{"id":4074,"date":"2018-09-14T14:29:37","date_gmt":"2018-09-14T11:29:37","guid":{"rendered":"https:\/\/umram.bilkent.edu.tr\/?p=4074"},"modified":"2018-09-17T14:30:07","modified_gmt":"2018-09-17T11:30:07","slug":"university-of-illinois-urbana-champainden-mert-hidayetoglu-doktora-calismalarini-sundu","status":"publish","type":"post","link":"https:\/\/umram.bilkent.edu.tr\/index.php\/tr\/2018\/09\/14\/university-of-illinois-urbana-champainden-mert-hidayetoglu-doktora-calismalarini-sundu\/","title":{"rendered":"University of Illinois-Urbana Champain’den\u00a0Mert Hidayeto\u011flu doktora \u00e7al\u0131\u015fmalar\u0131n\u0131 sundu"},"content":{"rendered":"

Mert Hidayetoglu<\/strong> of University of Illinois at Urbana Champain<\/a> presented on his PhD studies.<\/p>\n

Talk was entitled \u201cSUPERCOMPUTING FOR FULL-WAVE TOMOGRAPHIC IMAGE RECONSTRUCTION IN NEAR-REAL TIME.\u201d <\/strong><\/p>\n

The abstract of the talk is:<\/span><\/p>\n

Full-wave methods incorporate all wave phenomena into the image\u00a0reconstructions by solving the Helmholtz equation with no fundamental\u00a0approximation. These phenomena include refraction, absorption,\u00a0diffraction, and multiple-scattering of propagating waves.\u00a0 Although\u00a0full-wave image reconstruction has promising features and can lead to\u00a0new imaging technologies, it has been regarded as impractical due to its\u00a0high computational burden.\u00a0 This talk will be on our effort of making\u00a0full-wave imaging attainable with two approaches:\u00a0 fast algorithms and\u00a0supercomputing.<\/p>\n

We formulate image reconstruction as inverse multiple-scattering problem\u00a0which is solved by nonlinear optimization method. The\u00a0mathematically-exact functional derivative is found with the\u00a0distorted-Born approximation.\u00a0 However, a large image reconstruction\u00a0requires solving hundreds of thousands of forward-scattering problems,\u00a0i.e., inversion of large _N_-by-_N_ dense matrices.\u00a0 As a remedy, we\u00a0employ the multilevel fast multipole algorithm for solving these forward\u00a0problems with O(_N_) computational complexity.\u00a0 Furthermore, we use\u00a0NCSA\u2019s Blue Waters supercomputing facility with CPU+GPU node
\narchitecture for massively-parallel reconstructions.\u00a0 For efficient\u00a0implementation, we seek low-level GPU optimizations and effective\u00a0heterogeneous computing.<\/p>\n

This talk provides an overview of computational methods, efficient\u00a0parallelization strategies on large supercomputers, some GPU\u00a0optimizations, and performance results.\u00a0 Results show good scaling up to\u00a04,096 GPU nodes which provides the largest full-wave image\u00a0reconstructions to date in near-real time. Several real-life scenarios\u00a0will be provided where the proposed methodology is especially useful and\u00a0outperforms conventional approaches in terms of image quality.<\/p>\n

Shirt bio of Mert Hidayetoglu<\/strong>:<\/p>\n

Mert Hidayetoglu is a Ph.D. candidate ECE Illinois, where he works with
\nProf. Weng Cho Chew and Prof. Wen-Mei Hwu. Mert\u2019s research interests
\ninclude electromagnetics, fast algorithms for integral-equation methods,
\ninverse scattering and imaging, and parallel & high-performance
\ncomputing. He is among the first two authors of more than a dozen of
\nconference papers and of several journal papers. He was awarded with the
\nprofessor Kung Chie Yeh and Dan Vivoli endowed fellowships by the ECE
\ndepartment, and with computational science and engineering fellowship by
\nthe college of engineering at Illinois. Recently, he won the second
\nplace at the 2018 IEEE IPDPS Ph.D. Forum poster competition by popular
\nvote. In Summer 2018, he was a Givens fellow at Argonne National
\nLaboratory.<\/p>\n","protected":false},"excerpt":{"rendered":"

Mert Hidayetoglu of University of Illinois at Urbana Champain presented on his PhD studies. Talk was entitled \u201cSUPERCOMPUTING FOR FULL-WAVE TOMOGRAPHIC IMAGE RECONSTRUCTION IN NEAR-REAL TIME.\u201d The abstract of the talk is: Full-wave methods incorporate all wave phenomena into the image\u00a0reconstructions by solving the Helmholtz equation with no fundamental\u00a0approximation. These phenomena include refraction, absorption,\u00a0diffraction, and […]<\/p>\n","protected":false},"author":1,"featured_media":4073,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[445],"tags":[],"_links":{"self":[{"href":"https:\/\/umram.bilkent.edu.tr\/index.php\/wp-json\/wp\/v2\/posts\/4074"}],"collection":[{"href":"https:\/\/umram.bilkent.edu.tr\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/umram.bilkent.edu.tr\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/umram.bilkent.edu.tr\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/umram.bilkent.edu.tr\/index.php\/wp-json\/wp\/v2\/comments?post=4074"}],"version-history":[{"count":1,"href":"https:\/\/umram.bilkent.edu.tr\/index.php\/wp-json\/wp\/v2\/posts\/4074\/revisions"}],"predecessor-version":[{"id":4075,"href":"https:\/\/umram.bilkent.edu.tr\/index.php\/wp-json\/wp\/v2\/posts\/4074\/revisions\/4075"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/umram.bilkent.edu.tr\/index.php\/wp-json\/wp\/v2\/media\/4073"}],"wp:attachment":[{"href":"https:\/\/umram.bilkent.edu.tr\/index.php\/wp-json\/wp\/v2\/media?parent=4074"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/umram.bilkent.edu.tr\/index.php\/wp-json\/wp\/v2\/categories?post=4074"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/umram.bilkent.edu.tr\/index.php\/wp-json\/wp\/v2\/tags?post=4074"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}