b_eff = 915.478 MB/s = 38.145 * 24 PEs with 1024 MB/PE on HI-UX/MPP himiko 03-00 ad2b0 SR8000 SECTION-HEAD-BEGIN b_eff.c, Revision 3.3 from Nov. 29, 1999 MEMORY_PER_PROCESSOR = 1024 MBytes [1M = 1024*1024] 1-dim-paterns: size = 24 2-dim-paterns: size = 6 * 4 3-dim-paterns: size = 4 * 3 * 2 Used message lengths: msglng= 1 2 4 8 16 32 64 128 256 512 1024 2048 4096 10624 27554 71468 185364 480774 1246974 3234251 8388608 Used methods: 0=Sndrcv 1=Alltoal 2=non-blk Used patterns: 0=ring-12*2fix 1=ring-6*4fix 2=ring-3*8fix 3=ring-1*24fix 4=ring-1*24fix 5=ring-1*24fix 6=random-cyc-1dim 7=random-cyc-1dim 8=random-cyc-1dim 9=random-cyc-1dim 10=random-cyc-1dim 11=random-cyc-1dim 12=random-cyc-1dim 13=random-cyc-1dim 14=random-cyc-1dim 15=random-cyc-1dim 16=random-cyc-1dim 17=random-cyc-1dim 18=random-cyc-1dim 19=random-cyc-1dim 20=random-cyc-1dim 21=random-cyc-1dim 22=random-cyc-1dim 23=random-cyc-1dim 24=random-cyc-1dim 25=random-cyc-1dim 26=random-cyc-1dim 27=random-cyc-1dim 28=random-cyc-1dim 29=random-cyc-1dim 30=random-cyc-1dim 31=random-cyc-1dim 32=random-cyc-1dim 33=random-cyc-1dim 34=random-cyc-1dim 35=random-cyc-1dim 36=worst-cyc-1dim 37=best bi-section 38=worst bi-section 39=one PingPong Pair 40=acyclic-2dim-all 41=acyclic-3dim-all 42=cyclic-2dim-x 43=cyclic-2dim-y 44=cyclic-2dim-all 45=cyclic-3dim-x 46=cyclic-3dim-y 47=cyclic-3dim-z 48=cyclic-3dim-all SECTION-HEAD-END SECTION-LOOP-BEGIN measurment loop: 130.967 sec SECTION-LOOP-END SECTION-LOOPLNGS-BEGIN method=0 | ------- repetition 0 --------- ------- repetition 1 --------- ------- repetition 2 --------- [sec] (= Sndrcv)| loop total per pattrn&mthd loop total per pattrn&mthd loop total per pattrn&mthd msg length| lng time minimum maximum lng time minimum maximum lng time minimum maximum ----------| ---- ------- ------- --------- ---- ------- ------- --------- ---- ------- ------- --------- 1 300 1.0e+00 9.0e-03 4.7e-02 124 4.3e-01 3.7e-03 1.9e-02 124 4.3e-01 3.7e-03 1.9e-02 2 150 5.2e-01 4.5e-03 2.4e-02 83 2.9e-01 2.5e-03 1.3e-02 82 2.8e-01 2.4e-03 1.3e-02 4 83 2.9e-01 2.5e-03 1.3e-02 83 2.9e-01 2.5e-03 1.3e-02 83 2.9e-01 2.5e-03 1.3e-02 8 81 2.8e-01 2.4e-03 1.3e-02 82 2.8e-01 2.5e-03 1.3e-02 81 2.8e-01 2.4e-03 1.3e-02 16 83 3.3e-01 2.7e-03 1.4e-02 83 3.0e-01 2.7e-03 1.4e-02 84 3.0e-01 2.7e-03 1.4e-02 32 78 2.8e-01 2.6e-03 1.3e-02 76 2.7e-01 2.5e-03 1.3e-02 78 2.8e-01 2.5e-03 1.3e-02 64 76 2.8e-01 2.5e-03 1.3e-02 76 2.8e-01 2.5e-03 1.3e-02 77 2.8e-01 2.5e-03 1.3e-02 128 75 2.8e-01 2.5e-03 1.3e-02 75 2.9e-01 2.7e-03 1.3e-02 77 2.9e-01 2.6e-03 1.4e-02 256 73 2.9e-01 2.6e-03 1.3e-02 70 2.8e-01 2.5e-03 1.3e-02 74 2.9e-01 2.6e-03 1.4e-02 512 69 2.8e-01 2.5e-03 1.3e-02 70 2.8e-01 2.5e-03 1.3e-02 71 2.8e-01 2.6e-03 1.3e-02 1024 70 2.9e-01 2.7e-03 1.4e-02 70 2.9e-01 2.5e-03 1.3e-02 68 2.7e-01 2.4e-03 1.3e-02 2048 64 5.0e-01 3.5e-03 2.0e-02 69 5.0e-01 3.6e-03 2.2e-02 70 5.1e-01 3.9e-03 2.2e-02 4096 46 4.5e-01 3.3e-03 1.9e-02 47 4.9e-01 3.4e-03 2.1e-02 45 4.4e-01 3.2e-03 1.9e-02 10624 27 6.5e-01 3.2e-03 3.2e-02 26 6.3e-01 3.1e-03 3.0e-02 26 6.2e-01 3.0e-03 3.1e-02 27554 16 7.1e-01 3.3e-03 3.6e-02 16 7.1e-01 3.3e-03 3.5e-02 16 7.1e-01 3.3e-03 3.7e-02 71468 9 9.2e-01 3.6e-03 3.9e-02 9 9.2e-01 3.6e-03 3.9e-02 9 9.2e-01 3.6e-03 3.9e-02 185364 4 7.5e-01 3.2e-03 3.5e-02 4 7.8e-01 3.2e-03 3.5e-02 4 7.6e-01 3.2e-03 3.4e-02 480774 2 9.2e-01 3.1e-03 4.6e-02 2 9.3e-01 3.2e-03 3.9e-02 2 9.3e-01 3.1e-03 4.0e-02 1246974 1 1.1e+00 3.7e-03 4.3e-02 1 1.0e+00 3.7e-03 4.7e-02 1 1.0e+00 3.8e-03 4.4e-02 3234251 1 1.3e+00 9.2e-03 1.0e-01 M 1 1.6e+00 9.3e-03 1.2e-01 M 1 1.5e+00 9.1e-03 1.1e-01 M 8388608 1 3.2e+00 2.3e-02 3.1e-01 R 1 4.1e+00 2.3e-02 2.9e-01 R 0 0.0e+00 1.0e+30 0.0e+00 R Explanation: M = only the best method is used R = only best method and only 2 repetitions method=1 | ------- repetition 0 --------- ------- repetition 1 --------- ------- repetition 2 --------- [sec] (=Alltoal)| loop total per pattrn&mthd loop total per pattrn&mthd loop total per pattrn&mthd msg length| lng time minimum maximum lng time minimum maximum lng time minimum maximum ----------| ---- ------- ------- --------- ---- ------- ------- --------- ---- ------- ------- --------- 1 300 1.0e+01 2.1e-01 2.4e-01 27 9.3e-01 1.9e-02 2.1e-02 4 1.4e-01 2.7e-03 3.1e-03 2 150 5.1e+00 1.0e-01 1.2e-01 13 4.5e-01 8.9e-03 1.1e-02 3 1.0e-01 2.1e-03 2.4e-03 4 75 2.6e+00 5.2e-02 5.9e-02 6 2.1e-01 4.1e-03 4.8e-03 3 1.0e-01 2.0e-03 2.4e-03 8 37 1.3e+00 2.5e-02 2.9e-02 3 1.0e-01 2.0e-03 2.4e-03 3 1.0e-01 2.0e-03 2.3e-03 16 18 6.2e-01 1.2e-02 1.4e-02 3 1.0e-01 2.0e-03 2.4e-03 3 1.0e-01 2.1e-03 2.4e-03 32 9 3.1e-01 6.2e-03 7.1e-03 3 1.0e-01 2.1e-03 2.4e-03 3 1.0e-01 2.0e-03 2.4e-03 64 4 1.4e-01 2.8e-03 3.8e-03 3 1.1e-01 2.1e-03 2.4e-03 3 1.0e-01 2.0e-03 2.4e-03 128 3 1.1e-01 2.1e-03 2.5e-03 3 1.1e-01 2.1e-03 3.3e-03 3 1.1e-01 2.1e-03 2.4e-03 256 3 1.1e-01 2.1e-03 2.4e-03 3 1.1e-01 2.1e-03 2.4e-03 3 1.1e-01 2.1e-03 2.5e-03 512 3 1.1e-01 2.1e-03 2.5e-03 3 1.1e-01 2.1e-03 2.5e-03 3 1.1e-01 2.1e-03 2.5e-03 1024 3 1.1e-01 2.1e-03 2.4e-03 3 1.1e-01 2.1e-03 2.5e-03 3 1.1e-01 2.1e-03 2.5e-03 2048 3 1.4e-01 2.2e-03 4.3e-03 3 1.3e-01 2.2e-03 3.2e-03 3 1.3e-01 2.3e-03 3.2e-03 4096 3 1.5e-01 2.3e-03 3.9e-03 3 1.6e-01 2.3e-03 4.1e-03 3 1.5e-01 2.3e-03 3.9e-03 10624 2 1.5e-01 2.0e-03 4.7e-03 2 1.5e-01 2.0e-03 4.7e-03 2 1.5e-01 1.9e-03 8.0e-03 27554 1 1.2e-01 1.1e-03 4.1e-03 1 1.0e-01 1.1e-03 3.0e-03 1 1.0e-01 1.0e-03 3.4e-03 71468 1 2.0e-01 1.5e-03 1.1e-02 1 1.9e-01 1.4e-03 7.1e-03 1 1.8e-01 1.3e-03 6.4e-03 185364 1 3.5e-01 2.0e-03 1.5e-02 1 3.3e-01 3.2e-03 1.1e-02 1 3.2e-01 4.0e-03 1.0e-02 480774 1 7.6e-01 5.2e-03 2.6e-02 1 7.4e-01 4.4e-03 2.6e-02 1 7.4e-01 5.3e-03 2.4e-02 1246974 1 1.7e+00 1.2e-02 6.0e-02 1 1.6e+00 1.1e-02 6.0e-02 