<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Publishing DTD v1.3 20210610//EN" "JATS-journalpublishing1-3.dtd">
<article article-type="research-article" dtd-version="1.3" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xml:lang="ru"><front><journal-meta><journal-id journal-id-type="publisher-id">trudyniisi</journal-id><journal-title-group><journal-title xml:lang="ru">Труды НИИСИ</journal-title><trans-title-group xml:lang="en"><trans-title>SRISA Proceedings</trans-title></trans-title-group></journal-title-group><issn pub-type="ppub">2225-7349</issn><issn pub-type="epub">3033-6422</issn><publisher><publisher-name>НИЦ «КУРЧАТОВСКИЙ ИНСТИТУТ» - НИИСИ</publisher-name></publisher></journal-meta><article-meta><article-id custom-type="elpub" pub-id-type="custom">trudyniisi-68</article-id><article-categories><subj-group subj-group-type="heading"><subject>Research Article</subject></subj-group><subj-group subj-group-type="section-heading" xml:lang="ru"><subject>ВЫСОКОПРОИЗВОДИТЕЛЬНЫЕ ВЫЧИСЛЕНИЯ НА ГРАФИЧЕСКИХ УСКОРИТЕЛЯХ</subject></subj-group><subj-group subj-group-type="section-heading" xml:lang="en"><subject>HIGH-PERFORMANCE COMPUTING ON GPU</subject></subj-group></article-categories><title-group><article-title>Тенденции в графических ускорителях для высокопроизводительных вычислений</article-title><trans-title-group xml:lang="en"><trans-title>Graphics Accelerators for High-Performance Computing</trans-title></trans-title-group></title-group><contrib-group><contrib contrib-type="author" corresp="yes"><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Шмелёв</surname><given-names>А. С.</given-names></name><name name-style="western" xml:lang="en"><surname>Shmelev</surname><given-names>A. S.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Москва</p></bio><email xlink:type="simple">guest8993@rambler.ru</email><xref ref-type="aff" rid="aff-1"/></contrib></contrib-group><aff-alternatives id="aff-1"><aff xml:lang="ru">МСЦ РАН филиал ФГУ ФНЦ НИИСИ РАН<country>Россия</country></aff></aff-alternatives><pub-date pub-type="collection"><year>2023</year></pub-date><pub-date pub-type="epub"><day>21</day><month>10</month><year>2025</year></pub-date><volume>13</volume><issue>4</issue><fpage>117</fpage><lpage>122</lpage><permissions><copyright-statement>Copyright &amp;#x00A9; Шмелёв А.С., 2025</copyright-statement><copyright-year>2025</copyright-year><copyright-holder xml:lang="ru">Шмелёв А.С.</copyright-holder><copyright-holder xml:lang="en">Shmelev A.S.</copyright-holder><license license-type="creative-commons-attribution" xlink:href="https://creativecommons.org/licenses/by/4.0/" xlink:type="simple"><license-p>This work is licensed under a Creative Commons Attribution 4.0 License.</license-p></license></permissions><self-uri xlink:href="https://www.t-niisi.ru/jour/article/view/68">https://www.t-niisi.ru/jour/article/view/68</self-uri><abstract><p>Графические карты, построенные на основе большого количества простых и однотипных исполнительных устройств и обладающие высокой пиковой производительностью уже давно используются в области высокопроизводительных вычислений в качестве ускорителей вычислений. В настоящее время выпускаются отдельные продукты, ориентированные на применение в вычислительных центрах. В данной работе приводится обзор современных ускорителей для вычислительных центров, приведены их показатели производительности, а также приведены анонсы перспективных ускорителей вычислений и показаны тенденции в данной области.