Rev'd to 5.1.4 to show results differernce

[opsiff]

The following commits change results very huge,
but use the same version number in some test result.
More and more people use this version which not
release for likely riscv arch,and post it to ther Internet.
Simplely change the version ,we can compare results
not confused.

commit 81e9de5 ("fstime.c - Seperate r/w files for each parallel") and
commit fb4521c ("Allocate unique fd for each parallel to avoid unexpected syscall test")

[opsiff]

I have a story of one person who used 'git clone' to download the repository's code for benchmarking, while another person downloaded the repository's release for benchmarking, revealing a significant difference in scores :) .
And this story has happened repeatedly within a year, occurring in the company's internal tests, among video website bloggers, and in online chats.

[gstrauss]

Interesting. That version appears to have been last updated in Jan 2011 in 7df8999

[opsiff]

Interesting. That version appears to have been last updated in Jan 2011 in 7df8999

Yes.
The benchmark results change so different with #85 and #82 ,bring the big change in FileCopy and Syscall Benchmark.
They are just happened in the last two years.
People share their benchmark score with differnet parallelism program code, and the same unixbench version number:).

[opsiff]

I think that if not a release version, version number are not important, but show user "Oh ,the benchmanrk happened in difference program version.".