================================================================================================
BYTE_STREAM_SPLIT INT32
================================================================================================

OpenJDK 64-Bit Server VM 21.0.12+8-LTS on Linux 6.17.0-1020-azure
AMD EPYC 7763 64-Core Processor
BYTE_STREAM_SPLIT INT32:                  Best Time(ms)   Avg Time(ms)   Stdev(ms)    Rate(M/s)   Per Row(ns)   Relative
------------------------------------------------------------------------------------------------------------------------
Spark vectorized readIntegers                         2              2           0        532.5           1.9       1.0X
Spark vectorized readIntegersAsLongs                  2              3           1        430.2           2.3       0.8X
Spark vectorized readIntegersAsDoubles                2              2           1        476.2           2.1       0.9X
parquet-mr readInteger (per-value)                    9              9           1        122.7           8.1       0.2X


================================================================================================
BYTE_STREAM_SPLIT INT64
================================================================================================

OpenJDK 64-Bit Server VM 21.0.12+8-LTS on Linux 6.17.0-1020-azure
AMD EPYC 7763 64-Core Processor
BYTE_STREAM_SPLIT INT64:                  Best Time(ms)   Avg Time(ms)   Stdev(ms)    Rate(M/s)   Per Row(ns)   Relative
------------------------------------------------------------------------------------------------------------------------
Spark vectorized readLongs                            6              6           1        171.5           5.8       1.0X
Spark vectorized readLongsAsInts                      6              6           1        171.7           5.8       1.0X
parquet-mr readLong (per-value)                      13             13           1         82.2          12.2       0.5X


================================================================================================
BYTE_STREAM_SPLIT FLOAT
================================================================================================

OpenJDK 64-Bit Server VM 21.0.12+8-LTS on Linux 6.17.0-1020-azure
AMD EPYC 7763 64-Core Processor
BYTE_STREAM_SPLIT FLOAT:                  Best Time(ms)   Avg Time(ms)   Stdev(ms)    Rate(M/s)   Per Row(ns)   Relative
------------------------------------------------------------------------------------------------------------------------
Spark vectorized readFloats                           2              2           1        527.7           1.9       1.0X
Spark vectorized readFloatsAsDoubles                  2              2           1        459.2           2.2       0.9X
parquet-mr readFloat (per-value)                      8              9           0        123.7           8.1       0.2X


================================================================================================
BYTE_STREAM_SPLIT DOUBLE
================================================================================================

OpenJDK 64-Bit Server VM 21.0.12+8-LTS on Linux 6.17.0-1020-azure
AMD EPYC 7763 64-Core Processor
BYTE_STREAM_SPLIT DOUBLE:                 Best Time(ms)   Avg Time(ms)   Stdev(ms)    Rate(M/s)   Per Row(ns)   Relative
------------------------------------------------------------------------------------------------------------------------
Spark vectorized readDoubles                          6              6           1        171.4           5.8       1.0X
parquet-mr readDouble (per-value)                    13             13           1         82.0          12.2       0.5X


================================================================================================
BYTE_STREAM_SPLIT FLBA
================================================================================================

OpenJDK 64-Bit Server VM 21.0.12+8-LTS on Linux 6.17.0-1020-azure
AMD EPYC 7763 64-Core Processor
BYTE_STREAM_SPLIT FLBA:                                  Best Time(ms)   Avg Time(ms)   Stdev(ms)    Rate(M/s)   Per Row(ns)   Relative
---------------------------------------------------------------------------------------------------------------------------------------
Spark vectorized readBinary (width=12)                              24             26           2         43.0          23.2       1.0X
Spark per-value readBinary (old updater path, width=12)             29             29           1         36.3          27.5       0.8X
parquet-mr readBinary (per-value, width=12)                         37             38           2         28.3          35.3       0.7X