{"id":2133,"date":"2021-01-18T19:06:21","date_gmt":"2021-01-18T17:06:21","guid":{"rendered":"http:\/\/journals.khnu.km.ua\/vestnik\/?p=2133"},"modified":"2021-04-01T11:54:34","modified_gmt":"2021-04-01T08:54:34","slug":"%d0%bf%d1%80%d0%be%d0%b8%d0%b7%d0%b2%d0%be%d0%b4%d0%b8%d1%82%d0%b5%d0%bb%d1%8c%d0%bd%d0%be%d1%81%d1%82%d1%8c-gpu-%d0%b8-cpu-%d0%b4%d0%bb%d1%8f-%d0%bc%d0%b0%d1%82%d1%80%d0%b8%d1%87%d0%bd%d0%be%d0%b3","status":"publish","type":"post","link":"https:\/\/journals.khnu.km.ua\/vestnik\/?p=2133","title":{"rendered":"\u041f\u0440\u043e\u0438\u0437\u0432\u043e\u0434\u0438\u0442\u0435\u043b\u044c\u043d\u043e\u0441\u0442\u044c gpu \u0438 cpu \u0434\u043b\u044f \u043c\u0430\u0442\u0440\u0438\u0447\u043d\u043e\u0433\u043e \u0443\u043c\u043d\u043e\u0436\u0435\u043d\u0438\u044f"},"content":{"rendered":"

\u041f\u0420\u041e\u0418\u0417\u0412\u041e\u0414\u0418\u0422\u0415\u041b\u042c\u041d\u041e\u0421\u0422\u042c GPU \u0418 CPU \u0414\u041b\u042f \u041c\u0410\u0422\u0420\u0418\u0427\u041d\u041e\u0413\u041e \u0423\u041c\u041d\u041e\u0416\u0415\u041d\u0418\u042f<\/p>\n

GPU AND CPU PERFORMANCE FOR MATRIX MULTIPLICATION<\/p>\n

<\/a> \u0421\u0442\u043e\u0440\u0456\u043d\u043a\u0438: 102-110. \u041d\u043e\u043c\u0435\u0440: \u21162, 2019 (271)<\/strong>
\n\u0410\u0432\u0442\u043e\u0440\u0438:<\/strong>
\n\u0410. \u0410. \u041c\u042f\u0421\u0418\u0429\u0415\u0412, \u0412. \u041c. \u041f\u041e\u041b\u041e\u0417\u041e\u0412\u0410
\n\u0425\u043c\u0435\u043b\u044c\u043d\u0438\u0446\u043a\u0438\u0439 \u043d\u0430\u0446\u0438\u043e\u043d\u0430\u043b\u044c\u043d\u044b\u0439 \u0443\u043d\u0438\u0432\u0435\u0440\u0441\u0438\u0442\u0435\u0442
\nA. A. MYASISCHEV, V. M. POLOZOVA
\nKhmelnytskyi National University
\nDOI:<\/strong> https:\/\/www.doi.org\/10.31891\/2307-5732-2019-271-2-102-110<\/a>
\n\u0420\u0435\u0446\u0435\u043d\u0437\u0456\u044f\/Peer review :<\/strong> 02.03.2019 \u0440.
