$(25$ points$)$ You are given four $(4)$ different benchmarks and three $(3)$ different hardware platforms. The details are as follows:

$\backslash$table$[[$Benchmark A$,$Benchmark B$,$Benchmark C$,$Benchmark D$],[1000$ instructions,$1500$ instructions,$2000$ instructions,$3000$ instructions$],[80\%$ of instructions,$75\%$ of instructions,$25\%$ of instructions,$60\%$ of instructions$],[$are memory ops.,are memory ops.,are memory ops.,are memory ops.$],[75\%$ are reads, the rest are write ops.,$60\%$ are reads, the rest are write ops.,$40\%$ are reads, the rest are write ops.,$25\%$ are reads, the rest are write ops.$]]$

The hardware configurations are as follows:

Computer $1$

$\backslash$table$[[$Cache Level,Hit rate,Miss penalty $($cycles$)],[$L$1,0\mathrm{.}88,4],[$L$2,0\mathrm{.}92,8]]$

Computer $2$

$\backslash$table$[[$Cache Level,Hit rate,Miss penalty $($cycles$)],[$L$1,0\mathrm{.}92,6],[$L$2,0\mathrm{.}94,12]]$

Computer $3$

$\backslash$table$[[$Cache Level,Hit rate,Miss penalty $($cycles$)],[$L$1,0\mathrm{.}94,8],[$L$2,0\mathrm{.}96,10]]$

A read op takes $2$ cycles.

A write op takes $3$ cycles.

Any other op takes $1$ cycle.

What is the best hardware configuration across all benchmarks? Show your work and provide a data$-$driven answer.