Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756364Ab0LASzc (ORCPT ); Wed, 1 Dec 2010 13:55:32 -0500 Received: from casper.infradead.org ([85.118.1.10]:39103 "EHLO casper.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756344Ab0LASza (ORCPT ); Wed, 1 Dec 2010 13:55:30 -0500 Date: Wed, 1 Dec 2010 16:54:52 -0200 From: Arnaldo Carvalho de Melo To: Stephane Eranian Cc: linux-kernel@vger.kernel.org, peterz@infradead.org, mingo@elte.hu, paulus@samba.org, davem@davemloft.net, fweisbec@gmail.com, perfmon2-devel@lists.sf.net, eranian@gmail.com, robert.richter@amd.com Subject: Re: [PATCH] perf: add csv-style output to perf stat (v2) Message-ID: <20101201185452.GB26071@ghostprotocols.net> References: <4cf68aa7.0fedd80a.5294.1203@mx.google.com> <20101201181001.GA26071@ghostprotocols.net> MIME-Version: 1.0 Content-Type: multipart/mixed; boundary="bp/iNruPH9dso1Pn" Content-Disposition: inline In-Reply-To: <20101201181001.GA26071@ghostprotocols.net> X-Url: http://acmel.wordpress.com User-Agent: Mutt/1.5.19 (2009-01-05) X-SRS-Rewrite: SMTP reverse-path rewritten from by casper.infradead.org See http://www.infradead.org/rpr.html Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 14536 Lines: 450 --bp/iNruPH9dso1Pn Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Em Wed, Dec 01, 2010 at 04:10:01PM -0200, Arnaldo Carvalho de Melo escreveu: > Em Wed, Dec 01, 2010 at 06:49:05PM +0200, Stephane Eranian escreveu: > > This patch adds an option (-x/--field-separator) to print counts using > > a CSV-style output. The user can pass a custom separator. This makes it > > very easy to import counts directly into your favorite spreadsheet without > > having to write scripts. > > [root@mica ~]# perf stat -x, -a -A -- sleep 1 > -B option not supported with -x > > usage: perf stat [] [] > > -e, --event event selector. use 'perf list' to list > > Guess we'll have to disable big_num automatically when -x is specified, > but at the same time notice if the user _explicitely_ asked for -B. > > I'll fix this now. Should be OK now, patch attached. Please check if it works as expected and then I'll push it to Ingo, thanks. [root@mica ~]# perf stat -x, -a -A -- sleep 1 2>&1 | head -5 CPU0,998.225789,task-clock-msecs CPU1,998.385393,task-clock-msecs CPU2,998.370913,task-clock-msecs CPU3,998.356865,task-clock-msecs CPU0,2272,context-switches [root@mica ~]# perf stat -B -x, -a -A -- sleep 1 2>&1 | head -5 -B option not supported with -x usage: perf stat [] [] -e, --event event selector. use 'perf list' to list available events [root@mica ~]# perf stat -B -a -- sleep 1 2>&1 | head -10 Performance counter stats for 'sleep 1': 3992.488962 task-clock-msecs # 3.988 CPUs 9,082 context-switches # 0.002 M/sec 15 CPU-migrations # 0.000 M/sec 160 page-faults # 0.000 M/sec 84,802,643 cycles # 21.241 M/sec (scaled from 70.03%) 45,056,634 instructions # 0.531 IPC (scaled from 80.02%) 8,672,889 branches # 2.172 M/sec (scaled from 80.01%) [root@mica ~]# perf stat -B -A -a -- sleep 1 2>&1 | head -10 Performance counter stats for 'sleep 1': CPU0 998.135082 task-clock-msecs # 0.997 CPUs CPU1 998.137405 task-clock-msecs # 0.997 CPUs CPU2 998.144046 task-clock-msecs # 0.997 CPUs CPU3 998.146202 task-clock-msecs # 0.997 CPUs CPU0 2,280 context-switches # 0.002 M/sec CPU1 2,282 context-switches # 0.002 M/sec CPU2 2,244 context-switches # 0.002 M/sec [root@mica ~]# perf stat --no-big-num -A -a -- sleep 1 2>&1 | head -10 Performance counter stats for 'sleep 1': CPU0 998.125848 task-clock-msecs # 0.997 CPUs CPU1 