On 12/08/2016 05:39 PM, Martin Sebor wrote:
> On 12/08/2016 05:55 AM, Martin Liška wrote:
>> With the patch applied, one can distinguish between PROB_VERY_UNLIKELY and
>> real zero probability:
>
> I tried to see if formatting the expression
>
> e->probability * 100.0 / REG_BR_PROB_BASE
>
> with "%.2f" is guaranteed to output non-zero when e->probability
> is non-zero. If I got the values right then the smallest non-zero
> e->probability can be as low as 1, and REG_BR_PROB_BASE is 10000.
> That evaluates to 0.0099999999999999985 which Glibc formats as
> 0.01 because (AFAIK) its printf rounds to nearest.
Thank you for playing with that.
>
> If this is guaranteed then I think it's fine. Otherwise, if there
> is a chance that the printed result could be 0.00% for a non-zero
> probability it might be worth to detect it and make sure it's at
> least 0.01% to avoid the same confusion I had with the 0.00%.
Yep, I like that assumption.
>
> FWIW, if GCC uses integers rather than floats internally to make
> decisions (I don't know) then also printing integers would give
> the most accurate results. E.g., something like
> printf ("%2u.%u%%", e->probability / 100, e->probability % 100)
That's bit tricky as you need "%2u.%02u%%", where '02u' depends on result of
REG_BR_PROB_BASE / 100. We plan with Honza to utilize probably sreal type
instead of the scaled integers. So that I prepared a new patch (not tested yet).
Ideas?
Thanks,
Martin
>
> Thanks
> Martin
>
>>
>> f ()
>> {
>> int _1;
>>
>> <bb 2> [100.00%]:
>> _1 = __builtin_sprintf (&d2, "%i", 12);
>> if (_1 != 2)
>> goto <bb 3>; [0.04%]
>> else
>> goto <bb 4>; [99.96%]
>>
>> <bb 3> [0.04%]:
>> __builtin_abort ();
>>
>> <bb 4> [99.96%]:
>> return;
>>
>> }
>>
>> Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
>>
>> Ready to be installed?
>> Martin
>>
>
>From afbc46f8ebcdf00d320a066b692537399cd189c0 Mon Sep 17 00:00:00 2001
From: marxin <mli...@suse.cz>
Date: Thu, 1 Dec 2016 10:55:55 +0100
Subject: [PATCH] Print 2 digits after decimal delimiter for BB frequencies
gcc/ChangeLog:
2016-12-01 Martin Liska <mli...@suse.cz>
* gimple-pretty-print.c (dump_probability): New function.
(dump_edge_probability): Use the function.
(dump_gimple_label): Likewise.
(dump_gimple_bb_header): Likewise.
gcc/testsuite/ChangeLog:
2016-12-02 Martin Liska <mli...@suse.cz>
* gcc.dg/tree-ssa/20040703-1.c: Update scanned pattern.
* gcc.dg/tree-ssa/dump-2.c: Likewise.
---
gcc/gimple-pretty-print.c | 30 ++++++++++++++++++++++++------
gcc/testsuite/gcc.dg/tree-ssa/20040703-1.c | 2 +-
gcc/testsuite/gcc.dg/tree-ssa/dump-2.c | 2 +-
3 files changed, 26 insertions(+), 8 deletions(-)
diff --git a/gcc/gimple-pretty-print.c b/gcc/gimple-pretty-print.c
index b5e866d..27a22c4 100644
--- a/gcc/gimple-pretty-print.c
+++ b/gcc/gimple-pretty-print.c
@@ -72,13 +72,32 @@ debug_gimple_stmt (gimple *gs)
print_gimple_stmt (stderr, gs, 0, TDF_VOPS|TDF_MEMSYMS);
}
+
+/* Return formatted string of a VALUE probability
+ (biased by REG_BR_PROB_BASE). Returned string is allocated
+ by xstrdup_for_dump. */
+
+static const char *
+dump_probability (int value)
+{
+ char buf[16];
+ float minimum = 0.01f;
+
+ gcc_assert (0 <= value && value <= REG_BR_PROB_BASE);
+ float fvalue = value * 100.0f / REG_BR_PROB_BASE;
+ if (fvalue < minimum && value > 0)
+ return "[0.01%]";
+
+ sprintf (buf, "[%.2f%%]", fvalue);
+ return xstrdup_for_dump (buf);
+}
+
/* Dump E probability to BUFFER. */
static void
dump_edge_probability (pretty_printer *buffer, edge e)
{
- pp_scalar (buffer, " [%.1f%%]",
- e->probability * 100.0 / REG_BR_PROB_BASE);
+ pp_scalar (buffer, " %s", dump_probability (e->probability));
}
/* Print GIMPLE statement G to FILE using SPC indentation spaces and
@@ -1023,8 +1042,7 @@ dump_gimple_label (pretty_printer *buffer, glabel *gs, int spc, int flags)
dump_generic_node (buffer, label, spc, flags, false);
basic_block bb = gimple_bb (gs);
if (bb && !(flags & TDF_GIMPLE))
- pp_scalar (buffer, " [%.1f%%]",
- bb->frequency * 100.0 / REG_BR_PROB_BASE);
+ pp_scalar (buffer, " %s", dump_probability (bb->frequency));
pp_colon (buffer);
}
if (flags & TDF_GIMPLE)
@@ -2590,8 +2608,8 @@ dump_gimple_bb_header (FILE *outf, basic_block bb, int indent, int flags)
if (flags & TDF_GIMPLE)
fprintf (outf, "%*sbb_%d:\n", indent, "", bb->index);
else
- fprintf (outf, "%*s<bb %d> [%.1f%%]:\n", indent, "", bb->index,
- bb->frequency * 100.0 / REG_BR_PROB_BASE);
+ fprintf (outf, "%*s<bb %d> %s:\n",
+ indent, "", bb->index, dump_probability (bb->frequency));
}
}
}
diff --git a/gcc/testsuite/gcc.dg/tree-ssa/20040703-1.c b/gcc/testsuite/gcc.dg/tree-ssa/20040703-1.c
index 2980047..eb9fb56 100644
--- a/gcc/testsuite/gcc.dg/tree-ssa/20040703-1.c
+++ b/gcc/testsuite/gcc.dg/tree-ssa/20040703-1.c
@@ -9,4 +9,4 @@ float foo(float x)
}
/* We should *not* fold the arithmetic. */
-/* { dg-final { scan-tree-dump-times "0\\.0\[^%\]" 0 "dom2"} } */
+/* { dg-final { scan-tree-dump-times "0\\.0\[^%0\]" 0 "dom2"} } */
diff --git a/gcc/testsuite/gcc.dg/tree-ssa/dump-2.c b/gcc/testsuite/gcc.dg/tree-ssa/dump-2.c
index 11cde92..8a63af4 100644
--- a/gcc/testsuite/gcc.dg/tree-ssa/dump-2.c
+++ b/gcc/testsuite/gcc.dg/tree-ssa/dump-2.c
@@ -6,4 +6,4 @@ int f(void)
return 0;
}
-/* { dg-final { scan-tree-dump "<bb \[0-9\]> \\\[100\\\.0%\\\]:" "optimized" } } */
+/* { dg-final { scan-tree-dump "<bb \[0-9\]> \\\[100\\\.00%\\\]:" "optimized" } } */
--
2.10.2