</p></abstract><trans-abstract xml:lang="en"><p>GPUs, built from a large number of relatively simple an similar execution units and having high peak performance, have been used long time in high-performance computing as accelerators. Nowday there are GPUs that were specially designed for use in supercomputer centers, and not as graphics processors.Current trends in their development are presented.</p></trans-abstract><kwd-group xml:lang="ru"><kwd>высокопроизводительные вычисления</kwd><kwd>графические ускорители</kwd><kwd>ускорители вычислений</kwd><kwd>суперЭВМ</kwd><kwd>память с высокой пропускной способностью</kwd></kwd-group><kwd-group xml:lang="en"><kwd>High performance computing (HPC)</kwd><kwd>Graphics processing unit (GPU)</kwd><kwd>Accelerated Processing Unit (APU)</kwd><kwd>High Bandwidth Memory (HBM)</kwd></kwd-group><funding-group xml:lang="ru"><funding-statement>Работа была выполнена в МСЦ РАН в рамках государственного задания по теме FNEF-2022-0016. В исследованиях использовался суперкомпьютер МВС-10П.</funding-statement></funding-group></article-meta></front><back><ref-list><title>References</title><ref id="cit1"><label>1</label><citation-alternatives><mixed-citation xml:lang="ru">Lal Shimpi, Anand; Wilson, Derek (June 16, 2008). "Lots More Compute, a Leetle More Texturing - Nvidia's 1.4 Billion Transistor GPU: GT200 Arrives as the GeForce GTX 280 &amp; 260". Anandtech.com. Retrieved December 11, 2015. https://www.anandtech.com/show/2549/3</mixed-citation><mixed-citation xml:lang="en">Lal Shimpi, Anand; Wilson, Derek (June 16, 2008). "Lots More Compute, a Leetle More Texturing - Nvidia's 1.4 Billion Transistor GPU: GT200 Arrives as the GeForce GTX 280 &amp; 260". Anandtech.com. Retrieved December 11, 2015. https://www.anandtech.com/show/2549/3</mixed-citation></citation-alternatives></ref><ref id="cit2"><label>2</label><citation-alternatives><mixed-citation xml:lang="ru">Radeon R5xx Acceleration Revision 1.1 February 22, 2008 © 2008 Advanced Micro Devices, Inc.</mixed-citation><mixed-citation xml:lang="en">Radeon R5xx Acceleration Revision 1.1 February 22, 2008 © 2008 Advanced Micro Devices, Inc.</mixed-citation></citation-alternatives></ref><ref id="cit3"><label>3</label><citation-alternatives><mixed-citation xml:lang="ru">D.B. Kirk, W.W. Hwu. Programming Massively Parallel Processors. A Hands-on Approach, Morgan Kaufmann Publishers, 2010, 258 p.</mixed-citation><mixed-citation xml:lang="en">D.B. Kirk, W.W. Hwu. Programming Massively Parallel Processors. A Hands-on Approach, Morgan Kaufmann Publishers, 2010, 258 p.</mixed-citation></citation-alternatives></ref><ref id="cit4"><label>4</label><citation-alternatives><mixed-citation xml:lang="ru">HBM: MemorySolutionforBandwidth-HungryProcessors, JoonyoungKimandYounsuKim, SKhynix // Hot Chips 26, August 2014</mixed-citation><mixed-citation xml:lang="en">HBM: MemorySolutionforBandwidth-HungryProcessors, JoonyoungKimandYounsuKim, SKhynix // Hot Chips 26, August 2014</mixed-citation></citation-alternatives></ref><ref id="cit5"><label>5</label><citation-alternatives><mixed-citation xml:lang="ru">JEDEC Updates Groundbreaking High Bandwidth Memory (HBM) Standard (Пресс релиз). JEDEC. 2018-12-17.