{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T08:48:02Z","timestamp":1782809282002,"version":"3.54.5"},"reference-count":0,"publisher":"ECMS","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,6,23]]},"abstract":"<jats:p>The microkernel of dense matrix blocks multiplication is considered in relation to direct supernodal solvers designed to solve sparse symmetric systems of linear algebraic equations using the Cholesky method. The method is intended for computers with processors that support AVX512 instructions. The reason for creating this approach was that the Intel MKL library does not always utilize the capabilities of the latest AMD processor generations correctly, resulting in a significant decrease in the performance of the dgemm matrix multiplication library procedure. To achieve high performance, the proposed microkernel utilizes blocking the processor cache, blocking ZMM registers, and repacking elements of multiplied blocks to enable the use of fast loading of these elements into ZMM registers. This approach minimizes processor pipeline delays and leverages OpenMP-based multithreading. As a result, the developed microkernel enabled a significant performance increase in matrix block multiplication on the aforementioned AMD processors, while maintaining the ability to effectively utilize other MKL procedures. At the same time, this microkernel on Intel processors for desktops, laptops, and workstations with shared memory delivers performance only slightly inferior to the dgemm procedure from the Intel MKL library, which is optimized for these processors as much as possible.<\/jats:p>","DOI":"10.7148\/2026-0215","type":"proceedings-article","created":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T08:36:46Z","timestamp":1782808606000},"page":"215-220","source":"Crossref","is-referenced-by-count":0,"title":["Application of avx512 technique to the finite element solver parfes"],"prefix":"10.7148","author":[{"given":"Sergiy","family":"Fialko","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"4144","published-online":{"date-parts":[[2026,6,23]]},"event":{"name":"40th ECMS International Conference on Modelling and Simulation"},"container-title":["ECMS 2026 Proceedings edited by Filippo Sanfilippo, Florenc Demrozi, Fabio Sgarbossa, Mohammad Poursina"],"original-title":[],"deposited":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T08:36:54Z","timestamp":1782808614000},"score":1,"resource":{"primary":{"URL":"http:\/\/www.scs-europe.net\/dlib\/2026\/ecms2026acceptedpapers\/0215_fdem_ecms2026_0006.pdf"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,23]]},"references-count":0,"URL":"https:\/\/doi.org\/10.7148\/2026-0215","relation":{},"subject":[],"published":{"date-parts":[[2026,6,23]]}}}