\n\u041d\u0430\u0434\u0440\u0443\u043a\u043e\u0432\u0430\u043d\u0430\/Printed :<\/strong> 10.04.2019 \u0440.<\/p>\n

\u0410\u043d\u043e\u0442\u0430\u0446\u0456\u044f \u043c\u043e\u0432\u043e\u044e \u043e\u0440\u0438\u0433\u0456\u043d\u0430\u043b\u0443<\/strong><\/p>\n

\u0412 \u0440\u0430\u0431\u043e\u0442\u0435 \u0438\u0441\u0441\u043b\u0435\u0434\u0443\u0435\u0442\u0441\u044f \u0446\u0435\u043b\u0435\u0441\u043e\u043e\u0431\u0440\u0430\u0437\u043d\u043e\u0441\u0442\u044c \u043f\u0440\u0438\u043c\u0435\u043d\u0435\u043d\u0438\u044f \u0433\u0440\u0430\u0444\u0438\u0447\u0435\u0441\u043a\u0438\u0445 \u043f\u0440\u043e\u0446\u0435\u0441\u0441\u043e\u0440\u043e\u0432 \u043f\u0440\u0438 \u0440\u0435\u0448\u0435\u043d\u0438\u0438 \u0437\u0430\u0434\u0430\u0447 \u043c\u0430\u0442\u0440\u0438\u0447\u043d\u043e\u0433\u043e \u0443\u043c\u043d\u043e\u0436\u0435\u043d\u0438\u044f \u043f\u043e \u0441\u0440\u0430\u0432\u043d\u0435\u043d\u0438\u044e \u0441 \u043e\u0431\u044b\u0447\u043d\u044b\u043c\u0438 \u043c\u043d\u043e\u0433\u043e\u044f\u0434\u0435\u0440\u043d\u044b\u043c\u0438 \u043f\u0440\u043e\u0446\u0435\u0441\u0441\u043e\u0440\u0430\u043c\u0438. \u0423\u043a\u0430\u0437\u044b\u0432\u0430\u044e\u0442\u0441\u044f \u043e\u0441\u043e\u0431\u0435\u043d\u043d\u043e\u0441\u0442\u0438 \u0438\u0441\u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u043d\u0438\u044f \u0438 \u043f\u0440\u043e\u0431\u043b\u0435\u043c\u044b \u0443\u0441\u0442\u0430\u043d\u043e\u0432\u043a\u0438 \u0431\u0438\u0431\u043b\u0438\u043e\u0442\u0435\u043a\u0438 MAGMA. \u0414\u043b\u044f \u043f\u0440\u043e\u0432\u0435\u0434\u0435\u043d\u0438\u044f \u0432\u044b\u0447\u0438\u0441\u043b\u0438\u0442\u0435\u043b\u044c\u043d\u044b\u0445 \u044d\u043a\u0441\u043f\u0435\u0440\u0438\u043c\u0435\u043d\u0442\u043e\u0432 \u0440\u0430\u0441\u0441\u043c\u043e\u0442\u0440\u0435\u043d\u044b \u0434\u0432\u0435 \u0441\u0438\u0441\u0442\u0435\u043c\u044b. \u0412 \u043a\u0430\u0436\u0434\u043e\u0439 \u0438\u0437 \u043d\u0438\u0445 \u0443\u0441\u0442\u0430\u043d\u043e\u0432\u043b\u0435\u043d \u0448\u0435\u0441\u0442\u0438\u044f\u0434\u0435\u0440\u043d\u044b\u0439 \u043f\u0440\u043e\u0446\u0435\u0441\u0441\u043e\u0440 (CPU)\u00a0 AMD. \u0412 \u043f\u0435\u0440\u0432\u043e\u0439 \u0441\u0438\u0441\u0442\u0435\u043c\u0435 \u0438\u0441\u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u043d \u0433\u0440\u0430\u0444\u0438\u0447\u0435\u0441\u043a\u0438\u0439 \u043f\u0440\u043e\u0446\u0435\u0441\u0441\u043e\u0440 (GPU) Tesla C2075, \u0432\u043e \u0432\u0442\u043e\u0440\u043e\u0439 \u2013 GeForce GTX 480 \u0444\u0438\u0440\u043c\u044b \u00abNVIDIA\u00bb. GPU \u0432\u044b\u043f\u043e\u043b\u043d\u044f\u044e\u0442 \u0440\u043e\u043b\u044c \u0432\u044b\u0447\u0438\u0441\u043b\u0438\u0442\u0435\u043b\u044c\u043d\u044b\u0445 \u0443\u0441\u043a\u043e\u0440\u0438\u0442\u0435\u043b\u0435\u0439 \u0434\u043b\u044f \u0440\u0435\u0448\u0435\u043d\u0438\u044f \u0437\u0430\u0434\u0430\u0447 \u043c\u0430\u0442\u0440\u0438\u0447\u043d\u043e\u0433\u043e \u0443\u043c\u043d\u043e\u0436\u0435\u043d\u0438\u044f. \u041f\u0440\u0438\u0447\u0435\u043c \u0432 \u043f\u0435\u0440\u0432\u043e\u043c \u0441\u043b\u0443\u0447\u0430\u0435 \u0440\u0430\u0441\u0447\u0435\u0442 \u0432\u044b\u043f\u043e\u043b\u043d\u044f\u0435\u0442\u0441\u044f \u0441 \u0443\u0447\u0435\u0442\u043e\u043c \u0440\u0430\u0441\u043f\u0430\u0440\u0430\u043b\u043b\u0435\u043b\u0438\u0432\u0430\u043d\u0438\u044f \u043f\u043e 6 \u044f\u0434\u0440\u0430\u043c \u043f\u0440\u043e\u0446\u0435\u0441\u0441\u043e\u0440\u0430 \u0441 \u0438\u0441\u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u043d\u0438\u0435\u043c \u0431\u0438\u0431\u043b\u0438\u043e\u0442\u0435\u043a MPI, ScaLAPACK \u0438 ATLAS. \u0412\u043e \u0432\u0442\u043e\u0440\u043e\u043c \u0438 \u0442\u0440\u0435\u0442\u044c\u0435\u043c \u0441\u043b\u0443\u0447\u0430\u044f\u0445 \u2013 \u0440\u0430\u0441\u043f\u0430\u0440\u0430\u043b\u043b\u0435\u043b\u0438\u0432\u0430\u043d\u0438\u0435\u043c \u043f\u043e \u044f\u0434\u0440\u0430\u043c GPU Tesla C2075 \u0438 GeForce GTX 480 \u0441 \u0438\u0441\u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u043d\u0438\u0435\u043c \u0442\u0435\u0445\u043d\u043e\u043b\u043e\u0433\u0438\u0438 CUDA. \u0412\u044b\u0447\u0438\u0441\u043b\u0438\u0442\u0435\u043b\u044c\u043d\u044b\u0435 \u0441\u0438\u0441\u0442\u0435\u043c\u044b \u0440\u0430\u0431\u043e\u0442\u0430\u044e\u0442 \u043f\u043e\u0434 \u0443\u043f\u0440\u0430\u0432\u043b\u0435\u043d\u0438\u0435\u043c \u043e\u043f\u0435\u0440\u0430\u0446\u0438\u043e\u043d\u043d\u043e\u0439 \u0441\u0438\u0441\u0442\u0435\u043c\u044b Linux Ubuntu. \u041d\u0430 \u043d\u0438\u0445 \u0443\u0441\u0442\u0430\u043d\u043e\u0432\u043b\u0435\u043d\u044b \u043a\u043e\u043c\u043f\u0438\u043b\u044f\u0442\u043e\u0440\u044b \u0444\u043e\u0440\u0442\u0440\u0430\u043d \u0438 C++ \u0441 \u043f\u0435\u0440\u0435\u0447\u0438\u0441\u043b\u0435\u043d\u043d\u044b\u043c\u0438 \u0432\u044b\u0448\u0435 \u0431\u0438\u0431\u043b\u0438\u043e\u0442\u0435\u043a\u0430\u043c\u0438 \u0434\u043b\u044f 6-\u044f\u0434\u0435\u0440\u043d\u043e\u0433\u043e \u043f\u0440\u043e\u0446\u0435\u0441\u0441\u043e\u0440\u0430. \u0414\u043b\u044f \u043f\u0440\u043e\u0433\u0440\u0430\u043c\u043c\u0438\u0440\u043e\u0432\u0430\u043d\u0438\u044f \u043d\u0430 GPU Tesla C2075 \u0438 GeForce GTX 480 \u0438\u043d\u0441\u0442\u0430\u043b\u043b\u0438\u0440\u043e\u0432\u0430\u043d\u044b \u0432\u0438\u0434\u0435\u043e\u0434\u0440\u0430\u0439\u0432\u0435\u0440 nvidia \u0438 \u043f\u0440\u043e\u0433\u0440\u0430\u043c\u043c\u043d\u043e\u0435 \u043e\u0431\u0435\u0441\u043f\u0435\u0447\u0435\u043d\u0438\u0435 CUDA Toolkit. \u0423\u0441\u0442\u0430\u043d\u043e\u0432\u043b\u0435\u043d\u043e, \u0447\u0442\u043e \u043f\u0440\u043e\u0438\u0437\u0432\u043e\u0434\u0438\u0442\u0435\u043b\u044c\u043d\u043e\u0441\u0442\u044c GPU GeForce GTX 480 \u0438 GPU Tesla C2075 \u0432\u044b\u0448\u0435 \u043f\u0440\u043e\u0438\u0437\u0432\u043e\u0434\u0438\u0442\u0435\u043b\u044c\u043d\u043e\u0441\u0442\u0438 CPU AMD \u043f\u0440\u0438\u043c\u0435\u0440\u043d\u043e \u0432 3.5 \u0438 6.3 \u0440\u0430\u0437 \u0441\u043e\u043e\u0442\u0432\u0435\u0442\u0441\u0442\u0432\u0435\u043d\u043d\u043e \u0434\u043b\u044f \u0447\u0438\u0441\u0435\u043b \u0441 \u0434\u0432\u043e\u0439\u043d\u043e\u0439 \u0442\u043e\u0447\u043d\u043e\u0441\u0442\u044c\u044e. \u0410 \u043f\u0440\u043e\u0438\u0437\u0432\u043e\u0434\u0438\u0442\u0435\u043b\u044c\u043d\u043e\u0441\u0442\u044c GPU GeForce GTX 480 \u0432 1.3 \u0440\u0430\u0437\u0430 \u0432\u044b\u0448\u0435 \u043f\u0440\u043e\u0438\u0437\u0432\u043e\u0434\u0438\u0442\u0435\u043b\u044c\u043d\u043e\u0441\u0442\u0438 GPU Tesla C2075 \u0434\u043b\u044f \u0447\u0438\u0441\u0435\u043b \u0441 \u043e\u0434\u0438\u043d\u0430\u0440\u043d\u043e\u0439 \u0442\u043e\u0447\u043d\u043e\u0441\u0442\u044c\u044e. \u041f\u043e\u043a\u0430\u0437\u0430\u043d\u043e, \u0447\u0442\u043e \u0434\u043b\u044f \u0434\u043e\u0441\u0442\u0438\u0436\u0435\u043d\u0438\u044f \u043c\u0430\u043a\u0441\u0438\u043c\u0430\u043b\u044c\u043d\u043e\u0439 \u043f\u0440\u043e\u0438\u0437\u0432\u043e\u0434\u0438\u0442\u0435\u043b\u044c\u043d\u043e\u0441\u0442\u0438 GPU NVIDIA CUDA \u043d\u0435\u043e\u0431\u0445\u043e\u0434\u0438\u043c\u043e \u0438\u0441\u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u043d\u0438\u0435 \u0431\u0438\u0431\u043b\u0438\u043e\u0442\u0435\u043a MAGMA \u0438\u043b\u0438 CUBLAS, \u043a\u043e\u0442\u043e\u0440\u044b\u0435 \u0434\u0430\u044e\u0442 \u0443\u0441\u043a\u043e\u0440\u0435\u043d\u0438\u0435 \u0440\u0430\u0441\u0447\u0435\u0442\u043e\u0432 \u043f\u0440\u0438\u043c\u0435\u0440\u043d\u043e \u0432 6.4 \u0440\u0430\u0437\u0430 \u043f\u043e \u0441\u0440\u0430\u0432\u043d\u0435\u043d\u0438\u044e \u0441 \u0442\u0440\u0430\u0434\u0438\u0446\u0438\u043e\u043d\u043d\u044b\u043c \u0441\u043f\u043e\u0441\u043e\u0431\u043e\u043c \u043f\u0440\u043e\u0433\u0440\u0430\u043c\u043c\u0438\u0440\u043e\u0432\u0430\u043d\u0438\u044f \u0441 \u0438\u0441\u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u043d\u0438\u0435\u043c \u0433\u043b\u043e\u0431\u0430\u043b\u044c\u043d\u043e\u0439 \u043f\u0430\u043c\u044f\u0442\u0438. \u0420\u0430\u0441\u0441\u043c\u043e\u0442\u0440\u0435\u043d\u043e \u0440\u0435\u0448\u0435\u043d\u0438\u0435 \u0437\u0430\u0434\u0430\u0447\u0438 \u043c\u0430\u0442\u0440\u0438\u0447\u043d\u043e\u0433\u043e \u0443\u043c\u043d\u043e\u0436\u0435\u043d\u0438\u044f \u0441 \u0438\u0441\u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u043d\u0438\u0435\u043c \u0440\u0430\u0437\u0434\u0435\u043b\u044f\u0435\u043c\u043e\u0439 \u043f\u0430\u043c\u044f\u0442\u0438. \u041f\u043e\u043a\u0430\u0437\u0430\u043d\u043e, \u0447\u0442\u043e \u0432 \u044d\u0442\u043e\u043c \u0441\u043b\u0443\u0447\u0430\u0435 \u043f\u0440\u043e\u0438\u0437\u0432\u043e\u0434\u0438\u0442\u0435\u043b\u044c\u043d\u043e\u0441\u0442\u044c GPU \u043d\u0435\u0437\u043d\u0430\u0447\u0438\u0442\u0435\u043b\u044c\u043d\u043e \u043d\u0438\u0436\u0435 \u043f\u043e \u0441\u0440\u0430\u0432\u043d\u0435\u043d\u0438\u044e \u0441 \u0438\u0441\u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u043d\u0438\u0435\u043c \u0431\u0438\u0431\u043b\u0438\u043e\u0442\u0435\u043a MAGMA. \u041f\u043e\u043a\u0430\u0437\u0430\u043d\u043e, \u0447\u0442\u043e \u043f\u0440\u043e\u0438\u0437\u0432\u043e\u0434\u0438\u0442\u0435\u043b\u044c\u043d\u043e\u0441\u0442\u044c GPU \u043f\u0440\u0438 \u043c\u0430\u0442\u0440\u0438\u0447\u043d\u043e\u043c \u0443\u043c\u043d\u043e\u0436\u0435\u043d\u0438\u0438 \u0441 \u0438\u0441\u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u043d\u0438\u0435\u043c CUBLAS \u0434\u043b\u044f GeForce GTX 480 \u043f\u0440\u0430\u043a\u0442\u0438\u0447\u0435\u0441\u043a\u0438 \u0440\u0430\u0432\u043d\u0430 \u043f\u0438\u043a\u043e\u0432\u043e\u0439 \u043f\u0440\u043e\u0438\u0437\u0432\u043e\u0434\u0438\u0442\u0435\u043b\u044c\u043d\u043e\u0441\u0442\u0438 \u043f\u0440\u0438 \u0434\u0432\u043e\u0439\u043d\u043e\u0439 \u0442\u043e\u0447\u043d\u043e\u0441\u0442\u0438 \u0432\u044b\u0447\u0438\u0441\u043b\u0435\u043d\u0438\u0439 (\u043f\u0438\u043a\u043e\u0432\u0430\u044f – 168.1, \u043f\u043e\u043b\u0443\u0447\u0435\u043d\u043d\u0430\u044f 165.1 \u0413\u0438\u0433\u0430\u0444\u043b\u043e\u043f\u0441). \u0414\u043b\u044f GPU Tesla C2075 \u043f\u0438\u043a\u043e\u0432\u0430\u044f \u043f\u0440\u043e\u0438\u0437\u0432\u043e\u0434\u0438\u0442\u0435\u043b\u044c\u043d\u043e\u0441\u0442\u044c \u043f\u043e\u0447\u0442\u0438 \u0432 \u0434\u0432\u0430 \u0440\u0430\u0437\u0430 \u0432\u044b\u0448\u0435 \u043f\u043e\u043b\u0443\u0447\u0435\u043d\u043d\u043e\u0439 (\u043f\u0438\u043a\u043e\u0432\u0430\u044f \u2013 515.2, \u043f\u043e\u043b\u0443\u0447\u0435\u043d\u043d\u0430\u044f 300.6 \u0413\u0438\u0433\u0430\u0444\u043b\u043e\u043f\u0441). \u042d\u0442\u043e \u0443\u043a\u0430\u0437\u044b\u0432\u0430\u0435\u0442 \u043d\u0430 \u043d\u0435\u0434\u043e\u0441\u0442\u0430\u0442\u043e\u0447\u043d\u0443\u044e \u044d\u0444\u0444\u0435\u043a\u0442\u0438\u0432\u043d\u043e\u0441\u0442\u044c \u0431\u0438\u0431\u043b\u0438\u043e\u0442\u0435\u043a\u0438 CUBLAS \u0438 MAGMA \u0434\u043b\u044f GPU Tesla.
