Create Training Set

Train application_name system_name processors num_mpi num_omppn input frequency executable_name scenario frequency papi_tot_cyc papi_tot_ins papi_l1_tcm papi_l2_tcm papi_fml_ins papi_br_tkn papi_l2_ldm papi_l2_stm papi_fdv_ins papi_br_cn papi_tot_iis papi_ca_shr papi_ca_cln papi_ca_itv papi_br_ntk papi_br_msp papi_fp_ins papi_res_stl papi_l1_ldm papi_l1_stm papi_l2_tcw papi_l2_dca papi_l2_dcr papi_l2_dcw papi_l1_icm papi_br_ins papi_l1_dch papi_l1_dca papi_l1_dcm papi_l1_tca papi_l1_ica papi_fp_ops papi_l2_ica papi_tlb_dm papi_vec_ins papi_l2_dcm papi_l2_icm papi_l2_tch papi_l2_tca papi_ld_ins papi_sr_ins papi_tlb_im papi_hw_int papi_l3_tcm nruntime power_sys power_cpu power_mem num_mpi num_omppn scaling_factor
Horovod CANDLE P1B2 Benchmark ANL Theta 8 8 64 learning rate=0.001, batch size=50.0, epochs=384.0 1.301 p1b2-hybrid.py empirical 1.301 1 0.2194627787663067 0.0207273512644837 0.0166780991114403 0.0 0.0109096782919339 0.0138327175341051 0.0 0.0 0.0310046503788925 0.0 0.0 0.0 0.0 0.0200330020614886 0.002276267319605 0.0 0.2449978516098871 0.0162141134855922 0.0 0.0 0.0 0.0 0.0 0.0044790134573959 0.0408182334252496 0.0 0.1003401148749862 0.0162141134855922 0.0 0.1380417823731509 0.0 0.0 0.0150030205929445 0.0 0.0 0.0 0.0141343332338468 0.0308124323452871 0.0638833856054952 0.0364241607858809 0.0 0.0 0.0 8.75919882492e-09 133.426874 85.98702 14.300914 8 64 312171455820.900802130801279553
Horovod CANDLE P1B2 Benchmark ANL Theta 12 12 64 learning rate=0.001, batch size=50.0, epochs=384.0 1.301 p1b2-hybrid.py empirical 1.301 1 0.2426823994770412 0.0190495175834377 0.0141003490198952 0.0 0.0124533112081076 0.0114580524756682 0.0 0.0 0.0360253331629272 0.0 0.0 0.0 0.0 0.0235472826626525 0.0025594389885583 0.0 0.2878810057973457 0.0140170322355168 0.0 0.0 0.0 0.0 0.0 0.0050212330636835 0.0476203372983171 0.0 0.1082478928113239 0.0140170322355168 0.0 0.1583058033845296 0.0 0.0 0.0129561413878193 0.0 0.0 0.0 0.013675497734571 0.0277758467544662 0.0700876650955817 0.0376665321669854 0.0 0.0 0.0 8.80876886101e-09 120.563788 69.623411 15.267207 12 64 258914064494.876165346467618972
Horovod CANDLE P1B2 Benchmark ANL Theta 16 16 64 learning rate=0.001, batch size=50.0, epochs=384.0 1.301 p1b2-hybrid.py empirical 1.301 1 0.2593570356364023 0.0180243771662356 0.012297247949993 0.0 0.0136218016281658 0.0099076615773893 0.0 0.0 0.039710979954666 0.0 0.0 0.0 0.0 0.0261689477708203 0.0027467345254142 0.0 0.3151057518339384 0.0126363305146475 0.0 0.0 0.0 0.0 0.0 0.0053600401214308 0.0526099677735305 0.0 0.1145949551218172 0.0126363305146475 0.0 0.1714795628963123 0.0 0.0 0.011368848979376 0.0 0.0 0.0 0.0134104055637006 0.0257076535136935 0.0749955182871708 0.0388578997719516 0.0 0.0 0.0 6.54089670723e-09 136.69695 87.871021 12.989721 16 64 230481700060.11825696090644608
Horovod CANDLE P1B2 Benchmark ANL Theta 24 24 64 learning rate=0.001, batch size=50.0, epochs=384.0 1.301 p1b2-hybrid.py empirical 1.301 1 0.2770970885649838 0.016329998566187 0.0100490081333866 0.0 0.0146815912078139 0.0078359177418873 0.0 0.0 0.0438325302586155 0.0 0.0 0.0 0.0 0.02911226582378 0.0030505707586547 0.0 0.3558659169088086 0.0105395597713707 0.0 0.0 0.0 0.0 0.0 0.0057105170954645 0.0577790424133318 0.0 0.120154831615424 0.0105395597713707 0.0 0.1867475414739725 0.0 0.0 0.0095103036336559 0.0 0.0 0.0 0.0130221923913283 0.0230712005247149 0.0801906421687345 0.0398246193646737 0.0 0.0 0.0 6.28143152986e-09 128.276845 77.819211 13.641156 24 64 203799921548.859418836463910104