1 1.6e+00 1.1e-02 5.9e-02 3234251 1 0.0e+00 0.0e+00 0.0e+00 M 1 0.0e+00 0.0e+00 0.0e+00 M 1 0.0e+00 0.0e+00 0.0e+00 M 8388608 1 0.0e+00 0.0e+00 0.0e+00 R 1 0.0e+00 0.0e+00 0.0e+00 R 0 0.0e+00 1.0e+30 0.0e+00 R Explanation: M = only the best method is used R = only best method and only 2 repetitions method=2 | ------- repetition 0 --------- ------- repetition 1 --------- ------- repetition 2 --------- [sec] (=non-blk)| loop total per pattrn&mthd loop total per pattrn&mthd loop total per pattrn&mthd msg length| lng time minimum maximum lng time minimum maximum lng time minimum maximum ----------| ---- ------- ------- --------- ---- ------- ------- --------- ---- ------- ------- --------- 1 300 1.7e+00 1.7e-02 6.9e-02 66 3.6e-01 3.6e-03 1.5e-02 64 3.5e-01 3.5e-03 1.5e-02 2 150 8.3e-01 8.4e-03 3.5e-02 45 2.4e-01 2.5e-03 1.0e-02 45 2.4e-01 2.5e-03 1.0e-02 4 75 4.1e-01 4.3e-03 1.8e-02 45 2.4e-01 2.5e-03 1.1e-02 45 2.4e-01 2.5e-03 1.1e-02 8 44 2.4e-01 2.5e-03 1.0e-02 44 2.3e-01 2.5e-03 1.0e-02 44 2.3e-01 2.4e-03 1.0e-02 16 44 2.5e-01 2.6e-03 1.1e-02 43 2.4e-01 2.5e-03 1.0e-02 45 2.5e-01 2.6e-03 1.1e-02 32 41 2.4e-01 2.4e-03 1.0e-02 43 2.4e-01 2.5e-03 1.0e-02 42 2.4e-01 2.4e-03 1.0e-02 64 42 2.4e-01 2.5e-03 1.0e-02 42 2.4e-01 2.6e-03 1.0e-02 42 2.4e-01 2.5e-03 1.0e-02 128 41 2.4e-01 2.5e-03 1.0e-02 41 2.4e-01 2.5e-03 1.0e-02 42 2.4e-01 2.5e-03 1.1e-02 256 40 2.5e-01 2.4e-03 1.1e-02 40 2.4e-01 2.4e-03 1.1e-02 41 2.5e-01 2.5e-03 1.1e-02 512 41 2.6e-01 2.5e-03 1.1e-02 41 2.5e-01 2.5e-03 1.1e-02 41 2.5e-01 2.5e-03 1.1e-02 1024 40 2.5e-01 2.5e-03 1.1e-02 41 2.5e-01 2.6e-03 1.1e-02 41 2.5e-01 2.5e-03 1.1e-02 2048 39 3.5e-01 3.2e-03 1.4e-02 40 3.5e-01 3.4e-03 1.5e-02 41 3.6e-01 3.4e-03 1.5e-02 4096 30 3.3e-01 3.1e-03 1.4e-02 29 3.5e-01 3.1e-03 1.4e-02 30 3.4e-01 3.2e-03 1.4e-02 10624 18 4.2e-01 2.9e-03 1.8e-02 17 3.9e-01 2.6e-03 1.8e-02 18 3.9e-01 2.9e-03 1.7e-02 27554 12 4.7e-01 2.9e-03 2.0e-02 12 4.6e-01 2.7e-03 2.0e-02 12 4.5e-01 3.0e-03 2.0e-02 71468 8 7.9e-01 3.4e-03 4.0e-02 8 7.7e-01 3.4e-03 3.4e-02 7 6.7e-01 3.0e-03 3.1e-02 185364 4 7.6e-01 3.6e-03 3.3e-02 4 7.4e-01 3.8e-03 3.2e-02 4 7.3e-01 3.7e-03 3.2e-02 480774 2 9.1e-01 6.0e-03 4.4e-02 2 9.0e-01 4.3e-03 4.3e-02 2 8.5e-01 4.6e-03 4.1e-02 1246974 1 1.1e+00 5.3e-03 5.6e-02 1 1.1e+00 4.9e-03 5.5e-02 1 1.0e+00 5.0e-03 5.7e-02 3234251 1 1.2e+00 1.3e-02 9.0e-02 M 1 8.5e-01 1.2e-02 9.4e-02 M 1 1.0e+00 1.4e-02 7.7e-02 M 8388608 1 2.9e+00 3.3e-02 1.9e-01 R 1 2.0e+00 3.6e-02 2.0e-01 R 0 0.0e+00 1.0e+30 0.0e+00 R Explanation: M = only the best method is used R = only best method and only 2 repetitions SECTION-LOOPLNGS-END SECTION-ELAPSED-BEGIN measurment loop: elapsed time = 130.967 sec sum of max elapsed time per entries above = 131.172 sec difference to elapsed time = -0.205 sec = 0.2% sum based on fastest repetition = 110.500 sec difference to elapsed time = 20.467 sec = 15.6% CAUTION: A difference above is more than 5 %. There may be problems with the MPI implementation or processes weren't on dedicated processors SECTION-ELAPSED-END SECTION-PATTERN-BEGIN Pattern parameters: sendrecv total number messages/ messages/ unused best method calls of messages used node node&call nodes per repetition ---- ----------------- ----- ------------ --------- --------- ------ -------------- p00 ring-12*2fix 1 24 1.00 1.00 0 ( 2 0 2 ) p01 ring-6*4fix 2 48 2.00 1.00 0 ( 0 2 2 ) p02 ring-3*8fix 2 48 2.00 1.00 0 ( 0 0 2 ) p03 ring-1*24fix 2 48 2.00 1.00 0 ( 0 0 0 ) p04 ring-1*24fix 2 48 2.00 1.00 0 ( 0 0 0 ) p05 ring-1*24fix 2 48 2.00 1.00 0 ( 0 0 0 ) p06 random-cyc-1dim 2 48 2.00 1.00 0 ( 0 0 0 ) p07 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 2 2 ) p08 random-cyc-1dim 2 48 2.00 1.00 0 ( 0 0 0 ) p09 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 0 0 ) p10 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 2 2 ) p11 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 0 0 ) p12 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 2 2 ) p13 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 0 0 ) p14 random-cyc-1dim 2 48 2.00 1.00 0 ( 0 0 0 ) p15 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 0 2 ) p16 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 2 2 ) p17 random-cyc-1dim 2 48 2.00 1.00 0 ( 0 0 0 ) p18 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 2 2 ) p19 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 2 2 ) p20 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 2 2 ) p21 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 2 0 ) p22 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 0 0 ) p23 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 0 0 ) p24 random-cyc-1dim 2 48 2.00 1.00 0 ( 0 2 0 ) p25 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 2 2 ) p26 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 2 2 ) p27 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 0 2 ) p28 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 0 2 ) p29 random-cyc-1dim 2 48 2.00 1.00 0 ( 0 0 0 ) p30 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 0 2 ) p31 random-cyc-1dim 2 48 2.00 1.00 0 ( 0 0 0 ) p32 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 2 2 ) p33 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 2 0 ) p34 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 2 2 ) p35 random-cyc-1dim 2 48 2.00 1.00 0 ( 2 0 0 ) p36 worst-cyc-1dim 2 48 2.00 1.00 0 ( 0 0 0 ) p37 best bi-section 2 24 1.00 0.50 0 ( 0 2 2 ) p38 worst bi-section 2 24 1.00 0.50 0 ( 2 2 2 ) p39 one PingPong Pair 2 2 1.00 0.50 22 ( 0 0 0 ) p40 acyclic-2dim-all 4 76 3.17 0.79 0 ( 0 2 0 ) p41 acyclic-3dim-all 6 92 3.83 0.64 0 ( 2 2 2 ) p42 cyclic-2dim-x 2 48 2.00 1.00 0 ( 0 0 0 ) p43 cyclic-2dim-y 2 48 2.00 1.00 0 ( 0 0 0 ) p44 