\n\u041a\u043b\u044e\u0447\u0435\u0432\u044b\u0435 \u0441\u043b\u043e\u0432\u0430:<\/strong> GPU Tesla C2075, MAGMA, CUBLAS, CUDA, \u0433\u0440\u0430\u0444\u0438\u0447\u0435\u0441\u043a\u0438\u0439 \u043f\u0440\u043e\u0446\u0435\u0441\u0441\u043e\u0440, NVIDIA Tesla V100, SMP \u0441\u0438\u0441\u0442\u0435\u043c\u0430, MPI, ScaLAPACK, ATLAS<\/p>\n

\u0420\u043e\u0437\u0448\u0438\u0440\u0435\u043d\u0430 \u0430\u043d\u043e\u0442\u0430\u0446\u0456\u044f \u0430\u043d\u0433\u043b\u0456\u0439\u0441\u044c\u043a\u043e\u044e \u043c\u043e\u0432\u043e\u044e<\/strong><\/p>\n

The paper investigates the expediency of using graphics processors when solving problems of matrix multiplication compared to conventional multi-core processors. Indicates the features of the use and installation problems of the library MAGMA. To carry out computational experiments, two systems were considered. Each of them has a six-core AMD processor. The first system uses a graphics processor (GPU) Tesla C2075, the second GeForce GTX 480 from NVIDIA. GPUs play the role of computational accelerators for solving problems of matrix multiplication. Moreover, in the first case, the calculation is performed taking into account parallelization across the 6 cores of the processor using the MPI, ScaLAPACK and ATLAS libraries. In the second and third cases – parallelization across the cores of the Tesla C2075 GPU and GeForce GTX 480 using the CUDA technology. Computing systems run under the Linux Ubuntu operating system. The compilers Fortran and C ++ are installed on them with the libraries listed above for a 6-core processor. For programming on the Tesla C2075 GPU and GeForce GTX 480, the nvidia video driver and CUDA Toolkit software are installed. It has been established that the performance of the GPU GeForce GTX 480 and the GPU Tesla C2075 is higher than the performance of the AMD CPU by approximately 3.5 and 6.3 times, respectively, for double-precision numbers. And the performance of the GPU GeForce GTX 480 is 1.3 times higher than the performance of the GPU Tesla C2075 for numbers with single precision. It is shown that in order to achieve maximum performance of the NVIDIA CUDA GPU, it is necessary to use the MAGMA or CUBLAS libraries, which accelerate the calculations by about 6.4 times compared to the traditional programming method using global memory. The solution of the problem of matrix multiplication using shared memory is considered. It is shown that in this case the performance of the GPU is not significantly lower compared with the use of the MAGMA libraries. It is shown that GPU performance with matrix multiplication using CUBLAS for GeForce GTX 480 is almost equal to peak performance with double computational accuracy (peak – 168.1, obtained 165.1 Gigaflops). For the Tesla C2075 GPU, the peak performance is almost twice as high as obtained (peak – 515.2, obtained 300.6 Gigaflops). This indicates a lack of effectiveness of the CUBLAS and MAGMA libraries for the Tesla GPU.