Horovod CANDLE P1B2 Benchmark ANL Theta 48 48 64 learning rate=0.001, batch size=50.0, epochs=384.0 1.301 p1b2-hybrid.py empirical 1.301 1 0.3052075095946454 0.0146958835184862 0.0071666654288275 0.0 0.0167088469198685 0.0052441245379145 0.0 0.0 0.0500624981848328 0.0 0.0 0.0 0.0 0.0335194276228382 0.0033799739570605 0.0 0.4051247968494543 0.0083068024746309 0.0 0.0 0.0 0.0 0.0 0.0063510651346302 0.0668485696979414 0.0 0.1299314369692822 0.0083068024746309 0.0 0.2097524542615433 0.0 0.0 0.0072725837493035 0.0 0.0 0.0 0.0126562638704207 0.0198229292992482 0.0889311425912258 0.0419238582510361 0.0 0.0 0.0 4.22600160459e-09 136.721733 84.648612 14.98849 48 64 174790403580.943283969289156583
Horovod CANDLE P1B2 Benchmark ANL Theta 64 64 64 learning rate=0.001, batch size=50.0, epochs=384.0 1.301 p1b2-hybrid.py empirical 1.301 1 0.3136173202944304 0.0140552650971639 0.0062472144840848 0.0 0.0171222423808436 0.0044298453831607 0.0 0.0 0.0519681382619688 0.0 0.0 0.0 0.0 0.0349281389798281 0.0035100362140579 0.0 0.422394599411545 0.0075448783723362 0.0 0.0 0.0 0.0 0.0 0.0065512144162128 0.0687579883432039 0.0 0.1338183511672328 0.0075448783723362 0.0 0.2180319560413266 0.0 0.0 0.0064761724151607 0.0 0.0 0.0 0.0125116516947119 0.0187588661787966 0.0905559280775366 0.0421284645847172 0.0 0.0 0.0 3.25832178503e-09 128.341328 75.860701 13.804428 64 64 166272373554.109183477523452919
Horovod CANDLE P1B2 Benchmark ANL Theta 96 96 64 learning rate=0.001, batch size=50.0, epochs=384.0 1.301 p1b2-hybrid.py empirical 1.301 1 0.3234951913214683 0.0134555563059862 0.0053498044594349 0.0 0.0174680376104916 0.0035608507991122 0.0 0.0 0.0540983823721492 0.0 0.0 0.0 0.0 0.0363859229991721 0.0035644633335437 0.0 0.4380998973137833 0.0067362499265048 0.0 0.0 0.0 0.0 0.0 0.0067212708798013 0.070548413785522 0.0 0.1364974094380007 0.0067362499265048 0.0 0.2272615875438861 0.0 0.0 0.0056503063533451 0.0 0.0 0.0 0.0123610554707085 0.0177108599301434 0.0930775752079069 0.0426227974420666 0.0 0.0 0.0 3.104217808e-09 127.664315 81.366935 11.989919 96 64 159693760445.046709170866273182
Horovod CANDLE P1B2 Benchmark ANL Theta 128 128 64 learning rate=0.001, batch size=50.0, epochs=384.0 1.301 p1b2-hybrid.py empirical 1.301 1 0.3219513293564553 0.0130881969316108 0.0048002716603894 0.0 0.017924760095704 0.0031137124357843 0.0 0.0 0.0540853998007578 0.0 0.0 0.0 0.0 0.0363986710984225 0.0036490096980751 0.0 0.4452324500131826 0.0062790369956747 0.0 0.0 0.0 0.0 0.0 0.0068640918834467 0.0725759167911072 0.0 0.1385196189354937 0.0062790369956747 0.0 0.2295601618613685 0.0 0.0 0.0052394796633953 0.0 0.0 0.0 0.0122594006319852 0.0170596722923746 0.0944656762662969 0.0428389080651211 0.0 0.0 0.0 2.30131391146e-09 128.51087 79.807065 12.452446 128 64 159694213018.877745797155717508
Horovod CANDLE P1B2 Benchmark ANL Theta 192 192 64 learning rate=0.001, batch size=50.0, epochs=384.0 1.301 p1b2-hybrid.py empirical 1.301 1 0.3290357812782111 0.0129559612926375 0.0042125278945482 0.0 0.0178922462576706 0.0026351883624212 0.0 0.0 0.0556288687432379 0.0 0.0 0.0 0.0 0.0377062404457433 0.0037265944240215 0.0 0.4544376724544207 0.0059657890999998 0.0 0.0 0.0 0.0 0.0 0.006987113852202 0.0736935362281575 0.0 0.1376183922716386 0.0059657890999998 0.0 0.2317540339811728 0.0 0.0 0.0048625151328294 0.0 0.0 0.0 0.0120622891640699 0.0162748170586181 0.0947702241575708 0.0430673355268176 0.0 0.0 0.0 1.91407104427e-09 125.355556 77.782906 11.979487 192 64 153098599906.860812437611683886
Horovod CANDLE P1B2 Benchmark ANL Theta 384 384 64 learning rate=0.001, batch size=50.0, epochs=384.0 1.301 p1b2-hybrid.py empirical 1.301 1 0.3312805160334313 0.0124356582896457 0.0037224548378301 0.0 0.018171400649179 0.0021246283006387 0.0 0.0 0.0563044916065636 0.0 0.0 0.0 0.0 0.0382209389044284 0.0038297986958417 0.0 0.4678319735742594 0.0054046969440388 0.0 0.0 0.0 0.0 0.0 0.0070743142803516 0.0750986316751332 0.0 0.1398920746013438 0.0054046969440388 0.0 0.2358841807094143 0.0 0.0 0.0044629450521462 0.0 0.0 0.0 0.0120248004285654 0.0157472552663955 0.0963966269540223 0.0434952754863205 0.0 0.0 0.0 1.54237240351e-09 129.045064 82.553648 12.783262 384 64 151094234744.902875625491552206