998.244633 task-clock-msecs # 0.997 CPUs CPU2 998.337969 task-clock-msecs # 0.997 CPUs CPU3 998.313756 task-clock-msecs # 0.997 CPUs CPU0 2268 context-switches # 0.002 M/sec CPU1 2218 context-switches # 0.002 M/sec CPU2 2218 context-switches # 0.002 M/sec [root@mica ~]# --bp/iNruPH9dso1Pn Content-Type: text/plain; charset=us-ascii Content-Disposition: attachment; filename="0001-perf-stat-Add-csv-style-output.patch" >From d7470b6afca85ed4388fff57fc9d89f5a3be02ff Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Wed, 1 Dec 2010 18:49:05 +0200 Subject: [PATCH] perf stat: Add csv-style output This patch adds an option (-x/--field-separator) to print counts using a CSV-style output. The user can pass a custom separator. This makes it very easy to import counts directly into your favorite spreadsheet without having to write scripts. Example: $ perf stat --field-separator=, -a -- sleep 1 4009.961740,task-clock-msecs 13,context-switches 2,CPU-migrations 189,page-faults 9596385684,cycles 3493659441,instructions 872897069,branches 41562,branch-misses 22424,cache-references 1289,cache-misses Works also in non-aggregated mode: $ perf stat -x , -a -A -- sleep 1 CPU0,1002.526168,task-clock-msecs CPU1,1002.528365,task-clock-msecs CPU2,1002.523360,task-clock-msecs CPU3,1002.519878,task-clock-msecs CPU0,1,context-switches CPU1,5,context-switches CPU2,5,context-switches CPU3,6,context-switches CPU0,0,CPU-migrations CPU1,1,CPU-migrations CPU2,0,CPU-migrations CPU3,1,CPU-migrations CPU0,2,page-faults CPU1,6,page-faults CPU2,9,page-faults CPU3,174,page-faults CPU0,2399439771,cycles CPU1,2380369063,cycles CPU2,2399142710,cycles CPU3,2373161192,cycles CPU0,872900618,instructions CPU1,873030960,instructions CPU2,872714525,instructions CPU3,874460580,instructions CPU0,221556839,branches CPU1,218134342,branches CPU2,218161730,branches CPU3,218284093,branches CPU0,18556,branch-misses CPU1,1449,branch-misses CPU2,3447,branch-misses CPU3,12714,branch-misses CPU0,8330,cache-references CPU1,313844,cache-references CPU2,47993728,cache-references CPU3,826481,cache-references CPU0,272,cache-misses CPU1,5360,cache-misses CPU2,1342193,cache-misses CPU3,13992,cache-misses This second version adds the ability to name a separator and uses field-separator as the long option to be consistent with perf report. Commiter note: Since we enabled --big-num by default in 201e0b0 and -x can't be used with it, we need to notice if the user explicitely enabled or disabled -B, add code to disable big_num if the user didn't explicitely set --big_num when -x is used. Cc: David S. Miller Cc: Frederik Weisbecker Cc: Ingo Molnar Cc: paulus@samba.org Cc: Peter Zijlstra Cc: Robert Richter LKML-Reference: <4cf68aa7.0fedd80a.5294.1203@mx.google.com> Signed-off-by: Stephane Eranian Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/Documentation/perf-stat.txt | 5 + tools/perf/builtin-stat.c | 144 +++++++++++++++++++++++--------- 2 files changed, 109 insertions(+), 40 deletions(-) diff --git a/tools/perf/Documentation/perf-stat.txt b/tools/perf/Documentation/perf-stat.txt index c469ba5..b6da7af 100644 --- a/tools/perf/Documentation/perf-stat.txt +++ b/tools/perf/Documentation/perf-stat.txt @@ -78,6 +78,11 @@ This option is only valid in system-wide mode. --verbose:: be more verbose (show counter open errors, etc) +-x SEP:: +--field-separator SEP:: +print counts using a CSV-style output to make it easy to import directly into +spreadsheets. Columns are separated by the string specified in SEP. + EXAMPLES -------- diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index acbf7cc..7ff746d 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -52,6 +52,8 @@ #include #include +#define DEFAULT_SEPARATOR " " + static struct perf_event_attr default_attrs[] = { { .type = PERF_TYPE_SOFTWARE, .config = PERF_COUNT_SW_TASK_CLOCK }, @@ -83,7 +85,10 @@ static int thread_num = 0; static pid_t child_pid = -1; static bool null_run = false; static bool big_num = true; +static int big_num_opt = -1; static const char *cpu_list; +static const char *csv_sep = NULL; +static bool csv_output = false; static int *fd[MAX_NR_CPUS][MAX_COUNTERS]; @@ -449,12 +454,18 @@ static void print_noise(int counter, double avg) static void nsec_printout(int cpu, int counter, double avg) { double msecs = avg / 1e6; + char cpustr[16] = { '\0', }; + const char *fmt = csv_output ? "%s%.6f%s%s" : "%s%18.6f%s%-24s"; if (no_aggr) - fprintf(stderr, "CPU%-4d %18.6f %-24s", - cpumap[cpu], msecs, event_name(counter)); - else - fprintf(stderr, " %18.6f %-24s", msecs, event_name(counter)); + sprintf(cpustr, "CPU%*d%s", + csv_output ? 0 : -4, + cpumap[cpu], csv_sep); + + fprintf(stderr, fmt, cpustr, msecs, csv_sep, event_name(counter)); + + if (csv_output) + return; if (MATCH_EVENT(SOFTWARE, SW_TASK_CLOCK, counter)) { fprintf(stderr, " # %10.3f CPUs ", @@ -466,18 +477,26 @@ static void abs_printout(int cpu, int counter, double avg) { double total, ratio = 0.0; char cpustr[16] = { '\0', }; + const char *fmt; + + if (csv_output) + fmt = "%s%.0f%s%s"; + else if (big_num) + fmt = "%s%'18.0f%s%-24s"; + else + fmt = "%s%18.0f%s%-24s"; if (no_aggr) - sprintf(cpustr, "CPU%-4d", cpumap[cpu]); + sprintf(cpustr, "CPU%*d%s", + csv_output ? 0 : -4, + cpumap[cpu], csv_sep); else cpu = 0; - if (big_num) - fprintf(stderr, "%s %'18.0f %-24s", - cpustr, avg, event_name(counter)); - else - fprintf(stderr, "%s %18.0f %-24s", - cpustr, avg, event_name(counter)); + fprintf(stderr, fmt, cpustr, avg, csv_sep, event_name(counter)); + + if (csv_output) + return; if (MATCH_EVENT(HARDWARE, HW_INSTRUCTIONS, counter)) { total = avg_stats(&runtime_cycles_stats[cpu]); @@ -515,8 +534,9 @@ static void print_counter_aggr(int counter) int scaled = event_scaled[counter]; if (scaled == -1) { - fprintf(stderr, " %18s %-24s\n", - "", event_name(counter)); + fprintf(stderr, "%*s%s%-24s\n", + csv_output ? 0 : 18, + "", csv_sep, event_name(counter)); return; } @@ -525,6 +545,11 @@ static void print_counter_aggr(int counter) else abs_printout(-1, counter, avg); + if (csv_output) { + fputc('\n', stderr); + return; + } + print_noise(counter, avg); if (scaled) { @@ -554,8 +579,12 @@ static void print_counter(int counter) ena = cpu_counts[cpu][counter].ena; run = cpu_counts[cpu][counter].run; if (run == 0 || ena == 0) { - fprintf(stderr, "CPU%-4d %18s %-24s", cpumap[cpu], - "", event_name(counter)); + fprintf(stderr, "CPU%*d%s%*s%s%-24s", + csv_output ? 0 : -4, + cpumap[cpu], csv_sep, + csv_output ? 0 : 18, + "", csv_sep, + event_name(counter)); fprintf(stderr, "\n"); continue; @@ -566,11 +595,13 @@ static void print_counter(int counter) else abs_printout(cpu, counter, val); - print_noise(counter, 1.0); + if (!csv_output) { + print_noise(counter, 1.0); - if (run != ena) { - fprintf(stderr, " (scaled from %.2f%%)", + if (run != ena) { + fprintf(stderr, " (scaled from %.2f%%)", 100.0 * run / ena); + } } fprintf(stderr, "\n"); } @@ -582,21 +613,23 @@ static void print_stat(int argc, const char **argv) fflush(stdout); - fprintf(stderr, "\n"); - fprintf(stderr, " Performance counter