\nKeywords:<\/strong> GPU Tesla C2075, MAGMA, CUBLAS, CUDA, graphics processor, NVIDIA Tesla V100, SMP system, MPI, ScaLAPACK, ATLAS.<\/p>\n

References<\/strong><\/p>\n

    \n
  1. The ScaLAPACK Project. 2017 [Jelektronnyj resurs]. \u2013 Rezhim dostupa : http:\/\/www.netlib.org<\/li>\n
  2. Automatically Tuned Linear Algebra Software (ATLAS). 2016 [Jelektronnyj resurs]. \u2013 Rezhim dostupa : http:\/\/math-atlas.sourceforge.net\/<\/li>\n
  3. NVIDIA TESLA V100 TENSOR CORE GPU. 2018 [Jelektronnyj resurs]. \u2013 Rezhim dostupa : https:\/\/www.nvidia.com\/en-us\/data-center\/tesla-v100\/<\/li>\n
  4. Antonov A.S. Parallel’noe programmirovanie s ispol’zovaniem tehnologii MPI : uchebnoe posobie \/ Antonov A.S. \u2013 M. : Izd-vo MGU, 2004. \u2013 71 s.<\/li>\n
  5. Mjasishhev A.A. Dostizhenie naibol’shej proizvoditel’nosti peremnozhenija matric na sistemah s mnogojadernymi processorami. T. 3: Teor\u0456ja ta metodika navchannja \u0456nformatiki \/ Mjasishhev A.A. \u2013 Krivij R\u0456g : Vidavnichij v\u0456dd\u0456l NmetAU, 2010. \u2013 303 s.<\/li>\n
  6. Mjas\u0456shhev O.A. Oc\u0456nka produktivnost\u0456 GPU NVIDIA CUDA pri vir\u0456shenn\u0456 zadach matrichnogo mnozhennja \/ O.A. Mjas\u0456shhev \/\/ Vim\u0456rjuval’na ta obchisljuval’na tehn\u0456ka v tehnolog\u0456chnih procesah. \u2013 Hmel’nic’kij, 2012. \u2013 \u2116 1. \u2013 S. 73\u201379.<\/li>\n
  7. Parallel’nye vychislenija na GPU. Arhitektura i programmnaja model’ GPU : ucheb. posobie \/ A.V. Boreskov i dr. ; predisl.: V.A. Sadovnichij. \u2013 M. : Izdatel’stvo Moskovskogo universiteta, 2012. \u2013 336 s., ill. \u2013 (Serija “Superkomp’juternoe obrazovanie”).<\/li>\n
  8. Gergel’ V.P. Teorija i praktika parallel’nyh vychislenij. 2016 [Jelektronnyj resurs]. \u2013 Rezhim dostupa : https:\/\/www.twirpx.com\/file\/1978282\/<\/li>\n
  9. Parallel’nye vychislenija na GPU. Arhitektura i programmnaja model’ CUDA : ucheb. posobie \/ Boreskov A.V., Harlamov A.A., Markovskij N.D. i dr. \u2013 M. : Izdatel’stvo Moskovskogo universiteta, 2012. \u2013 336 s.<\/li>\n
  10. Matrix Algebra on GPU and Multicore Architectures. 2018 [Jelektronnyj resurs]. \u2013 Rezhim dostupa : http:\/\/icl.cs.utk.edu\/magma\/index.html<\/li>\n
  11. CUBLAS (NVIDIA CUDA Basic Linear Algebra Subroutines). 2018 [Jelektronnyj resurs]. \u2013 Rezhim dostupa : http:\/\/developer.nvidia.com\/cublas<\/li>\n
  12. Ustanovka CUDA Toolkit i drajvera NVIDIA dlja razrabotchikov. 2012 [Jelektronnyj resurs]. \u2013 Rezhim dostupa : http:\/\/forum.ubuntu.ru\/index.php?topic=114802.0<\/li>\n
  13. Mjasishhev A.A. Vychislitel’nye vozmozhnosti mikrokontrollerov STM32F4\/ A.A. Mjasishhev, S.V. Lenkov \/\/ Zb\u0456rnik naukovih prac’ V\u0456js’kovogo \u0456nstitutu Ki\u0457vs’kogo nac\u0456onal’nogo un\u0456versitetu \u0456meny Tarasa Shevchenka. \u2013 Ki\u0457v, 2016. \u2013 Vipusk \u2116 51. \u2013 S. 185\u2013192.<\/li>\n
  14. Top 500 the list. 2018 [Jelektronnyj resurs]. \u2013 Rezhim dostupa : https:\/\/www.top500.org\/lists\/2018\/06\/<\/li>\n
  15. NVIDIA TURING. 2018 [Jelektronnyj resurs]. \u2013 Rezhim dostupa : https:\/\/www.nvidia.com\/ru-ru\/geforce\/turing\/<\/li>\n
  16. Myasischev A.A. Computing capabilities stm32f429i-disco for matrix multiplication \/ A.A. Myasischev \/\/ Materials of the XI International scientific and practical conference. \u2013 2015. \u2013 Vol 17. Sheffild. \u2013 S. 51\u201357.<\/li>\n
