Re: [PATCH i-g-t 3/4] scripts/trace.pl: Calculate stats only after all munging

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 1/22/2018 2:54 AM, Tvrtko Ursulin wrote:

On 20/01/2018 00:24, John.C.Harrison@xxxxxxxxx wrote:
From: John Harrison <John.C.Harrison@xxxxxxxxx>

There are various statistics being calculated multiple times in
multiple places while the log file is being read in. Some of these are
then re-calculated when the database is munged to correct various
issues with the logs. This patch consolidates the calculations into a
separate pass after all the reading and munging has been done.

Note that this actually produces a different final output as the
'execute-delay' values were not previously being re-calculated after
all the fixups. Thus were based on an incorrect calculation.

Signed-off-by: John Harrison <John.C.Harrison@xxxxxxxxx>
Cc: Tvrtko Ursulin <tvrtko.ursulin@xxxxxxxxx>
---
  scripts/trace.pl | 54 ++++++++++++++++++++++++++++--------------------------
  1 file changed, 28 insertions(+), 26 deletions(-)

diff --git a/scripts/trace.pl b/scripts/trace.pl
index cf841b7e..c5d822aa 100755
--- a/scripts/trace.pl
+++ b/scripts/trace.pl
@@ -437,8 +437,7 @@ while (<>) {
          $req{'global'} = $tp{'global'};
          $req{'port'} = $tp{'port'};
          $req{'queue'} = $queue{$key};
-        $req{'submit-delay'} = $submit{$key} - $queue{$key};
-        $req{'execute-delay'} = $req{'start'} - $submit{$key};
+        $req{'submit'} = $submit{$key};

You only use the submit key to look up things two times, so you could use the %submit hash directly to save some lookups, but on the other hand maybe you were going for nicer split between data structures over performance.

When I look at it again, I was using the same approach for 'queue' when %queue was also available so never mind.. maybe an optimisation opportunity with both for later, if would be worth it.
Yeah, I wasn't 100% certain the queue/submit things were safe to use later given that you weren't. However, I think it definitely makes the code easier to maintain if %db is assumed to be the one true source of all information. Once that is constructed, all else can be forgotten about and all subsequent processing only needs to operate on that one data structure.



          $rings{$ring} = $gid++ unless exists $rings{$ring};
          $ringmap{$rings{$ring}} = $ring;
          $db{$key} = \%req;
@@ -458,8 +457,6 @@ while (<>) {
              $db{$key}->{'notify'} = $db{$key}->{'end'};
              $db{$key}->{'no-notify'} = 1;
          }
-        $db{$key}->{'duration'} = $db{$key}->{'notify'} - $db{$key}->{'start'}; -        $db{$key}->{'context-complete-delay'} = $db{$key}->{'end'} - $db{$key}->{'notify'};
      } elsif ($tp_name eq 'i915:intel_engine_notify:') {
          $notify{global_key($ring, $seqno)} = $time;
      } elsif ($tp_name eq 'i915:intel_gpu_freq_change:') {
@@ -493,16 +490,11 @@ foreach my $key (keys %db) {
              $db{$key}->{'notify'} = $db{$key}->{'end'};
              $db{$key}->{'incomplete'} = 1;
          }
-
-        $db{$key}->{'duration'} = $db{$key}->{'notify'} - $db{$key}->{'start'}; -        $db{$key}->{'context-complete-delay'} = $db{$key}->{'end'} - $db{$key}->{'notify'};
      } else {
          # Notify arrived after context complete.
          if (exists $db{$key}->{'no-notify'} and exists $notify{$gkey}) {
              delete $db{$key}->{'no-notify'};
              $db{$key}->{'notify'} = $notify{$gkey};
-            $db{$key}->{'duration'} = $db{$key}->{'notify'} - $db{$key}->{'start'}; -            $db{$key}->{'context-complete-delay'} = $db{$key}->{'end'} - $db{$key}->{'notify'};
          }
      }
  }
@@ -529,8 +521,6 @@ foreach my $key (@keys) {
      if (exists $db{$next_key}) {
          $db{$key}->{'notify'} = $db{$next_key}->{'end'};
          $db{$key}->{'end'} = $db{$key}->{'notify'};
-        $db{$key}->{'duration'} = $db{$key}->{'notify'} - $db{$key}->{'start'}; -        $db{$key}->{'context-complete-delay'} = $db{$key}->{'end'} - $db{$key}->{'notify'};
      }
  }
  @@ -565,19 +555,14 @@ die "Database changed size?!" unless scalar(@sorted_keys) == $keyCount;
  foreach my $key (@sorted_keys) {
      my $ring = $db{$key}->{'ring'};
      my $end = $db{$key}->{'end'};
+    my $start = $db{$key}->{'start'};
+    my $notify = $db{$key}->{'notify'};

Why did you move/put these two here since they are not used outside the $correct_durations block? I'd avoid hash queries if not needed.

