Permalink
Browse files

Address CF's comment (sort of; well not really, I suppose).

  • Loading branch information...
ltratt committed Jan 19, 2016
1 parent 4bf12c7 commit dd6930a60098e3ca87fac0fcf4111ec84aa20f62
Showing with 4 additions and 5 deletions.
  1. +4 −5 warmup.tex
View
@@ -278,12 +278,11 @@ \subsubsection{Ensuring Determinism}
fixed this simply by moving the random seed initialisation to the start
of the in-process iteration main loop.
For the Java benchmarks, we noticed that classes \laurie{just the benchmarking
class, or others too?}\cfbolz{just the benchmarking class, which I guess could
be a problem?} were lazily loaded during the first iteration of
benchmarks, distorting timings. We thus ensured that the benchmarking class
For the Java benchmarks, we noticed that the benchmarking class used to record
timings was lazily loaded during the first iteration of
benchmarks, distorting timings \laurie{how do we know that other classes in java benchmarks aren't lazily loaded}. We thus ensured that the benchmarking class
is fully loaded before the first iteration by calling a static
method on the benchmarking class, which forces it to be loaded.
method \laurie{does each benchmark call this static method? from its main method, or from a static initializer or somethign?} on the benchmarking class, which forces it to be loaded.
Bearing in mind surprising
results such as the importance of link order~\cite{mytkowicz09surprising}, we

0 comments on commit dd6930a

Please sign in to comment.