This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Take branch misprediction effects into account when RTL loop unrolling (issue 6099055) File loop-unroll.c (right):
loop-unroll.c:156: static bool
An empty line here.
loop-unroll.c:182: static bool
add empty line after comment.
loop-unroll.c:213: /* Compute the number of branches in LOOP, weighted
by execution counts.  */
same here.
loop-unroll.c:215: compute_weighted_branches(struct loop *loop)
Missing space. Similarly for other functions.
loop-unroll.c:255: data exists, simply return the current NUNROLL
factor.  */
same here.
loop-unroll.c:281: while (loop_outer(outer))
Should the check start from the current loop? Or the handling of the
current loop should be peeled out -- max_unroll_factor =
loop-unroll.c:317: return (PARAM_VALUE
The number of branches may be larger than budget --> leading to overflow
-- need to guard against it.
loop-unroll.c:318: weighted_outer_branches)/(weighted_num_branches - 1)
+ 1;
Should it continue walking up the loop tree and find the minmal unroll
loop-unroll.c:747: && loop_has_FP_comp(loop)
missing space
loop-unroll.c:749: && desc->niter < (unsigned) PARAM_VALUE
line too long.
loop-unroll.c:1057: && loop_has_FP_comp(loop)
Refactor the check into a helper function?
File params.def (right):
Missing comment.
missing comment.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]