stats for "); - if(target_pid == -1 && target_tid == -1) { - fprintf(stderr, "\'%s", argv[0]); - for (i = 1; i < argc; i++) - fprintf(stderr, " %s", argv[i]); - } else if (target_pid != -1) - fprintf(stderr, "process id \'%d", target_pid); - else - fprintf(stderr, "thread id \'%d", target_tid); + if (!csv_output) { + fprintf(stderr, "\n"); + fprintf(stderr, " Performance counter stats for "); + if(target_pid == -1 && target_tid == -1) { + fprintf(stderr, "\'%s", argv[0]); + for (i = 1; i < argc; i++) + fprintf(stderr, " %s", argv[i]); + } else if (target_pid != -1) + fprintf(stderr, "process id \'%d", target_pid); + else + fprintf(stderr, "thread id \'%d", target_tid); - fprintf(stderr, "\'"); - if (run_count > 1) - fprintf(stderr, " (%d runs)", run_count); - fprintf(stderr, ":\n\n"); + fprintf(stderr, "\'"); + if (run_count > 1) + fprintf(stderr, " (%d runs)", run_count); + fprintf(stderr, ":\n\n"); + } if (no_aggr) { for (counter = 0; counter < nr_counters; counter++) @@ -606,15 +639,17 @@ static void print_stat(int argc, const char **argv) print_counter_aggr(counter); } - fprintf(stderr, "\n"); - fprintf(stderr, " %18.9f seconds time elapsed", - avg_stats(&walltime_nsecs_stats)/1e9); - if (run_count > 1) { - fprintf(stderr, " ( +- %7.3f%% )", + if (!csv_output) { + fprintf(stderr, "\n"); + fprintf(stderr, " %18.9f seconds time elapsed", + avg_stats(&walltime_nsecs_stats)/1e9); + if (run_count > 1) { + fprintf(stderr, " ( +- %7.3f%% )", 100*stddev_stats(&walltime_nsecs_stats) / avg_stats(&walltime_nsecs_stats)); + } + fprintf(stderr, "\n\n"); } - fprintf(stderr, "\n\n"); } static volatile int signr = -1; @@ -644,6 +679,13 @@ static const char * const stat_usage[] = { NULL }; +static int stat__set_big_num(const struct option *opt __used, + const char *s __used, int unset) +{ + big_num_opt = unset ? 0 : 1; + return 0; +} + static const struct option options[] = { OPT_CALLBACK('e', "event", NULL, "event", "event selector. use 'perf list' to list available events", @@ -664,12 +706,15 @@ static const struct option options[] = { "repeat command and print average + stddev (max: 100)"), OPT_BOOLEAN('n', "null", &null_run, "null run - dont start any counters"), - OPT_BOOLEAN('B', "big-num", &big_num, - "print large numbers with thousands\' separators"), + OPT_CALLBACK_NOOPT('B', "big-num", NULL, NULL, + "print large numbers with thousands\' separators", + stat__set_big_num), OPT_STRING('C', "cpu", &cpu_list, "cpu", "list of cpus to monitor in system-wide"), OPT_BOOLEAN('A', "no-aggr", &no_aggr, "disable CPU count aggregation"), + OPT_STRING('x', "field-separator", &csv_sep, "separator", + "print counts with custom separator"), OPT_END() }; @@ -682,6 +727,25 @@ int cmd_stat(int argc, const char **argv, const char *prefix __used) argc = parse_options(argc, argv, options, stat_usage, PARSE_OPT_STOP_AT_NON_OPTION); + + if (csv_sep) + csv_output = true; + else + csv_sep = DEFAULT_SEPARATOR; + + /* + * let the spreadsheet do the pretty-printing + */ + if (csv_output) { + /* User explicitely passed -B? */ + if (big_num_opt == 1) { + fprintf(stderr, "-B option not supported with -x\n"); + usage_with_options(stat_usage, options); + } else /* Nope, so disable big number formatting */ + big_num = false; + } else if (big_num_opt == 0) /* User passed --no-big-num */ + big_num = false; + if (!argc && target_pid == -1 && target_tid == -1) usage_with_options(stat_usage, options); if (run_count <= 0) -- 1.7.1.rc2 --bp/iNruPH9dso1Pn-- -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/