[foreign-memaccess+abi] RFR: Add benchmark

[Per Minborg]

On Fri, 16 Dec 2022 12:35:23 GMT, Per Minborg <pminborg at openjdk.org> wrote:

> This PR proposed to add a new benchmark outlining potential performance gains with the FFM API in internal JDK classes.

Here is a run on a MacBook Pro (16-inch, 2021) M1 Max, , macOS 12.6.1:


Benchmark                                  (size)  Mode  Cnt    Score   Error  Units
MemorySegmentVsBits.bitsEquivalent              2  avgt   30    3,597 ? 0,021  ns/op
MemorySegmentVsBits.bitsEquivalent             16  avgt   30   20,074 ? 0,082  ns/op
MemorySegmentVsBits.bitsEquivalent             64  avgt   30   81,934 ? 0,328  ns/op
MemorySegmentVsBits.bitsEquivalent            256  avgt   30  322,905 ? 1,041  ns/op
MemorySegmentVsBits.panamaHeap                  2  avgt   30    4,538 ? 0,021  ns/op
MemorySegmentVsBits.panamaHeap                 16  avgt   30    7,040 ? 0,078  ns/op
MemorySegmentVsBits.panamaHeap                 64  avgt   30   16,027 ? 0,022  ns/op
MemorySegmentVsBits.panamaHeap                256  avgt   30   43,778 ? 0,090  ns/op
MemorySegmentVsBits.panamaNative                2  avgt   30    4,564 ? 0,040  ns/op
MemorySegmentVsBits.panamaNative               16  avgt   30    7,036 ? 0,059  ns/op
MemorySegmentVsBits.panamaNative               64  avgt   30   17,237 ? 0,066  ns/op
MemorySegmentVsBits.panamaNative              256  avgt   30   31,617 ? 0,129  ns/op
MemorySegmentVsBits.panamaNativeUnaligned       2  avgt   30    4,334 ? 0,016  ns/op
MemorySegmentVsBits.panamaNativeUnaligned      16  avgt   30    6,944 ? 0,044  ns/op
MemorySegmentVsBits.panamaNativeUnaligned      64  avgt   30   17,220 ? 0,102  ns/op
MemorySegmentVsBits.panamaNativeUnaligned     256  avgt   30   31,272 ? 0,184  ns/op

-------------

PR: https://git.openjdk.org/panama-foreign/pull/761