</mixed-citation><mixed-citation xml:lang="en">JEDEC Updates Groundbreaking High Bandwidth Memory (HBM) Standard (Пресс релиз). JEDEC. 2018-12-17.</mixed-citation></citation-alternatives></ref><ref id="cit6"><label>6</label><citation-alternatives><mixed-citation xml:lang="ru">Intel Data Center GPU Max Series Technical Overview https://www.intel.com/content/www/us/en/developer/articles/technical/intel-data-center-gpu-max-series-overview.html Обращение сентябрь 2023.</mixed-citation><mixed-citation xml:lang="en">Intel Data Center GPU Max Series Technical Overview https://www.intel.com/content/www/us/en/developer/articles/technical/intel-data-center-gpu-max-series-overview.html Обращение сентябрь 2023.</mixed-citation></citation-alternatives></ref><ref id="cit7"><label>7</label><citation-alternatives><mixed-citation xml:lang="ru">Cutress, Ian (August 24, 2021). "Hot Chips 2021 Live Blog: Graphics (Intel, AMD, Google, Xilinx)". AnandTech. Retrieved August 29, 2021.</mixed-citation><mixed-citation xml:lang="en">Cutress, Ian (August 24, 2021). "Hot Chips 2021 Live Blog: Graphics (Intel, AMD, Google, Xilinx)". AnandTech. Retrieved August 29, 2021.</mixed-citation></citation-alternatives></ref><ref id="cit8"><label>8</label><citation-alternatives><mixed-citation xml:lang="ru">Intel® Data Center GPU Max 1550 Specifications https://www.intel.com/content/www/us/en/products/sku/232873/intel-data-center-gpu-max-1550/specifications.html Обращение сентябрь 2023.</mixed-citation><mixed-citation xml:lang="en">Intel® Data Center GPU Max 1550 Specifications https://www.intel.com/content/www/us/en/products/sku/232873/intel-data-center-gpu-max-1550/specifications.html Обращение сентябрь 2023.</mixed-citation></citation-alternatives></ref><ref id="cit9"><label>9</label><citation-alternatives><mixed-citation xml:lang="ru">Argonne installs final components of Aurora supercomputer. Пресс релиз 22.06.2023 Jim Collins https://www.anl.gov/article/argonne-installs-final-components-of-aurora-supercomputer</mixed-citation><mixed-citation xml:lang="en">Argonne installs final components of Aurora supercomputer. Пресс релиз 22.06.2023 Jim Collins https://www.anl.gov/article/argonne-installs-final-components-of-aurora-supercomputer</mixed-citation></citation-alternatives></ref><ref id="cit10"><label>10</label><citation-alternatives><mixed-citation xml:lang="ru">Intel Xeon MAX 9480 Deep-Dive 64GB HBM2e Onboard Like a GPU or AI Accelerator. Patrick Kennedy 12.09.2023 https://www.servethehome.com/intel-xeon-max-9480-deep-dive-intel-has-64gb-hbm2e-onboard-like-a-gpu-or-ai-accelerator/3/</mixed-citation><mixed-citation xml:lang="en">Intel Xeon MAX 9480 Deep-Dive 64GB HBM2e Onboard Like a GPU or AI Accelerator. Patrick Kennedy 12.09.2023 https://www.servethehome.com/intel-xeon-max-9480-deep-dive-intel-has-64gb-hbm2e-onboard-like-a-gpu-or-ai-accelerator/3/</mixed-citation></citation-alternatives></ref><ref id="cit11"><label>11</label><citation-alternatives><mixed-citation xml:lang="ru">INTRODUCING AMD CDNA™ 2 ARCHITECTURE © 2021 Advanced Micro Device https://www.amd.com/system/files/documents/amd-cdna2-white-paper.pdf</mixed-citation><mixed-citation xml:lang="en">INTRODUCING AMD CDNA™ 2 ARCHITECTURE © 2021 Advanced Micro Device https://www.amd.com/system/files/documents/amd-cdna2-white-paper.pdf</mixed-citation></citation-alternatives></ref><ref id="cit12"><label>12</label><citation-alternatives><mixed-citation xml:lang="ru">https://www.top500.org/lists/ Обращение сентябрь 2023.