cyclic-2dim-all 4 96 4.00 1.00 0 ( 0 0 0 ) p45 cyclic-3dim-x 2 48 2.00 1.00 0 ( 2 2 2 ) p46 cyclic-3dim-y 2 48 2.00 1.00 0 ( 0 0 2 ) p47 cyclic-3dim-z 1 24 1.00 1.00 0 ( 0 0 0 ) p48 cyclic-3dim-all 5 120 5.00 1.00 0 ( 0 0 0 ) SECTION-PATTERN-END Printing accumulated bandwidth / number of processes = bandwidth per process (in MBytes/sec = 1e6 Bytes/sec) Additionally the accumulated value processes is printed in the last column Only the cyclic and random patterns are used to compute b_eff (except p00), the other patterns are only as information. SECTION-BY-METHODS-BEGIN 2nd analysis path, only as information, last row: for all mthd: logavg_pat ( avg_L ( max_rep ( b(L) ) ) ) || logavg_pat ( max_mthd ( avg_L ( max_rep ( b(L) ) ) ) ) pattern / methods: Sndrcv Alltoal non-blk -> best method -> accumulated p00 ring-12*2fix : 30.804 18.449 30.085 -> 30.804 -> 739.287 MByte/s p01 ring-6*4fix : 44.127 26.820 40.710 -> 44.127 -> 1059.052 MByte/s p02 ring-3*8fix : 32.820 21.161 31.047 -> 32.820 -> 787.672 MByte/s p03 ring-1*24fix : 34.224 21.343 32.138 -> 34.224 -> 821.381 MByte/s p04 ring-1*24fix : 34.906 21.935 32.215 -> 34.906 -> 837.741 MByte/s p05 ring-1*24fix : 34.936 21.846 32.326 -> 34.936 -> 838.468 MByte/s p06 random-cyc-1dim : 33.908 21.048 32.477 -> 33.908 -> 813.786 MByte/s p07 random-cyc-1dim : 44.168 25.243 42.863 -> 44.168 -> 1060.040 MByte/s p08 random-cyc-1dim : 41.024 24.195 39.509 -> 41.024 -> 984.576 MByte/s p09 random-cyc-1dim : 48.610 27.509 46.203 -> 48.610 -> 1166.644 MByte/s p10 random-cyc-1dim : 39.890 23.816 38.008 -> 39.890 -> 957.348 MByte/s p11 random-cyc-1dim : 39.457 23.065 36.297 -> 39.457 -> 946.957 MByte/s p12 random-cyc-1dim : 37.844 22.642 35.649 -> 37.844 -> 908.245 MByte/s p13 random-cyc-1dim : 39.702 23.612 37.702 -> 39.702 -> 952.847 MByte/s p14 random-cyc-1dim : 40.478 23.729 38.030 -> 40.478 -> 971.483 MByte/s p15 random-cyc-1dim : 40.870 23.846 39.665 -> 40.870 -> 980.885 MByte/s p16 random-cyc-1dim : 37.401 22.534 35.812 -> 37.401 -> 897.619 MByte/s p17 random-cyc-1dim : 42.192 24.550 39.502 -> 42.192 -> 1012.603 MByte/s p18 random-cyc-1dim : 41.850 24.524 39.823 -> 41.850 -> 1004.397 MByte/s p19 random-cyc-1dim : 37.694 22.569 36.115 -> 37.694 -> 904.667 MByte/s p20 random-cyc-1dim : 39.794 22.706 38.279 -> 39.794 -> 955.067 MByte/s p21 random-cyc-1dim : 41.248 24.445 40.485 -> 41.248 -> 989.944 MByte/s p22 random-cyc-1dim : 37.996 22.292 36.526 -> 37.996 -> 911.892 MByte/s p23 random-cyc-1dim : 42.288 24.612 41.366 -> 42.288 -> 1014.918 MByte/s p24 random-cyc-1dim : 47.741 27.323 44.416 -> 47.741 -> 1145.784 MByte/s p25 random-cyc-1dim : 42.784 25.223 40.942 -> 42.784 -> 1026.824 MByte/s p26 random-cyc-1dim : 38.136 23.869 38.257 -> 38.257 -> 918.169 MByte/s p27 random-cyc-1dim : 43.507 25.779 42.730 -> 43.507 -> 1044.175 MByte/s p28 random-cyc-1dim : 43.219 25.929 43.083 -> 43.219 -> 1037.257 MByte/s p29 random-cyc-1dim : 38.023 22.294 35.779 -> 38.023 -> 912.554 MByte/s p30 random-cyc-1dim : 37.775 22.728 36.552 -> 37.775 -> 906.599 MByte/s p31 random-cyc-1dim : 34.828 20.931 33.211 -> 34.828 -> 835.879 MByte/s p32 random-cyc-1dim : 43.110 25.618 42.159 -> 43.110 -> 1034.632 MByte/s p33 random-cyc-1dim : 38.973 23.814 38.139 -> 38.973 -> 935.340 MByte/s p34 random-cyc-1dim : 36.099 21.026 34.895 -> 36.099 -> 866.373 MByte/s p35 random-cyc-1dim : 37.388 22.597 35.389 -> 37.388 -> 897.303 MByte/s -- additional patterns that are not used to compute b_eff: p36 worst-cyc-1dim : 40.196 24.164 38.672 -> 40.196 -> 964.707 MByte/s p37 best bi-section : 29.846 18.434 29.506 -> 29.846 -> 716.298 MByte/s p38 worst bi-section : 36.543 26.211 40.435 -> 40.435 -> 970.437 MByte/s p39 one PingPong Pair : 8.097 2.878 2.878 -> 8.097 -> 194.330 MByte/s p40 acyclic-2dim-all : 33.627 25.045 31.400 -> 33.627 -> 807.038 MByte/s p41 acyclic-3dim-all : 42.394 32.369 44.752 -> 44.752 -> 1074.058 MByte/s p42 cyclic-2dim-x : 34.574 21.846 32.527 -> 34.574 -> 829.781 MByte/s p43 cyclic-2dim-y : 46.053 28.810 42.325 -> 46.053 -> 1105.282 MByte/s p44 cyclic-2dim-all : 38.997 28.906 36.483 -> 38.997 -> 935.936 MByte/s p45 cyclic-3dim-x : 154.725 61.276 144.026 -> 154.725 -> 3713.392 MByte/s p46 cyclic-3dim-y : 32.697 23.040 30.478 -> 32.697 -> 784.728 MByte/s p47 cyclic-3dim-z : 30.344 18.494 29.301 -> 30.344 -> 728.250 MByte/s p48 cyclic-3dim-all : 47.824 36.410 46.044 -> 47.824 -> 1147.765 MByte/s log_avg of all rings : 35.074 21.792 32.919 || 35.074 -> 841.788 MByte/s log_avg of all random : 40.134 23.746 38.527 || 40.138 -> 963.314 MByte/s log_avg(ring,random) : 37.519 22.748 35.613 ||( 37.521 -> 900.503)MByte/s * size -> accumulated on all pr.: 900.456 545.946 854.706 ||(900.503)MByte/s SECTION-BY-METHODS-END SECTION-BY-REPETITIONS-BEGIN 3rd analysis path, only as information, last row: for all rep.:logavg_pat ( avg_L ( max_mthd ( b(L) ) ) ) || logavg_pat ( max_rep ( avg_L ( max_mthd ( b(L) ) ) ) ) pattern / repetition: rep.0 rep.1 rep.2 -> best repetition -> accumulated p00 ring-12*2fix : 29.843 30.484 31.116 -> 31.116 -> 746.779 MByte/s p01 ring-6*4fix : 41.393 42.755 43.818 -> 43.818 -> 1051.642 MByte/s p02 ring-3*8fix : 32.235 31.993 32.485 -> 32.485 -> 779.637 MByte/s p03 ring-1*24fix : 33.002 33.709 32.673 -> 33.709 -> 809.016 MByte/s p04 ring-1*24fix : 33.755 33.150 34.300 -> 34.300 -> 823.208 MByte/s p05 ring-1*24fix : 33.586 34.693 33.242 -> 34.693 -> 832.633 MByte/s p06 random-cyc-1dim : 32.267 33.803 33.737 -> 33.803 -> 811.272 MByte/s p07 random-cyc-1dim : 41.342 43.638 44.415 -> 44.415 -> 1065.965 MByte/s p08 random-cyc-1dim : 38.949 41.232 40.021 -> 41.232 -> 989.572 MByte/s p09 random-cyc-1dim : 44.304 48.262 48.718 -> 48.718 -> 1169.240 MByte/s p10 random-cyc-1dim : 37.047 