  17. Mjas\u0456shhev O.A. Efektivn\u0456st’ vikoristannja GPU NVIDIA pri vir\u0456shenn\u0456 sistem l\u0456n\u0456jnih r\u0456vnjan’ \/ O.A. Mjas\u0456shhev \/\/ Zb. nauk. prac’ V\u0456js’kovogo \u0456nstitutu Ki\u0457vs’kogo NU \u0456m. Taras Shevchenko. \u2013 K. : V\u0406KNU, 2012. \u2013 Vip. 38. \u2013 S. 76\u201380.<\/li>\n
  18. Mjasishhev A.A. Sopostavlenie proizvoditel’nostej GPU i CPU dlja matrichnogo umnozhenija s dvojnoj tochnost’ju \/ A.A. Mjasishhev \/\/ Teor\u0456ja ta metodika navchannja matematiki, f\u0456ziki, \u0456nformatiki : zb\u0456rnik naukovih prac’ : v 3-h t. T. 3: Teor\u0456ja ta metodika navchannja \u0456nformatiki. \u2013 Krivij R\u0456g : Vidavnichij v\u0456dd\u0456l NMetAU, 2012. \u2013 Vipusk X. \u2013 S. 99\u2013108.<\/li>\n
  19. Mjasishhev A.A. Sopostavlenie proizvoditel’nostej GPU TESLA C2075, GEFORCE 480 GTX s 6-jadernym CPU AMD pri reshenii sistem linejnyh uravnenij \/ A.A. Mjasishhev \/\/ Materialy mezhdunar. NPK \u00abKljuchevye problemy sovremennoj nauki \u2013 2012\u00bb. \u2013 Sofija : Bjal GRAD-BG, 2012. \u2013 Tom 30. \u2013 S. 28\u201340.<\/li>\n<\/ol>\n","protected":false},"excerpt":{"rendered":"

    \u041f\u0420\u041e\u0418\u0417\u0412\u041e\u0414\u0418\u0422\u0415\u041b\u042c\u041d\u041e\u0421\u0422\u042c GPU \u0418 CPU \u0414\u041b\u042f \u041c\u0410\u0422\u0420\u0418\u0427\u041d\u041e\u0413\u041e \u0423\u041c\u041d\u041e\u0416\u0415\u041d\u0418\u042f GPU AND CPU PERFORMANCE FOR MATRIX MULTIPLICATION \u0421\u0442\u043e\u0440\u0456\u043d\u043a\u0438: 102-110. \u041d\u043e\u043c\u0435\u0440: \u21162, 2019 (271) \u0410\u0432\u0442\u043e\u0440\u0438: \u0410. \u0410. \u041c\u042f\u0421\u0418\u0429\u0415\u0412, \u0412. \u041c. \u041f\u041e\u041b\u041e\u0417\u041e\u0412\u0410 \u0425\u043c\u0435\u043b\u044c\u043d\u0438\u0446\u043a\u0438\u0439 \u043d\u0430\u0446\u0438\u043e\u043d\u0430\u043b\u044c\u043d\u044b\u0439 \u0443\u043d\u0438\u0432\u0435\u0440\u0441\u0438\u0442\u0435\u0442 A. A. MYASISCHEV, V. M. POLOZOVA Khmelnytskyi National University DOI: https:\/\/www.doi.org\/10.31891\/2307-5732-2019-271-2-102-110 \u0420\u0435\u0446\u0435\u043d\u0437\u0456\u044f\/Peer review : 02.03.2019 \u0440. \u041d\u0430\u0434\u0440\u0443\u043a\u043e\u0432\u0430\u043d\u0430\/Printed : 10.04.2019 \u0440. \u0410\u043d\u043e\u0442\u0430\u0446\u0456\u044f \u043c\u043e\u0432\u043e\u044e \u043e\u0440\u0438\u0433\u0456\u043d\u0430\u043b\u0443 \u0412 \u0440\u0430\u0431\u043e\u0442\u0435 […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[22],"tags":[],"_links":{"self":[{"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=\/wp\/v2\/posts\/2133"}],"collection":[{"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=2133"}],"version-history":[{"count":3,"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=\/wp\/v2\/posts\/2133\/revisions"}],"predecessor-version":[{"id":5293,"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=\/wp\/v2\/posts\/2133\/revisions\/5293"}],"wp:attachment":[{"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=2133"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=2133"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=2133"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}