lkml.org 
[lkml]   [2022]   [Dec]   [19]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
From
Subject[PATCH 4/4] perf report: Add 'simd' sort field
Date
From: German Gomez <german.gomez@arm.com>

Add 'simd' sort field to visualize SIMD ops in perf-report.

Rows are labeled with the SIMD isa, and the type of predicate (if any):

- [p] partial predicate
- [e] empty predicate (no elements in the vector being used)

Example with Arm SPE and SVE (Scalable Vector Extension):

#include <arm_sve.h>

double src[1025], dst[1025];

int main(void) {
svfloat64_t vc = svdup_f64(1);
for(;;)
for(int i = 0; i < 1025; i += svcntd())
{
svbool_t pg = svwhilelt_b64(i, 1025);
svfloat64_t vsrc = svld1(pg, &src[i]);
svfloat64_t vdst = svadd_x(pg, vsrc, vc);
svst1(pg, &dst[i], vdst);
}
return 0;
}

... compiled using "gcc-11 -march=armv8-a+sve -O3"

Profiling on a platform that implements FEAT_SVE and FEAT_SPEv1p1:

$ perf record -e arm_spe_0// -- ./a.out
$ perf report --itrace=i1i -s overhead,pid,simd,sym

Overhead Pid:Command Simd Symbol
........ ................ ....... ......................

53.76% 10758:program [.] main
46.14% 10758:program [.] SVE [.] main
0.09% 10758:program [p] SVE [.] main

The report shows 0.09% of the sampled SVE operations use partial
predicates due to src and dst arrays not being multiples of the vector
register lengths.

Signed-off-by: German Gomez <german.gomez@arm.com>
Signed-off-by: James Clark <james.clark@arm.com>
---
tools/perf/Documentation/perf-report.txt | 1 +
tools/perf/util/hist.c | 1 +
tools/perf/util/hist.h | 1 +
tools/perf/util/sort.c | 47 ++++++++++++++++++++++++
tools/perf/util/sort.h | 2 +
5 files changed, 52 insertions(+)

diff --git a/tools/perf/Documentation/perf-report.txt b/tools/perf/Documentation/perf-report.txt
index 4fa509b15948..ff524d83a4a7 100644
--- a/tools/perf/Documentation/perf-report.txt
+++ b/tools/perf/Documentation/perf-report.txt
@@ -115,6 +115,7 @@ OPTIONS
- p_stage_cyc: On powerpc, this presents the number of cycles spent in a
pipeline stage. And currently supported only on powerpc.
- addr: (Full) virtual address of the sampled instruction
+ - simd: Flags describing a SIMD operation. "e" for empty Arm SVE predicate. "p" for partial Arm SVE predicate

By default, comm, dso and symbol keys are used.
(i.e. --sort comm,dso,symbol)
diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index 17a05e943b44..e2390114c495 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -742,6 +742,7 @@ __hists__add_entry(struct hists *hists,
.weight = sample->weight,
.ins_lat = sample->ins_lat,
.p_stage_cyc = sample->p_stage_cyc,
+ .simd_flags = sample->simd_flags,
}, *he = hists__findnew_entry(hists, &entry, al, sample_self);

if (!hists->has_callchains && he && he->callchain_size != 0)
diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
index ebd8a8f783ee..e6ecb4453053 100644
--- a/tools/perf/util/hist.h
+++ b/tools/perf/util/hist.h
@@ -80,6 +80,7 @@ enum hist_column {
HISTC_ADDR_FROM,
HISTC_ADDR_TO,
HISTC_ADDR,
+ HISTC_SIMD,
HISTC_NR_COLS, /* Last entry */
};

diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c
index 0ecc2cb13792..5c8bfea2ce34 100644
--- a/tools/perf/util/sort.c
+++ b/tools/perf/util/sort.c
@@ -131,6 +131,52 @@ struct sort_entry sort_thread = {
.se_width_idx = HISTC_THREAD,
};

+/* --sort simd */
+
+static int64_t
+sort__simd_cmp(struct hist_entry *left, struct hist_entry *right)
+{
+ if (left->simd_flags.arch != right->simd_flags.arch)
+ return (int64_t) left->simd_flags.arch - right->simd_flags.arch;
+
+ return (int64_t) left->simd_flags.pred - right->simd_flags.pred;
+}
+
+static const char *hist_entry__get_simd_name(struct simd_flags *simd_flags)
+{
+ u64 arch = simd_flags->arch;
+
+ if (arch & SIMD_OP_FLAGS_ARCH_SVE)
+ return "SVE";
+ else
+ return "n/a";
+}
+
+static int hist_entry__simd_snprintf(struct hist_entry *he, char *bf,
+ size_t size, unsigned int width __maybe_unused)
+{
+ const char *name;
+
+ if (!he->simd_flags.arch)
+ return repsep_snprintf(bf, size, "");
+
+ name = hist_entry__get_simd_name(&he->simd_flags);
+
+ if (he->simd_flags.pred & SIMD_OP_FLAGS_PRED_EMPTY)
+ return repsep_snprintf(bf, size, "[e] %s", name);
+ else if (he->simd_flags.pred & SIMD_OP_FLAGS_PRED_PARTIAL)
+ return repsep_snprintf(bf, size, "[p] %s", name);
+
+ return repsep_snprintf(bf, size, "[.] %s", name);
+}
+
+struct sort_entry sort_simd = {
+ .se_header = "Simd ",
+ .se_cmp = sort__simd_cmp,
+ .se_snprintf = hist_entry__simd_snprintf,
+ .se_width_idx = HISTC_SIMD,
+};
+
/* --sort comm */

/*
@@ -2042,6 +2088,7 @@ static struct sort_dimension common_sort_dimensions[] = {
DIM(SORT_LOCAL_PIPELINE_STAGE_CYC, "local_p_stage_cyc", sort_local_p_stage_cyc),
DIM(SORT_GLOBAL_PIPELINE_STAGE_CYC, "p_stage_cyc", sort_global_p_stage_cyc),
DIM(SORT_ADDR, "addr", sort_addr),
+ DIM(SORT_SIMD, "simd", sort_simd)
};

#undef DIM
diff --git a/tools/perf/util/sort.h b/tools/perf/util/sort.h
index 04ff8b61a2a7..8e69a2a53dc1 100644
--- a/tools/perf/util/sort.h
+++ b/tools/perf/util/sort.h
@@ -110,6 +110,7 @@ struct hist_entry {
u64 p_stage_cyc;
u8 cpumode;
u8 depth;
+ struct simd_flags simd_flags;

/* We are added by hists__add_dummy_entry. */
bool dummy;
@@ -237,6 +238,7 @@ enum sort_type {
SORT_LOCAL_PIPELINE_STAGE_CYC,
SORT_GLOBAL_PIPELINE_STAGE_CYC,
SORT_ADDR,
+ SORT_SIMD,

/* branch stack specific sort keys */
__SORT_BRANCH_STACK,
--
2.25.1
\
 
 \ /
  Last update: 2022-12-19 17:15    [W:0.062 / U:0.348 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site