The code had mutated a lot before arriving at this point, yer honour. It all made sense at one point in time!


        $first_ts = $db{$key}->{'queue'} if not defined $first_ts or $db{$key}->{'queue'} < $first_ts;
      $last_ts = $end if $end > $last_ts;
  -    $running{$ring} += $end - $db{$key}->{'start'} unless exists $db{$key}->{'no-end'};
-    $runnable{$ring} += $db{$key}->{'execute-delay'};
-    $queued{$ring} += $db{$key}->{'start'} - $db{$key}->{'execute-delay'} - $db{$key}->{'queue'};
-
-    $batch_count{$ring}++;
-
      # correct duration of merged batches
      if ($correct_durations and exists $db{$key}->{'no-end'}) {
-        my $start = $db{$key}->{'start'};
          my $ctx = $db{$key}->{'ctx'};
          my $seqno = $db{$key}->{'seqno'};
          my $next_key;
@@ -591,24 +576,41 @@ foreach my $key (@sorted_keys) {
          # 20us tolerance
          if (exists $db{$next_key} and $db{$next_key}->{'start'} < $start + 20) {
              $re_sort = 1;
-            $db{$next_key}->{'start'} = $start + $db{$key}->{'duration'};
+            $db{$next_key}->{'start'} = $notify;

Makes sense, mine was silly!

              $db{$next_key}->{'start'} = $db{$next_key}->{'end'} if $db{$next_key}->{'start'} > $db{$next_key}->{'end'}; -            $db{$next_key}->{'duration'} = $db{$next_key}->{'notify'} - $db{$next_key}->{'start'};
-            $end = $db{$key}->{'notify'};
              die if $db{$next_key}->{'start'} > $db{$next_key}->{'end'};
          }
-        die if $db{$key}->{'start'} > $db{$key}->{'end'};
+        die if $start > $end;
      }
+}
+
+@sorted_keys = sort sortStart keys %db if $re_sort;
+
+foreach my $key (@sorted_keys) {
+    my $ring = $db{$key}->{'ring'};
+    my $end = $db{$key}->{'end'};
+    my $start = $db{$key}->{'start'};
+    my $notify = $db{$key}->{'notify'};
+
+    $db{$key}->{'context-complete-delay'} = $end - $notify;
+    $db{$key}->{'execute-delay'} = $start - $db{$key}->{'submit'};
+    $db{$key}->{'submit-delay'} = $db{$key}->{'submit'} - $db{$key}->{'queue'};
+    $db{$key}->{'duration'} = $notify - $start;
+
+    $running{$ring} += $end - $start unless exists $db{$key}->{'no-end'};
+    $runnable{$ring} += $db{$key}->{'execute-delay'};
+    $queued{$ring} += $start - $db{$key}->{'execute-delay'} - $db{$key}->{'queue'};
+
+    $batch_count{$ring}++;
+
      $batch_avg{$ring} += $db{$key}->{'duration'};
-    $batch_total_avg{$ring} += $end - $db{$key}->{'start'};
+    $batch_total_avg{$ring} += $end - $start;
        $submit_avg{$ring} += $db{$key}->{'submit-delay'};
      $execute_avg{$ring} += $db{$key}->{'execute-delay'};
-    $ctxsave_avg{$ring} += $db{$key}->{'end'} - $db{$key}->{'notify'};
+    $ctxsave_avg{$ring} += $end - $notify;
  }
  -@sorted_keys = sort sortStart keys %db if $re_sort;
-
  foreach my $ring (sort keys %batch_avg) {
      $batch_avg{$ring} /= $batch_count{$ring};
      $batch_total_avg{$ring} /= $batch_count{$ring};


A great improvement! I didn't spot that you've changed anything which would break the flow.

I'd move those two locals in the if block, but otherwise:

Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@xxxxxxxxx>

Regards,

Tvrtko


_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx




[Index of Archives]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]
  Powered by Linux