This version has been modified to do its work in memory, rather than
reading and writing files.
Generates a large buffer-full of data, then compresses that into another
memory buffer, and then checks the result, decompresses back to a memory
buffer, etc.
Rough estimates of some of the basic hardware level activities: CPI, I-
and D-cache activity, and TLB activity. This data is only useful for comparing
the '95 benchmarks, not for comparing against any particular hardware.
gprof
Full gprof(1) outputs. [Procedure level profiling with call graphs.]
prof
The first 50 lines of the profile taken from the full gprof(1) output.