39.220 39.698 -> 39.698 -> 952.740 MByte/s p11 random-cyc-1dim : 35.103 38.446 38.391 -> 38.446 -> 922.712 MByte/s p12 random-cyc-1dim : 34.996 37.088 38.104 -> 38.104 -> 914.486 MByte/s p13 random-cyc-1dim : 36.780 38.056 39.790 -> 39.790 -> 954.957 MByte/s p14 random-cyc-1dim : 38.433 39.842 40.707 -> 40.707 -> 976.964 MByte/s p15 random-cyc-1dim : 38.537 40.658 40.816 -> 40.816 -> 979.586 MByte/s p16 random-cyc-1dim : 35.013 37.660 37.143 -> 37.660 -> 903.844 MByte/s p17 random-cyc-1dim : 38.914 41.744 41.916 -> 41.916 -> 1005.976 MByte/s p18 random-cyc-1dim : 39.784 40.961 42.281 -> 42.281 -> 1014.744 MByte/s p19 random-cyc-1dim : 35.685 37.679 37.154 -> 37.679 -> 904.290 MByte/s p20 random-cyc-1dim : 38.184 40.375 39.286 -> 40.375 -> 969.012 MByte/s p21 random-cyc-1dim : 39.128 41.728 41.405 -> 41.728 -> 1001.465 MByte/s p22 random-cyc-1dim : 36.069 37.702 38.104 -> 38.104 -> 914.499 MByte/s p23 random-cyc-1dim : 40.568 41.807 41.765 -> 41.807 -> 1003.379 MByte/s p24 random-cyc-1dim : 46.154 47.081 47.646 -> 47.646 -> 1143.498 MByte/s p25 random-cyc-1dim : 40.899 42.934 42.755 -> 42.934 -> 1030.422 MByte/s p26 random-cyc-1dim : 37.658 38.492 38.961 -> 38.961 -> 935.068 MByte/s p27 random-cyc-1dim : 42.640 43.983 43.404 -> 43.983 -> 1055.594 MByte/s p28 random-cyc-1dim : 42.253 43.632 43.848 -> 43.848 -> 1052.362 MByte/s p29 random-cyc-1dim : 36.068 36.429 37.894 -> 37.894 -> 909.449 MByte/s p30 random-cyc-1dim : 37.229 37.782 37.515 -> 37.782 -> 906.773 MByte/s p31 random-cyc-1dim : 34.028 33.933 35.120 -> 35.120 -> 842.884 MByte/s p32 random-cyc-1dim : 42.842 43.309 43.591 -> 43.591 -> 1046.183 MByte/s p33 random-cyc-1dim : 38.730 38.115 38.849 -> 38.849 -> 932.375 MByte/s p34 random-cyc-1dim : 34.820 36.290 35.488 -> 36.290 -> 870.952 MByte/s p35 random-cyc-1dim : 36.407 37.130 36.829 -> 37.130 -> 891.127 MByte/s -- additional patterns that are not used to compute b_eff: p36 worst-cyc-1dim : 39.431 40.038 40.131 -> 40.131 -> 963.138 MByte/s p37 best bi-section : 28.578 30.270 29.719 -> 30.270 -> 726.472 MByte/s p38 worst bi-section : 39.493 38.837 39.335 -> 39.493 -> 947.827 MByte/s p39 one PingPong Pair : 8.054 7.995 8.016 -> 8.054 -> 193.300 MByte/s p40 acyclic-2dim-all : 31.380 31.440 33.409 -> 33.409 -> 801.824 MByte/s p41 acyclic-3dim-all : 42.700 44.693 44.422 -> 44.693 -> 1072.622 MByte/s p42 cyclic-2dim-x : 32.114 34.209 33.742 -> 34.209 -> 821.005 MByte/s p43 cyclic-2dim-y : 44.349 45.459 45.499 -> 45.499 -> 1091.970 MByte/s p44 cyclic-2dim-all : 38.367 37.912 38.723 -> 38.723 -> 929.343 MByte/s p45 cyclic-3dim-x : 152.646 151.363 152.479 -> 152.646 -> 3663.502 MByte/s p46 cyclic-3dim-y : 32.813 31.827 30.808 -> 32.813 -> 787.501 MByte/s p47 cyclic-3dim-z : 30.025 29.079 29.925 -> 30.025 -> 720.608 MByte/s p48 cyclic-3dim-all : 47.598 49.431 48.291 -> 49.431 -> 1186.344 MByte/s log_avg of all rings : 33.795 34.258 34.375 || 34.802 -> 835.241 MByte/s log_avg of all random : 38.232 39.826 40.037 || 40.241 -> 965.775 MByte/s log_avg(ring,random) : 35.945 36.937 37.098 ||( 37.422 -> 898.140)MByte/s * size -> accumulated on all pr.: 862.686 886.493 890.359 ||(898.140)MByte/s SECTION-BY-REPETITIONS-END SECTION-BY-MTHD-MSGLNG-BEGIN 4th analysis path, only as information, last 3 rows: for all methods and some msglng L: logavg_pat ( max_rep ( b(L) ) ) for all methods: logavg_pat ( avg_L ( max_rep ( b(L) ) ) ) pattern & method / msg: 1 16 256 4096 185364 8388608 -> average -> accumulated (latency of one sendrecv (or equivalent method) in microsec) p00 ring-12*2fix p00 method 0 =Sndrcv :( 35.378) 0.028 0.435 6.448 36.798 69.302 90.705 -> 30.804 -> 739.287 MByte/s p00 method 1 =Alltoal :(692.517) 0.001 0.023 0.365 4.255 39.675 90.705 -> 18.449 -> 442.782 MByte/s p00 method 2 =non-blk :( 62.301) 0.016 0.248 3.788 32.751 67.930 90.705 -> 30.085 -> 722.043 MByte/s p01 ring-6*4fix p01 method 0 =Sndrcv :( 34.476) 0.029 0.447 6.569 42.092 97.259 147.356 -> 44.127 -> 1059.052 MByte/s p01 method 1 =Alltoal :(341.490) 0.003 0.046 0.707 7.445 56.912 147.356 -> 26.820 -> 643.681 MByte/s p01 method 2 =non-blk :( 56.886) 0.018 0.275 4.089 36.104 93.962 147.356 -> 40.710 -> 977.049 MByte/s p02 ring-3*8fix p02 method 0 =Sndrcv :( 35.173) 0.028 0.445 6.445 37.172 68.030 101.649 -> 32.820 -> 787.672 MByte/s p02 method 1 =Alltoal :(340.998) 0.003 0.046 0.711 7.243 48.386 101.649 -> 21.161 -> 507.854 MByte/s p02 method 2 =non-blk :( 57.219) 0.017 0.268 3.964 32.997 68.902 101.649 -> 31.047 -> 745.125 MByte/s p03 ring-1*24fix p03 method 0 =Sndrcv :( 35.070) 0.029 0.440 6.453 37.105 73.466 110.339 -> 34.224 -> 821.381 MByte/s p03 method 1 =Alltoal :(343.382) 0.003 0.046 0.706 6.925 44.866 110.339 -> 21.343 -> 512.228 MByte/s p03 method 2 =non-blk :( 56.924) 0.018 0.274 4.016 32.922 71.927 110.339 -> 32.138 -> 771.310 MByte/s p04 ring-1*24fix p04 method 0 =Sndrcv :( 34.972) 0.029 0.439 6.524 37.680 75.428 108.704 -> 34.906 -> 837.741 MByte/s p04 method 1 =Alltoal :(340.998) 0.003 0.046 0.702 6.695 45.101 108.704 -> 21.935 -> 526.450 MByte/s p04 method 2 =non-blk :( 57.182) 0.017 0.269 4.064 33.183 73.913 108.704 -> 32.215 -> 773.152 MByte/s p05 ring-1*24fix p05 method 0 =Sndrcv :( 35.068) 0.029 0.441 6.478 37.387 76.918 110.352 -> 34.936 -> 838.468 MByte/s p05 method 1 =Alltoal :(342.369) 0.003 0.046 0.699 6.838 45.707 110.352 -> 21.846 -> 524.293 MByte/s p05 method 2 =non-blk :( 56.992) 0.018 0.273 4.004 33.197 68.149 110.352 -> 32.326 -> 775.817 MByte/s p06 random-cyc-1dim p06 method 0 =Sndrcv :( 34.903) 0.029 0.439 6.499 38.364 76.510 99.885 -> 33.908 -> 813.786 MByte/s p06 method 1 =Alltoal :(357.628) 0.003 0.045 0.699 7.378 50.910 99.885 -> 21.048 -> 505.147 MByte/s p06 method 2 =non-blk :( 57.500) 0.017 