</mixed-citation><mixed-citation xml:lang="en">https://www.top500.org/lists/ Обращение сентябрь 2023.</mixed-citation></citation-alternatives></ref><ref id="cit13"><label>13</label><citation-alternatives><mixed-citation xml:lang="ru">https://top500.org/lists/top500/2023/06/ Обращение сентябрь 2023.</mixed-citation><mixed-citation xml:lang="en">https://top500.org/lists/top500/2023/06/ Обращение сентябрь 2023.</mixed-citation></citation-alternatives></ref><ref id="cit14"><label>14</label><citation-alternatives><mixed-citation xml:lang="ru">https://www.anandtech.com/show/18946/el-capitan-installation-begins-first-apu-exascale-system-shaping-up-for-2024 Обращение сентябрь 2023</mixed-citation><mixed-citation xml:lang="en">https://www.anandtech.com/show/18946/el-capitan-installation-begins-first-apu-exascale-system-shaping-up-for-2024 Обращение сентябрь 2023</mixed-citation></citation-alternatives></ref><ref id="cit15"><label>15</label><citation-alternatives><mixed-citation xml:lang="ru">https://top500.org/lists/green500/2023/06/ Обращение сентябрь 2023.</mixed-citation><mixed-citation xml:lang="en">https://top500.org/lists/green500/2023/06/ Обращение сентябрь 2023.</mixed-citation></citation-alternatives></ref><ref id="cit16"><label>16</label><citation-alternatives><mixed-citation xml:lang="ru">Константин Калгин. Аппаратная архитектура CUDA. Новосибирский национальный исследовательский государственный университет, Новосибирск, 2013. https://ssd.sscc.ru/sites/default/files/content/attach/332/cuda-3-arch.pdf</mixed-citation><mixed-citation xml:lang="en">Константин Калгин. Аппаратная архитектура CUDA. Новосибирский национальный исследовательский государственный университет, Новосибирск, 2013. https://ssd.sscc.ru/sites/default/files/content/attach/332/cuda-3-arch.pdf</mixed-citation></citation-alternatives></ref><ref id="cit17"><label>17</label><citation-alternatives><mixed-citation xml:lang="ru">NVIDIA H100 Tensor Core GPU Architecture whitepaper V1.04 © 2023 NVIDIA Corporation. https://resources.nvidia.com/en-us-tensor-core/gtc22-whitepaper-hopper</mixed-citation><mixed-citation xml:lang="en">NVIDIA H100 Tensor Core GPU Architecture whitepaper V1.04 © 2023 NVIDIA Corporation. https://resources.nvidia.com/en-us-tensor-core/gtc22-whitepaper-hopper</mixed-citation></citation-alternatives></ref><ref id="cit18"><label>18</label><citation-alternatives><mixed-citation xml:lang="ru">NVIDIA H100 Tensor Core GPU Datasheet © 2023 NVIDIA Corporation and affiliates. https://resources.nvidia.com/en-us-tensor-core/nvidia-tensor-core-gpu-datasheet</mixed-citation><mixed-citation xml:lang="en">NVIDIA H100 Tensor Core GPU Datasheet © 2023 NVIDIA Corporation and affiliates. https://resources.nvidia.com/en-us-tensor-core/nvidia-tensor-core-gpu-datasheet</mixed-citation></citation-alternatives></ref><ref id="cit19"><label>19</label><citation-alternatives><mixed-citation xml:lang="ru">NVIDIA Ada Lovelace Professional GPU Architecture Whitepaper v1.1 © 2023 NVIDIA Corporation. https://images.nvidia.com/aem-dam/en-zz/Solutions/technologies/NVIDIA-ADA-GPU-PROVIZ-Architecture-Whitepaper_1.1.pdf</mixed-citation><mixed-citation xml:lang="en">NVIDIA Ada Lovelace Professional