deletions | additions       diff --git a/artifact/session3.tex b/artifact/session3.tex  index 6846178..f76679e 100644  --- a/artifact/session3.tex  +++ b/artifact/session3.tex 
...
$ ./mcvm -jit_feval_opt false <  benchmarks/scripts/base/odeRK4  ***********************************************  McVM - The McLab Virtual Machine v1.0 Visit http://www.sable.mcgill.ca for more info.  *********************************************** 
...
\end{verbatim}  \end{small}  \noindent The experiment duration on our platform was $\approx2$m, with a time per trial of $\approx32.536$s $\approx32.537$s  (discarding the warm-up run). The resulting speedup for the base code caching mechanism was thus $32.866/32.536=1.010$, $32.867/32.537=1.010$,  slightly different than the one reported in \mytable\ref{tab:feval} on the Intel Xeon platform, for which we repeated each experiment $10$ times. We can now set an upper bound for speedups by measuring the running time when the code has been optimized by hand inserting direct calls in place of {\tt feval} instructions:  \begin{small}  \begin{verbatim}  $ ./mcvm < benchmarks/scripts/direct/odeRK4  ***********************************************  McVM - The McLab Virtual Machine v1.0  Visit http://www.sable.mcgill.ca for more info.  ***********************************************  >: >: Compiling function: "testSH_direct"  Compiling function: "odeRK4_testSHfun"  Compiling function: "testSHfun"  Compiling function: "rhsSteelHeat"  [TOC] Elapsed time: 11.776950 seconds  t y_RK4  0.0000 1.000000  20.0000 227.364633  \end{verbatim}  \end{small}  \noindent In this scenario McVM can compile the whole program ahead of time, as {\tt rhsSteelHeat} is not invoked through an {\tt feval} call anymore. A comparison of the running times suggests a rough $32.537/11.777=2.791$ speedup for by-hand optimization w.r.t. the baseline version.