.V. Issledovaniye razlichnykh modeley svertochnykh neyronnykh setey dlya klassifikatsii izobrazheniy defektov topologicheskogo risunka SBIS. Nanoindustriya. 2018. № S (82). S. 392–398. (in Russian).
19. Amirkhanov A.V., Aristov R.S., Vlasov A.I., Gladkikh A.A., Zakharchenko A., Makarchuk V.V. Ispolzovaniye sovremennykh mode-ley iskusstvennykh neyronnykh setey dlya analiza defektov pri izgotovlenii submikronnykh SBIS. Nanoindustriya. 2017. № S (74). S. 395–406. (in Russian).
20. Snapdragon 888 – Opening Presentation [Elektronnyj resurs] – Rezhim dostupa: https://www.qualcomm.com/media/documents/
    files/snapdragon-888-opening-presentation.pdf