0.265 3.977 33.287 76.712 99.885 -> 32.477 -> 779.456 MByte/s p07 random-cyc-1dim p07 method 0 =Sndrcv :( 34.424) 0.029 0.457 6.591 43.751 103.245 136.513 -> 44.168 -> 1060.040 MByte/s p07 method 1 =Alltoal :(365.376) 0.003 0.043 0.687 7.997 57.300 136.513 -> 25.243 -> 605.822 MByte/s p07 method 2 =non-blk :( 56.344) 0.018 0.274 4.056 37.231 108.139 136.513 -> 42.863 -> 1028.706 MByte/s p08 random-cyc-1dim p08 method 0 =Sndrcv :( 34.307) 0.029 0.446 6.628 42.575 91.662 121.322 -> 41.024 -> 984.576 MByte/s p08 method 1 =Alltoal :(364.006) 0.003 0.044 0.691 7.907 54.311 121.322 -> 24.195 -> 580.677 MByte/s p08 method 2 =non-blk :( 56.882) 0.018 0.270 4.085 36.323 97.548 121.322 -> 39.509 -> 948.221 MByte/s p09 random-cyc-1dim p09 method 0 =Sndrcv :( 34.317) 0.029 0.459 6.586 44.781 120.523 151.379 -> 48.610 -> 1166.644 MByte/s p09 method 1 =Alltoal :(389.874) 0.003 0.041 0.647 8.706 56.221 151.379 -> 27.509 -> 660.219 MByte/s p09 method 2 =non-blk :( 55.091) 0.018 0.278 4.036 37.676 119.110 151.379 -> 46.203 -> 1108.882 MByte/s p10 random-cyc-1dim p10 method 0 =Sndrcv :( 34.774) 0.029 0.443 6.547 41.143 94.201 120.683 -> 39.890 -> 957.348 MByte/s p10 method 1 =Alltoal :(364.006) 0.003 0.044 0.684 8.114 50.875 120.683 -> 23.816 -> 571.594 MByte/s p10 method 2 =non-blk :( 57.023) 0.018 0.269 4.059 33.342 93.335 120.683 -> 38.008 -> 912.196 MByte/s p11 random-cyc-1dim p11 method 0 =Sndrcv :( 34.728) 0.029 0.450 6.509 41.739 90.876 116.965 -> 39.457 -> 946.957 MByte/s p11 method 1 =Alltoal :(358.880) 0.003 0.044 0.699 7.765 50.160 116.965 -> 23.065 -> 553.570 MByte/s p11 method 2 =non-blk :( 57.803) 0.017 0.272 4.001 35.264 91.781 116.965 -> 36.297 -> 871.122 MByte/s p12 random-cyc-1dim p12 method 0 =Sndrcv :( 34.629) 0.029 0.444 6.490 41.443 86.893 111.176 -> 37.844 -> 908.245 MByte/s p12 method 1 =Alltoal :(354.990) 0.003 0.045 0.706 7.461 51.312 111.176 -> 22.642 -> 543.411 MByte/s p12 method 2 =non-blk :( 56.591) 0.018 0.270 3.957 35.846 89.047 111.176 -> 35.649 -> 855.574 MByte/s p13 random-cyc-1dim p13 method 0 =Sndrcv :( 29.649) 0.034 0.494 7.120 44.249 93.630 120.467 -> 39.702 -> 952.847 MByte/s p13 method 1 =Alltoal :(351.384) 0.003 0.045 0.709 7.599 52.878 120.467 -> 23.612 -> 566.679 MByte/s p13 method 2 =non-blk :( 51.690) 0.019 0.296 4.374 38.874 93.254 120.467 -> 37.702 -> 904.845 MByte/s p14 random-cyc-1dim p14 method 0 =Sndrcv :( 34.730) 0.029 0.446 6.466 41.881 94.447 119.603 -> 40.478 -> 971.483 MByte/s p14 method 1 =Alltoal :(352.129) 0.003 0.045 0.705 7.520 49.997 119.603 -> 23.729 -> 569.501 MByte/s p14 method 2 =non-blk :( 55.766) 0.018 0.276 3.992 36.120 94.791 119.603 -> 38.030 -> 912.726 MByte/s p15 random-cyc-1dim p15 method 0 =Sndrcv :( 34.453) 0.029 0.451 6.564 42.935 98.082 121.375 -> 40.870 -> 980.885 MByte/s p15 method 1 =Alltoal :(358.388) 0.003 0.044 0.666 7.887 52.378 121.375 -> 23.846 -> 572.305 MByte/s p15 method 2 =non-blk :( 56.053) 0.018 0.276 4.112 32.405 97.387 121.375 -> 39.665 -> 951.969 MByte/s p16 random-cyc-1dim p16 method 0 =Sndrcv :( 34.775) 0.029 0.441 6.306 40.946 82.761 111.988 -> 37.401 -> 897.619 MByte/s p16 method 1 =Alltoal :(366.136) 0.003 0.043 0.687 7.772 53.068 111.988 -> 22.534 -> 540.813 MByte/s p16 method 2 =non-blk :( 57.320) 0.017 0.267 3.911 33.296 89.983 111.988 -> 35.812 -> 859.496 MByte/s p17 random-cyc-1dim p17 method 0 =Sndrcv :( 34.348) 0.029 0.458 6.635 43.087 101.890 127.077 -> 42.192 -> 1012.603 MByte/s p17 method 1 =Alltoal :(370.756) 0.003 0.043 0.658 8.013 48.857 127.077 -> 24.550 -> 589.207 MByte/s p17 method 2 =non-blk :( 56.460) 0.018 0.275 3.974 35.726 105.127 127.077 -> 39.502 -> 948.039 MByte/s p18 random-cyc-1dim p18 method 0 =Sndrcv :( 34.657) 0.029 0.453 6.464 42.850 97.785 121.022 -> 41.850 -> 1004.397 MByte/s p18 method 1 =Alltoal :(358.254) 0.003 0.044 0.695 7.714 54.232 121.022 -> 24.524 -> 588.584 MByte/s p18 method 2 =non-blk :( 56.414) 0.018 0.274 4.028 35.872 101.368 121.022 -> 39.823 -> 955.755 MByte/s p19 random-cyc-1dim p19 method 0 =Sndrcv :( 34.673) 0.029 0.454 6.592 41.443 89.830 110.360 -> 37.694 -> 904.667 MByte/s p19 method 1 =Alltoal :(369.370) 0.003 0.043 0.685 7.714 51.191 110.360 -> 22.569 -> 541.655 MByte/s p19 method 2 =non-blk :( 56.054) 0.018 0.275 4.002 35.285 85.937 110.360 -> 36.115 -> 866.754 MByte/s p20 random-cyc-1dim p20 method 0 =Sndrcv :( 34.165) 0.029 0.457 6.525 42.304 87.199 121.116 -> 39.794 -> 955.067 MByte/s p20 method 1 =Alltoal :(354.379) 0.003 0.045 0.708 7.794 38.141 121.116 -> 22.706 -> 544.934 MByte/s p20 method 2 =non-blk :( 55.563) 0.018 0.278 4.079 35.684 94.967 121.116 -> 38.279 -> 918.701 MByte/s p21 random-cyc-1dim p21 method 0 =Sndrcv :( 32.617) 0.031 0.453 6.665 43.865 96.960 124.583 -> 41.248 -> 989.944 MByte/s p21 method 1 =Alltoal :(360.489) 0.003 0.044 0.697 8.055 56.060 124.583 -> 24.445 -> 586.668 MByte/s p21 method 2 =non-blk :( 52.098) 0.019 0.293 4.308 39.524 98.050 124.583 -> 40.485 -> 971.643 MByte/s p22 random-cyc-1dim p22 method 0 =Sndrcv :( 34.375) 0.029 0.453 6.567 40.942 87.876 111.939 -> 37.996 -> 911.892 MByte/s p22 method 1 =Alltoal :(361.264) 0.003 0.044 0.699 7.750 45.106 111.939 -> 22.292 -> 535.015 MByte/s p22 method 2 =non-blk :( 56.250) 0.018 0.274 4.048 36.051 91.374 111.939 -> 36.526 -> 876.627 MByte/s p23 random-cyc-1dim p23 method 0 =Sndrcv :( 34.302) 0.029 0.452 6.647 43.136 96.872 126.768 -> 42.288 -> 1014.918 MByte/s p23 method 1 =Alltoal :(367.746) 0.003 0.043 0.683 7.982 56.316 126.768 -> 24.612 -> 590.677 MByte/s p23 method 2 =non-blk :( 55.312) 0.018 0.275 4.051 36.588 101.354 126.768 -> 41.366 -> 992.794 MByte/s p24 random-cyc-1dim p24 method 0 =Sndrcv :( 34.072) 0.029 0.451 6.593 44.388 117.125 151.133 -> 47.741 -> 1145.784 MByte/s