GPU Architecture Whitepaper v1.1 © 2023 NVIDIA Corporation. https://images.nvidia.com/aem-dam/en-zz/Solutions/technologies/NVIDIA-ADA-GPU-PROVIZ-Architecture-Whitepaper_1.1.pdf</mixed-citation></citation-alternatives></ref><ref id="cit20"><label>20</label><citation-alternatives><mixed-citation xml:lang="ru">NVIDIA GH200 Grace Hopper Superchip Architecture Whitepaper V1.11 © 2023 NVIDIA Corporation. https://resources.nvidia.com/en-us-grace-cpu/nvidia-grace-hopper</mixed-citation><mixed-citation xml:lang="en">NVIDIA GH200 Grace Hopper Superchip Architecture Whitepaper V1.11 © 2023 NVIDIA Corporation. https://resources.nvidia.com/en-us-grace-cpu/nvidia-grace-hopper</mixed-citation></citation-alternatives></ref><ref id="cit21"><label>21</label><citation-alternatives><mixed-citation xml:lang="ru">NVIDIA Grace Hopper Superchip Data Sheet © 2023 NVIDIA Corporation. https://resources.nvidia.com/en-us-grace-cpu/grace-hopper-superchip</mixed-citation><mixed-citation xml:lang="en">NVIDIA Grace Hopper Superchip Data Sheet © 2023 NVIDIA Corporation. https://resources.nvidia.com/en-us-grace-cpu/grace-hopper-superchip</mixed-citation></citation-alternatives></ref><ref id="cit22"><label>22</label><citation-alternatives><mixed-citation xml:lang="ru">Samsung Develops Industry’s First High Bandwidth Memory with AI Processing Power (Пресс Релиз 17.02.2021) https://news.samsung.com/global/samsung-develops-industrys-first-high-bandwidth-memory-with-ai-processing-power</mixed-citation><mixed-citation xml:lang="en">Samsung Develops Industry’s First High Bandwidth Memory with AI Processing Power (Пресс Релиз 17.02.2021) https://news.samsung.com/global/samsung-develops-industrys-first-high-bandwidth-memory-with-ai-processing-power</mixed-citation></citation-alternatives></ref><ref id="cit23"><label>23</label><citation-alternatives><mixed-citation xml:lang="ru">Samsung Processing in Memory Technology at Hot Chips 2023 By Patrick Kennedy August 28, 2023 https://www.servethehome.com/samsung-processing-in-memory-technology-at-hot-chips-2023/</mixed-citation><mixed-citation xml:lang="en">Samsung Processing in Memory Technology at Hot Chips 2023 By Patrick Kennedy August 28, 2023 https://www.servethehome.com/samsung-processing-in-memory-technology-at-hot-chips-2023/</mixed-citation></citation-alternatives></ref><ref id="cit24"><label>24</label><citation-alternatives><mixed-citation xml:lang="ru">Яхонтов Д.Е., Левченко Н.Н., Окунев А.С. Принципы работы блока специальных операций модуля ассоциативной памяти параллельной потоковой вычислительной системы ППВС // Материалы Международной научно-технической конференции «Суперкомпьютерные технологии: разработка, программирование, применение» (СКТ-2010), Таганрог – Москва, 2010, Т. 1, С. 166-170.</mixed-citation><mixed-citation xml:lang="en">Яхонтов Д.Е., Левченко Н.Н., Окунев А.С. Принципы работы блока специальных операций модуля ассоциативной памяти параллельной потоковой вычислительной системы ППВС // Материалы Международной научно-технической конференции «Суперкомпьютерные технологии: разработка, программирование, применение» (СКТ-2010), Таганрог – Москва, 2010, Т. 1, С. 166-170.</mixed-citation></citation-alternatives></ref></ref-list><fn-group><fn fn-type="conflict"><p>The authors declare that there are no conflicts of interest present.</p></fn></fn-group></back></article>
