{"id":332,"date":"2019-02-22T11:58:37","date_gmt":"2019-02-22T10:58:37","guid":{"rendered":"https:\/\/fsr.cs.uni-potsdam.de\/?p=332"},"modified":"2019-02-22T11:58:37","modified_gmt":"2019-02-22T10:58:37","slug":"paralleles-rechnen-2005","status":"publish","type":"post","link":"https:\/\/fsr.cs.uni-potsdam.de\/?p=332","title":{"rendered":"Paralleles Rechnen 2005"},"content":{"rendered":"

Aufgabe 1<\/h1>\n
\n

Gitter 2^10 x 2^10 , Torus, 9-Point-Stencil<\/p>\n

1) 1-dim. Zerlegung in Streifen<\/p>\n

2) 2-dim. Zerlegung in Rechtecke f\u00fcr n = 2^k x 2^k Prozessoren<\/p>\n

a) Zerlegung skizzieren f\u00fcr k=2<\/p>\n

b) Task Interaction Graph zeichnen f\u00fcr k=2<\/p>\n

c) Formel f\u00fcr die Anzahl Nachrichten bzw. Datenmenge finden (pro Knoten 4Byte)<\/p>\n

d) Formel aus c) f\u00fcr k=2 und k=3 ausrechnen<\/p>\n

e) Welche Zerlegung ist effizienter und warum?<\/p>\n<\/div>\n

Aufgabe 2<\/h1>\n
\n

a) Was versteht man unter SPMD?<\/p>\n

b) Welche Semantik\/Parameter hat MPI_Barrier<\/p>\n

c) Anwendungsbeispiel f\u00fcr MPI_Barrier<\/p>\n

d) m\u00f6gliche Realisierung von MPI_Barrier durch MPI Laufzeitumgebung<\/p>\n<\/div>\n

Aufgabe 3<\/h1>\n
\n

geg: T(1)=1000 T(10)=700 T(25)=680<\/p>\n

a) ges: S(10) S(25)<\/p>\n

b) Karp-Flatt-Metrik f\u00fcr n=10 und n=25 bzw. max !SpeedUp<\/p>\n

c) E(10), E(25) und max. n damit E(n) \u00fcber 0,5<\/p>\n

d) Beurteilung der Parallelisierbarkeit<\/p>\n<\/div>\n

Aufgabe 4<\/h1>\n
\n

HPF-Programm kommentieren<\/p>\n

PROGRAM ABmult IMPLICIT NONE INTEGER, PARAMETER :: N = 100 INTEGER, DIMENSION (N,N) :: A, B, C INTEGER :: i, j\n\n!HPF$ PROCESSORS square(2,2)\n\n!HPF$ DISTRIBUTE (BLOCK,BLOCK) ONTO square :: C\n\n!HPF$ ALIGN A(i, * ) WITH C(i,j)\n\n!HPF$ ALIGN B( * ,j) WITH C(i,j)\n\nA = 1 B = 2 C = 0\n\nDO i = 1, N\n\n  DO\n\n    j = 1, N C(i,j) = DOT_PRODUCT(A(i,:), B(:,j))\n\n  END DO\n\nEND DO\n\nWRITE( * ,*) C\nEND<\/pre>\n<\/div>\n
Aufgabe 5<\/h1>\n
\n
Was sind Receive-Livelocks und wie werden sie im LINUX-Kernel vermieden?<\/p>\n<\/div>\n","protected":false},"excerpt":{"rendered":"
Aufgabe 1 Gitter 2^10 x 2^10 , Torus, 9-Point-Stencil 1) 1-dim. Zerlegung in Streifen 2) 2-dim. Zerlegung in Rechtecke f\u00fcr n = 2^k x 2^k Prozessoren a) Zerlegung skizzieren f\u00fcr k=2 b) Task Interaction Graph zeichnen f\u00fcr k=2 c) Formel f\u00fcr die Anzahl Nachrichten bzw. Datenmenge finden (pro Knoten 4Byte) d) Formel aus c) f\u00fcr weiterlesen…<\/a><\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-332","post","type-post","status-publish","format-standard","hentry","category-allgemein"],"_links":{"self":[{"href":"https:\/\/fsr.cs.uni-potsdam.de\/index.php?rest_route=\/wp\/v2\/posts\/332","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/fsr.cs.uni-potsdam.de\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/fsr.cs.uni-potsdam.de\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/fsr.cs.uni-potsdam.de\/index.php?rest_route=\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/fsr.cs.uni-potsdam.de\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=332"}],"version-history":[{"count":0,"href":"https:\/\/fsr.cs.uni-potsdam.de\/index.php?rest_route=\/wp\/v2\/posts\/332\/revisions"}],"wp:attachment":[{"href":"https:\/\/fsr.cs.uni-potsdam.de\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=332"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/fsr.cs.uni-potsdam.de\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=332"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/fsr.cs.uni-potsdam.de\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=332"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}