p24 method 1 =Alltoal :(361.368) 0.003 0.044 0.689 7.755 57.791 151.133 -> 27.323 -> 655.745 MByte/s p24 method 2 =non-blk :( 55.728) 0.018 0.274 4.056 37.140 109.005 151.133 -> 44.416 -> 1065.982 MByte/s p25 random-cyc-1dim p25 method 0 =Sndrcv :( 34.508) 0.029 0.445 6.530 42.775 104.387 126.936 -> 42.784 -> 1026.824 MByte/s p25 method 1 =Alltoal :(364.751) 0.003 0.044 0.694 8.021 54.906 126.936 -> 25.223 -> 605.363 MByte/s p25 method 2 =non-blk :( 55.547) 0.018 0.275 4.087 36.978 106.608 126.936 -> 40.942 -> 982.609 MByte/s p26 random-cyc-1dim p26 method 0 =Sndrcv :( 34.355) 0.029 0.449 6.473 42.563 89.278 120.994 -> 38.136 -> 915.262 MByte/s p26 method 1 =Alltoal :(352.737) 0.003 0.046 0.719 7.915 51.857 120.994 -> 23.869 -> 572.845 MByte/s p26 method 2 =non-blk :( 56.008) 0.018 0.271 4.016 35.899 95.988 120.994 -> 38.257 -> 918.169 MByte/s p27 random-cyc-1dim p27 method 0 =Sndrcv :( 34.419) 0.029 0.450 6.525 44.255 100.653 133.877 -> 43.507 -> 1044.175 MByte/s p27 method 1 =Alltoal :(364.453) 0.003 0.044 0.685 8.165 57.836 133.877 -> 25.779 -> 618.698 MByte/s p27 method 2 =non-blk :( 56.336) 0.018 0.274 4.043 37.169 109.175 133.877 -> 42.730 -> 1025.520 MByte/s p28 random-cyc-1dim p28 method 0 =Sndrcv :( 34.355) 0.029 0.454 6.579 44.255 98.271 135.825 -> 43.219 -> 1037.257 MByte/s p28 method 1 =Alltoal :(380.620) 0.003 0.042 0.671 8.097 57.513 135.825 -> 25.929 -> 622.296 MByte/s p28 method 2 =non-blk :( 55.780) 0.018 0.279 4.044 36.769 105.119 135.825 -> 43.083 -> 1033.988 MByte/s p29 random-cyc-1dim p29 method 0 =Sndrcv :( 34.552) 0.029 0.451 6.568 41.102 88.967 110.364 -> 38.023 -> 912.554 MByte/s p29 method 1 =Alltoal :(359.118) 0.003 0.045 0.696 7.606 49.876 110.364 -> 22.294 -> 535.048 MByte/s p29 method 2 =non-blk :( 56.341) 0.018 0.275 4.058 34.517 88.548 110.364 -> 35.779 -> 858.688 MByte/s p30 random-cyc-1dim p30 method 0 =Sndrcv :( 34.782) 0.029 0.440 6.388 40.630 86.644 113.720 -> 37.775 -> 906.599 MByte/s p30 method 1 =Alltoal :(356.868) 0.003 0.045 0.706 7.611 49.129 113.720 -> 22.728 -> 545.472 MByte/s p30 method 2 =non-blk :( 57.796) 0.017 0.269 4.012 34.614 91.138 113.720 -> 36.552 -> 877.247 MByte/s p31 random-cyc-1dim p31 method 0 =Sndrcv :( 35.056) 0.029 0.440 6.416 39.154 80.918 100.165 -> 34.828 -> 835.879 MByte/s p31 method 1 =Alltoal :(349.134) 0.003 0.045 0.706 7.380 49.669 100.165 -> 20.931 -> 502.336 MByte/s p31 method 2 =non-blk :( 57.446) 0.017 0.271 3.926 33.740 81.012 100.165 -> 33.211 -> 797.061 MByte/s p32 random-cyc-1dim p32 method 0 =Sndrcv :( 34.025) 0.029 0.458 6.623 44.733 99.839 129.010 -> 43.110 -> 1034.632 MByte/s p32 method 1 =Alltoal :(385.001) 0.003 0.041 0.653 8.214 56.068 129.010 -> 25.618 -> 614.833 MByte/s p32 method 2 =non-blk :( 56.189) 0.018 0.277 4.063 36.485 104.305 129.010 -> 42.159 -> 1011.816 MByte/s p33 random-cyc-1dim p33 method 0 =Sndrcv :( 34.647) 0.029 0.449 6.475 41.856 87.648 113.227 -> 38.973 -> 935.340 MByte/s p33 method 1 =Alltoal :(356.123) 0.003 0.045 0.712 7.892 52.908 113.227 -> 23.814 -> 571.546 MByte/s p33 method 2 =non-blk :( 56.908) 0.018 0.272 3.981 35.893 94.459 113.227 -> 38.139 -> 915.331 MByte/s p34 random-cyc-1dim p34 method 0 =Sndrcv :( 34.891) 0.029 0.444 6.463 40.751 79.881 107.787 -> 36.099 -> 866.373 MByte/s p34 method 1 =Alltoal :(369.385) 0.003 0.043 0.684 7.666 46.786 107.787 -> 21.026 -> 504.620 MByte/s p34 method 2 =non-blk :( 57.379) 0.017 0.270 3.996 34.629 91.054 107.787 -> 34.895 -> 837.473 MByte/s p35 random-cyc-1dim p35 method 0 =Sndrcv :( 34.750) 0.029 0.448 6.316 41.049 84.757 107.996 -> 37.388 -> 897.303 MByte/s p35 method 1 =Alltoal :(350.505) 0.003 0.045 0.710 7.658 48.354 107.996 -> 22.597 -> 542.332 MByte/s p35 method 2 =non-blk :( 57.273) 0.017 0.270 3.919 34.969 84.008 107.996 -> 35.389 -> 849.330 MByte/s -- additional patterns that are not used to compute b_eff: p36 worst-cyc-1dim p36 method 0 =Sndrcv :( 34.670) 0.029 0.445 6.588 41.772 94.393 118.276 -> 40.196 -> 964.707 MByte/s p36 method 1 =Alltoal :(343.636) 0.003 0.046 0.691 7.990 53.729 118.276 -> 24.164 -> 579.924 MByte/s p36 method 2 =non-blk :( 55.940) 0.018 0.273 4.072 35.705 97.087 118.276 -> 38.672 -> 928.131 MByte/s p37 best bi-section p37 method 0 =Sndrcv :( 28.689) 0.017 0.254 3.671 27.980 71.826 92.114 -> 29.846 -> 716.298 MByte/s p37 method 1 =Alltoal :(345.277) 0.001 0.023 0.364 4.462 40.348 92.114 -> 18.434 -> 442.413 MByte/s p37 method 2 =non-blk :( 30.078) 0.017 0.248 3.832 31.812 61.588 92.114 -> 29.506 -> 708.142 MByte/s p38 worst bi-section p38 method 0 =Sndrcv :( 28.726) 0.017 0.260 3.712 29.808 94.501 120.736 -> 36.543 -> 877.042 MByte/s p38 method 1 =Alltoal :(341.505) 0.001 0.023 0.366 5.389 46.295 120.736 -> 26.211 -> 629.074 MByte/s p38 method 2 =non-blk :( 30.462) 0.016 0.248 3.857 36.205 99.631 120.736 -> 40.435 -> 970.437 MByte/s p39 one PingPong Pair p39 method 0 =Sndrcv :( 23.077) 0.002 0.027 0.386 3.132 19.547 30.897 -> 8.097 -> 194.330 MByte/s p39 method 1 =Alltoal :(-------) 0.000 0.000 0.000 0.000 0.000 30.897 -> 2.878 -> 69.071 MByte/s p39 method 2 =non-blk :(-------) 0.000 0.000 0.000 0.000 0.000 30.897 -> 2.878 -> 69.071 MByte/s p40 acyclic-2dim-all p40 method 0 =Sndrcv :( 34.437) 0.023 0.354 5.200 33.021 74.201 102.963 -> 33.627 -> 807.038 MByte/s p40 method 1 =Alltoal :(172.888) 0.005 0.071 1.097 11.118 59.903 102.963 -> 25.045 -> 601.071 MByte/s p40 method 2 =non-blk :( 52.504) 0.015 0.233 3.398 28.407 73.081 102.963 -> 31.400 -> 753.601 MByte/s p41 acyclic-3dim-all p41 method 0 =Sndrcv :( 26.165) 0.024 0.372 5.323 38.124 106.360 135.585 -> 42.394 -> 1017.457 MByte/s p41 method 1 =Alltoal :(114.749) 0.006 0.087 1.327 12.419 79.092 135.585 -> 32.369 -> 776.865 MByte/s p41 method 2 =non-blk :( 