ELF>/@@ overhead.cstderrH5P_CLS_DATASET_XFER_gmain__intel_new_feature_proc_initdisplay_error_cbH5Eset_auto2H5openH5PcreateH5Pget_btree_ratiosH5Pclose__ctype_b_locstrtodfprintffputsexittestgetenvremoveprintfH5T_NATIVE_INT_gH5P_CLS_DATASET_CREATE_gH5P_CLS_FILE_ACCESS_gH5Pget_cacheH5Pset_cacheH5check_versionH5FcreateH5Pset_chunkH5Pset_btree_ratiosH5Screate_simpleH5Dcreate2opencallocHDrandH5Sselect_hyperslabH5DwriteH5Fflush__fxstatfreeH5DcloseH5ScloseH5FclosecloseabortstdoutputsH5Eprint2M @@  *@/^kr|(5EO\pUHHATAUAVAWSHXAHt$(3$E3Dt$ 3 $@3$Mc=ÅHt$HT$HL$iZHD$(E3L`IHD$0DM$LǾ:utW:VuHHu3u ALǾ:utW:VuHHu3u ALǾ:utW:VuHHu3uAtLǾ:utW:VuHHu3uA9LǾ:utW:VuHHu3uAHII;LǾ:utW:VuHHu3u@I}SI$HT$0LH2FNAt/L3 $ $ġ{DIHII;2<.tHD$(H=H3H5Dt$ AEu@Ht$3ɋ~~YH5HX[A_A^A]A\H]HD$(H=H3H5xHt3HX[A_A^A]A\H]ÿDt$ A΃|$ H33Ht$3DAǿHt$3DAſHt$3DAĿHt$3DÿHt$3DEEDADATAUAVAWSUHAEE3AD$$H4$$D$D$H$HDŽ$HDŽ$H$D$H$=ÅEu<3҉3H$E33҉3W$q 3҉AE>=AE"DH$=AEH$DX(PH$H$$$H$H$$jHDE3E$$$H$!33$$$uH$HH$HDŽ$HD$D$D$AEB$$LcJt>t‰ЙLcJt>uL$HHDKDtHcH+A} AAAEHEH$>Ht HH$(HH+H$H+H$ HH$3D3H$E3L$~DE苼$L$AQ`5M$u H$nD3'H$ H$WLL$0W*˿HL+H*^HH$H$H;6D$Ht H3틼$$$DDD$'$wR$$HDŽ$6HDŽ$(HDŽ$HDŽ$ HDŽ$H$}n$$$DDDHtH3$H][A_A^A]A\H$WHT$0W*ɿHH+и*^H$3$3H][A_A^A]A\D$@SH53[Ð*** ERRORS DETECTED *** forwardreverseinwardoutwardrandomcacheusage: %s [STYLE|cache] [LEFT [MIDDLE [RIGHT]]] HDF5_NOCLEANUPoverhead.h5%-7s %8s StyleBytes/Chunk----------------chunked%4lu %8.3f *** %-7s %8.3f *FAILED* STYLE is the order that the dataset is filled and should be one of: forward -- Fill the dataset from lowest address to highest address. This style tests the right split ratio. reverse -- Fill the dataset from highest address to lowest address. This is the reverse order of `forward' and tests the left split ratio. inward -- Fill beginning at both the lowest and highest addresses and work in toward the center of the dataset. This tests the middle split ratio. outward -- Start at the center of the dataset and work outward toward the lowest and highest addresses. This tests both left and right split ratios. random -- Write the chunks of the dataset in random order. This tests all split ratios. If no fill style is specified then all fill styles are tried and a single value is printed for each one. If the word `cache' is used instead of a fill style then the raw data cache is enabled. It is not possible to enable the raw data cache when a specific fill style is used because H5Fflush() is called after each chunk is written in order to calculate overhead during the test. If the cache is enabled then chunks are written to disk in different orders than the actual H5Dwrite() calls in the test due to collisions and the resulting B-tree will be split differently. LEFT, MIDDLE, and RIGHT are the ratios to use for splitting and should be values between zero and one, inclusive. STYLE is the order that the dataset is filled and should be one of: forward -- Fill the dataset from lowest address to highest address. This style tests the right split ratio. reverse -- Fill the dataset from highest address to lowest address. This is the reverse order of `forward' and tests the left split ratio. inward -- Fill beginning at both the lowest and highest addresses and work in toward the center of the dataset. This tests the middle split ratio. outward -- Start at the center of the dataset and work outward toward the lowest and highest addresses. This tests both left and right split ratios. random -- Write the chunks of the dataset in random order. This tests all split ratios. If no fill style is specified then all fill styles are tried and a single value is printed for each one. If the word `cache' is used instead of a fill style then the raw data cache is enabled. It is not possible to enable the raw data cache when a specific fill style is used because H5Fflush() is called after each chunk is written in order to calculate overhead during the test. If the cache is enabled then chunks are written to disk in different orders than the actual H5Dwrite() calls in the test due to collisions and the resulting B-tree will be split differently. LEFT, MIDDLE, and RIGHT are the ratios to use for splitting and should be values between zero and one, inclusive. x L@AC Q8  " 8  " 8  "8  "8  "ABBBF A 8  " 8  " 8  "8  "8  "FABBBF A 8  " 8  " 8  "8  "8  "l@BBB B(A0A8Gr8A0A(B BBBA\8A0A(B BBBA$  A\A.shstrtab.strtab.symtab.data.bss.text.rodata.str1.4.rodata.str1.32.rodata.note.GNU-stack.eh_frame.rela.text.rela.rodata.rela.eh_frame%2 OTZ _  $F , 4 <  Do L     " L) 38 ? DN Sq v      & !4"9 G#X^c$ %&&,8 DW'u () *+,<-X. /011  25 0A <O 4] ,k $y  . 0 1 1   2 / 3 D R 3k 4  6 5 7j  K( 0j 8q @c HU PG X9 p@ @P@ !& ,20;2@@ K `S c 0#m# xp. /H