35.992) 0.018 0.277 3.912 36.464 118.774 135.585 -> 44.752 -> 1074.058 MByte/s p42 cyclic-2dim-x p42 method 0 =Sndrcv :( 34.875) 0.029 0.433 6.280 37.685 74.740 112.575 -> 34.574 -> 829.781 MByte/s p42 method 1 =Alltoal :(342.503) 0.003 0.046 0.693 6.964 45.888 112.575 -> 21.846 -> 524.301 MByte/s p42 method 2 =non-blk :( 58.032) 0.017 0.265 4.016 32.490 68.413 112.575 -> 32.527 -> 780.656 MByte/s p43 cyclic-2dim-y p43 method 0 =Sndrcv :( 34.331) 0.029 0.451 6.613 42.455 106.082 150.968 -> 46.053 -> 1105.282 MByte/s p43 method 1 =Alltoal :(344.262) 0.003 0.046 0.717 7.592 58.163 150.968 -> 28.810 -> 691.432 MByte/s p43 method 2 =non-blk :( 56.422) 0.018 0.275 4.045 35.402 92.336 150.968 -> 42.325 -> 1015.791 MByte/s p44 cyclic-2dim-all p44 method 0 =Sndrcv :( 34.645) 0.029 0.442 6.414 40.327 82.669 124.721 -> 38.997 -> 935.936 MByte/s p44 method 1 =Alltoal :(173.019) 0.006 0.091 1.370 13.245 67.645 124.721 -> 28.906 -> 693.743 MByte/s p44 method 2 =non-blk :( 53.148) 0.019 0.291 4.212 34.840 85.316 124.721 -> 36.483 -> 875.591 MByte/s p45 cyclic-3dim-x p45 method 0 =Sndrcv :( 16.867) 0.059 0.946 11.706 113.150 383.074 465.671 -> 154.725 -> 3713.392 MByte/s p45 method 1 =Alltoal :(343.129) 0.003 0.046 0.729 10.653 68.198 465.671 -> 61.276 -> 1470.623 MByte/s p45 method 2 =non-blk :( 37.297) 0.027 0.430 6.189 75.408 386.778 465.671 -> 144.026 -> 3456.618 MByte/s p46 cyclic-3dim-y p46 method 0 =Sndrcv :( 35.322) 0.028 0.422 6.336 37.297 75.240 105.320 -> 32.697 -> 784.728 MByte/s p46 method 1 =Alltoal :(344.698) 0.003 0.046 0.714 8.029 52.526 105.320 -> 23.040 -> 552.963 MByte/s p46 method 2 =non-blk :( 57.704) 0.017 0.263 3.866 31.736 64.861 105.320 -> 30.478 -> 731.475 MByte/s p47 cyclic-3dim-z p47 method 0 =Sndrcv :( 34.952) 0.029 0.437 6.444 37.012 68.161 93.263 -> 30.344 -> 728.250 MByte/s p47 method 1 =Alltoal :(682.265) 0.001 0.023 0.365 4.452 38.754 93.263 -> 18.494 -> 443.867 MByte/s p47 method 2 =non-blk :( 62.289) 0.016 0.248 3.825 31.132 64.926 93.263 -> 29.301 -> 703.226 MByte/s p48 cyclic-3dim-all p48 method 0 =Sndrcv :( 27.434) 0.036 0.556 7.939 51.902 115.119 143.602 -> 47.824 -> 1147.765 MByte/s p48 method 1 =Alltoal :(138.767) 0.007 0.113 1.714 15.689 91.141 143.602 -> 36.410 -> 873.846 MByte/s p48 method 2 =non-blk :( 46.069) 0.022 0.332 4.787 43.895 119.324 143.602 -> 46.044 -> 1105.062 MByte/s log_avg of all rings - ring, method 0 = Sndrcv : 0.029 0.441 6.486 37.997 76.179 110.275 || 35.074 -> 841.788 MByte/s - ring, method 1 = Alltoal: 0.003 0.041 0.632 6.461 46.497 110.275 || 21.792 -> 522.997 MByte/s - ring, method 2 = non-blk: 0.017 0.268 3.986 33.506 73.630 110.275 || 32.919 -> 790.060 MByte/s log_avg of all random - random, method 0 = Sndrcv : 0.029 0.451 6.544 42.268 93.170 120.245 || 40.134 -> 963.213 MByte/s - random, method 1 = Alltoal: 0.003 0.044 0.691 7.832 51.883 120.245 || 23.746 -> 569.902 MByte/s - random, method 2 = non-blk: 0.018 0.274 4.041 35.738 95.887 120.245 || 38.527 -> 924.641 MByte/s log_avg(ring,random) - average, method 0 = Sndrcv : 0.029 0.446 6.515 40.076 84.248 115.152 || 37.519 -> 900.456 MByte/s - average, method 1 = Alltoal: 0.003 0.043 0.661 7.114 49.116 115.152 || 22.748 -> 545.946 MByte/s - average, method 2 = non-blk: 0.018 0.271 4.013 34.604 84.025 115.152 || 35.613 -> 854.706 MByte/s * size -> accumulated on all processes: - accumulated, mthd 0 = Sndrcv : 0.693 10.706 156.354 961.819 2021.942 2763.652 || 900.456 MByte/s - accumulated, mthd 1 = Alltoal: 0.064 1.021 15.857 170.726 1178.780 2763.652 || 545.946 MByte/s - accumulated, mthd 2 = non-blk: 0.421 6.505 96.319 830.493 2016.594 2763.652 || 854.706 MByte/s SECTION-BY-MTHD-MSGLNG-END SECTION-BY-MSGLNG-BEGIN 5th analysis path only as information: logavg_pat ( max_mthd ( max_rep ( b(L) ) ) ) and for all methods: logavg_pat ( max_rep ( b(L) ) ) msg length|accumulated| effective bandwidth per process: | effective| average rings random method_0 method_1 method_2 | bandwidth| crt,rnd only only Sndrcv Alltoal non-blk 1 0.693 0.029 0.029 0.029 0.029 0.003 0.018 2 1.386 0.058 0.057 0.058 0.058 0.005 0.035 4 2.736 0.114 0.113 0.115 0.114 0.011 0.069 8 5.552 0.231 0.230 0.233 0.231 0.021 0.141 16 10.706 0.446 0.441 0.451 0.446 0.043 0.271 32 21.314 0.888 0.881 0.896 0.888 0.085 0.538 64 42.175 1.757 1.746 1.769 1.757 0.169 1.069 128 81.565 3.399 3.387 3.410 3.399 0.333 2.098 256 156.354 6.515 6.486 6.544 6.515 0.661 4.013 512 310.708 12.946 12.825 13.069 12.946 1.317 8.025 1024 610.007 25.417 25.361 25.473 25.417 2.612 15.936 2048 651.758 27.157 26.161 28.191 27.157 4.144 22.096 4096 961.819 40.076 37.997 42.268 40.076 7.114 34.604 10624 1056.600 44.025 40.317 48.073 41.540 13.234 43.883 27554 1536.115 64.005 56.329 72.726 56.987 22.928 64.005 71468 1555.841 64.827 58.535 71.795 63.810 31.675 63.484 185364 2058.942 85.789 76.341 96.406 84.248 49.116 84.025 480774 2135.888 88.995 81.238 97.494 88.688 57.484 79.697 1246974 2567.824 106.993 102.154 112.061 106.069 59.953 96.013 3234251 2659.387 110.808 105.944 115.895 110.808 110.808 110.808 8388608 2763.652 115.152 110.275 120.245 115.152 115.152 115.152 SECTION-BY-MSGLNG-END SECTION-BY-PATTERN-MSGLNG-BEGIN official (1st) analysis path -- see last column: (only columns of some message lengths are printed) logavg_pat ( avg_L ( max_mthd ( max_rep ( b(L) ) ) ) ) pattern / msg-length: 1 16 256 4096 185364 8388608 -> average -> accumulated (latency of one sendrecv (or equivalent method) in microsec) p00 ring-12*2fix :( 35.378) 0.028 0.435 6.448 36.798 69.302 90.705 -> 31.769 -> 762.452 MByte/s p01 ring-6*4fix :( 34.476) 0.029 0.447 6.569 42.092 97.259 147.356 -> 44.466 -> 1067.190 MByte/s p02 ring-3*8fix :( 35.173) 0.028 0.445 6.445 37.172 68.902 101.649 -> 33.362 -> 800.690 MByte/s p03 ring-1*24fix :( 35.070) 0.029 0.440 6.453 37.105 73.466 110.339 -> 34.676 -> 832.231 MByte/s p04 ring-1*24fix :( 34.972) 0.029 0.439 6.524 37.680 75.428 108.704 -> 35.288 -> 846.908 MByte/s p05 ring-1*24fix :( 35.068) 0.029 0.441 6.478 37.387 76.918 110.352 -> 35.437 -> 850.476 MByte/s p06 random-cyc-1dim :( 34.903) 0.029 0.439 6.499 38.364 76.712 99.885 -> 34.262 -> 822.291 MByte/s p07 random-cyc-1dim :( 34.424) 0.029 0.457 6.591 43.751 108.139 136.513 -> 44.776 -> 1074.621 MByte/s p08 random-cyc-1dim :( 34.307) 0.029 0.446 6.628 42.575 97.548 121.322 -> 41.823 -> 1003.762 MByte/s p09 random-cyc-1dim :( 34.317) 0.029 0.459 6.586 44.781 120.523 151.379 -> 49.530 -> 1188.729 MByte/s p10 random-cyc-1dim :( 34.774) 0.029 0.443 6.547 41.143 94.201 120.683 -> 40.334 -> 968.016 MByte/s p11 random-cyc-1dim :( 34.728) 0.029 0.450 6.509 41.739 91.781 116.965 -> 39.786 -> 954.865 MByte/s p12 random-cyc-1dim :( 34.629) 0.029 0.444 6.490 41.443 89.047 111.176 -> 38.433 -> 922.399 MByte/s p13 random-cyc-1dim :( 29.649) 0.034 0.494 7.120 44.249 93.630 120.467 -> 40.209 -> 965.009 MByte/s p14 random-cyc-1dim :( 34.730) 0.029 0.446 6.466 41.881 94.791 119.603 -> 40.967 -> 983.199 MByte/s p15 random-cyc-1dim :( 34.453) 0.029 0.451 6.564 42.935 98.082 121.375 -> 41.597 -> 998.333 MByte/s p16 random-cyc-1dim :( 34.775) 0.029 0.441 6.306 40.946 89.983 111.988 -> 38.215 -> 917.150 MByte/s p17 random-cyc-1dim :( 34.348) 0.029 0.458 6.635 43.087 105.127 127.077 -> 42.714 -> 1025.127 MByte/s p18 random-cyc-1dim :( 34.657) 0.029 0.453 6.464 42.850 101.368 121.022 -> 42.395 -> 1017.478 MByte/s p19 random-cyc-1dim :( 34.673) 0.029 0.454 6.592 41.443 89.830 110.360 -> 38.305 -> 919.316 MByte/s p20 random-cyc-1dim :( 34.165) 0.029 0.457 6.525 42.304 94.967 121.116 -> 40.772 -> 978.522 MByte/s p21 random-cyc-1dim :( 32.617) 0.031 0.453 6.665 43.865 98.050 124.583 -> 42.052 -> 1009.240 MByte/s p22 random-cyc-1dim :( 34.375) 0.029 0.453 6.567 40.942 91.374 111.939 -> 38.534 -> 924.808 MByte/s p23 random-cyc-1dim :( 34.302) 0.029 0.452 6.647 43.136 101.354 126.768 -> 43.127 -> 1035.050 MByte/s p24 random-cyc-1dim :( 34.072) 0.029 0.451 6.593 44.388 117.125 151.133 -> 48.044 -> 1153.060 MByte/s p25 random-cyc-1dim :( 34.508) 0.029 0.445 6.530 42.775 106.608 126.936 -> 43.708 -> 1048.985 MByte/s p26 random-cyc-1dim :( 34.355) 0.029 0.449 6.473 42.563 95.988 120.994 -> 39.770 -> 954.477 MByte/s p27 random-cyc-1dim :( 34.419) 0.029 0.450 6.525 44.255 109.175 133.877 -> 44.667 -> 1072.010 MByte/s p28 random-cyc-1dim :( 34.355) 0.029 0.454 6.579 44.255 105.119 135.825 -> 44.667 -> 1072.020 MByte/s p29 random-cyc-1dim :( 34.552) 0.029 0.451 6.568 41.102 88.967 110.364 -> 38.354 -> 920.501 MByte/s p30 random-cyc-1dim :( 34.782) 0.029 0.440 6.388 40.630 91.138 113.720 -> 38.526 -> 924.633 MByte/s p31 random-cyc-1dim :( 35.056) 0.029 0.440 6.416 39.154 81.012 100.165 -> 35.255 -> 846.116 MByte/s p32 random-cyc-1dim :( 34.025) 0.029 0.458 6.623 44.733 104.305 129.010 -> 44.011 -> 1056.275 MByte/s p33 random-cyc-1dim :( 34.647) 0.029 0.449 6.475 41.856 94.459 113.227 -> 39.775 -> 954.602 MByte/s p34 random-cyc-1dim :( 34.891) 0.029 0.444 6.463 40.751 91.054 107.787 -> 37.094 -> 890.254 MByte/s p35 random-cyc-1dim :( 34.750) 0.029 0.448 6.316 41.049 84.757 107.996 -> 37.762 -> 906.295 MByte/s -- additional patterns that are not used to compute b_eff: p36 worst-cyc-1dim :( 34.670) 0.029 0.445 6.588 41.772 97.087 118.276 -> 41.074 -> 985.786 MByte/s p37 best bi-section :( 28.689) 0.017 0.254 3.832 31.812 71.826 92.114 -> 30.620 -> 734.870 MByte/s p38 worst bi-section :( 28.726) 0.017 0.260 3.857 36.205 99.631 120.736 -> 40.437 -> 970.488 MByte/s p39 one PingPong Pair :( 23.077) 0.002 0.027 0.386 3.132 19.547 30.897 -> 8.097 -> 194.330 MByte/s p40 acyclic-2dim-all :( 34.437) 0.023 0.354 5.200 33.021 74.201 102.963 -> 33.842 -> 812.218 MByte/s p41 acyclic-3dim-all :( 26.165) 0.024 0.372 5.323 38.124 118.774 135.585 -> 45.463 -> 1091.104 MByte/s p42 cyclic-2dim-x :( 34.875) 0.029 0.433 6.280 37.685 74.740 112.575 -> 35.050 -> 841.204 MByte/s p43 cyclic-2dim-y :( 34.331) 0.029 0.451 6.613 42.455 106.082 150.968 -> 46.456 -> 1114.933 MByte/s p44 cyclic-2dim-all :( 34.645) 0.029 0.442 6.414 40.327 85.316 124.721 -> 39.447 -> 946.738 MByte/s p45 cyclic-3dim-x :( 16.867) 0.059 0.946 11.706 113.150 386.778 465.671 -> 155.681 -> 3736.332 MByte/s p46 cyclic-3dim-y :( 35.322) 0.028 0.422 6.336 37.297 75.240 105.320 -> 33.456 -> 802.951 MByte/s p47 cyclic-3dim-z :( 34.952) 0.029 0.437 6.444 37.012 68.161 93.263 -> 31.302 -> 751.245 MByte/s p48 cyclic-3dim-all :( 27.434) 0.036 0.556 7.939 51.902 119.324 143.602 -> 50.131 -> 1203.137 MByte/s log_avg of all rings : 0.029 0.441 6.486 37.997 76.341 110.275 || 35.623 -> 854.956 MByte/s log_avg of all random : 0.029 0.451 6.544 42.268 96.406 120.245 || 40.845 -> 980.285 MByte/s log_avg(ring,random) : 0.029 0.446 6.515 40.076 85.789 115.152 || 38.145 -> 915.478 MByte/s * size -> accumulated on all pr.: 0.693 10.706 156.354 961.819 2058.942 2763.652 || 915.478 MByte/s SECTION-BY-PATTERN-MSGLNG-END SECTION-BEFF-BEGIN The effective bandwidth is b_eff = 915.478 MByte/s on 24 processes ( = 38.145 MByte/s * 24 processes) Ping-pong latency: 23.077 microsec Ping-pong bandwidth: 741.535 MByte/s at Lmax= 8.000 MByte (MByte/s=1e6 Byte/s) (MByte=2**20 Byte) system parameters : 24 nodes, 1024 MB/node system name : HI-UX/MPP hostname : himiko OS release : 03-00 OS version : ad2b0 machine : SR8000 Date of measurement: Mon Nov 29 13:36:35 1999 Total execution wall clock time = 132 seconds SECTION-BEFF-END b_eff = 915.478 MB/s = 38.145 * 24 PEs with 1024 MB/PE on HI-UX/MPP himiko 03-00 ad2b0 SR8000