Skip to contentSkip to conversation

perf c2c: Change the default coalesce setup

PATCH Applied Jiri Olsa 3 participants 6 messages Dec 28, 2018 – Jan 3, 2019 perf events

lore
Share & tools
lore

Jiri Olsa

author reviewer
#0 lore raw

PATCH 1/2 perf c2c: Change the default coalesce setup

Joe suggested to have the coalesce default set just to 'iaddr', because it's easier to read on the (default) c2c report run.

By removing pid from -c option, the c2c report will group all the relevant pids under instruction address bucket. User can always run -c pid,iaddr for more grained output on particular pids.

---
 tools/perf/builtin-c2c.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/builtin-c2c.c +1 -1
index 686fb60a8e11..321687f8f373 100644
@@ -68,7 +68,7 @@ struct c2c_hist_entry {
 	struct hist_entry	he;
 };
 
-static char const *coalesce_default = "pid,iaddr";
+static char const *coalesce_default = "iaddr";
 
 struct perf_c2c {
 	struct perf_tool	tool;
Signature · 1 line
2.17.2
author reviewer
#1 lore raw

PATCH 2/2 perf c2c: Increase the HITM ratio limit for displayed cachelines

The cachelines being reported go down to 0.05%. hat makes for very long output files. Raising that to 0.1%. The user can always specify --show-all if they want all the cachelines.

---
 tools/perf/builtin-c2c.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/builtin-c2c.c +1 -1
index 321687f8f373..30584cb44fb7 100644
@@ -1878,7 +1878,7 @@ static int c2c_hists__reinit(struct c2c_hists *c2c_hists,
 	return hpp_list__parse(&c2c_hists->list, output, sort);
 }
 
-#define DISPLAY_LINE_LIMIT  0.0005
+#define DISPLAY_LINE_LIMIT  0.001
 
 static bool he__display(struct hist_entry *he, struct c2c_stats *stats)
 {
Signature · 1 line
2.17.2
maintainer Applied
#2 lore raw

Re: [PATCH 2/2] perf c2c: Increase the HITM ratio limit for displayed cachelines

Em Fri, Dec 28, 2018 at 11:18:20AM +0100, Jiri Olsa escreveu:

Quoted 3 lines

The cachelines being reported go down to 0.05%. hat makes for very long output files. Raising that to 0.1%. The user can always specify --show-all if they want all the cachelines.

Thanks, applied.

- Arnaldo

Quoted 22 lines
---
 tools/perf/builtin-c2c.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/builtin-c2c.c +1 -1
index 321687f8f373..30584cb44fb7 100644
@@ -1878,7 +1878,7 @@ static int c2c_hists__reinit(struct c2c_hists *c2c_hists,
 	return hpp_list__parse(&c2c_hists->list, output, sort);
 }
 
-#define DISPLAY_LINE_LIMIT  0.0005
+#define DISPLAY_LINE_LIMIT  0.001
 
 static bool he__display(struct hist_entry *he, struct c2c_stats *stats)
 {
Signature · 1 line
2.17.2
Signature · 1 line
- Arnaldo
#3 lore raw

[tip:perf/urgent] perf c2c: Increase the HITM ratio limit for displayed cachelines

In reply to Jiri Olsa
perf c2c: Increase the HITM ratio limit for displayed cachelines
Commit-ID:  c4a75bb9488a9f7b3c7e79017609529d1aa5deab
Gitweb:     https://git.kernel.org/tip/c4a75bb9488a9f7b3c7e79017609529d1aa5deab
Author:     Jiri Olsa <[email protected]>
AuthorDate: Fri, 28 Dec 2018 11:18:20 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Fri, 28 Dec 2018 16:33:07 -0300

perf c2c: Increase the HITM ratio limit for displayed cachelines

The cachelines being reported are the ones with percentages all the way down to 0.05%. That makes for very long output files. Raising that to 0.1%. The user can always specify --show-all if they want all the cachelines with hits.

---
 tools/perf/builtin-c2c.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/builtin-c2c.c +1 -1
index 964f3f5ef420..d340d2e42776 100644
@@ -1878,7 +1878,7 @@ static int c2c_hists__reinit(struct c2c_hists *c2c_hists,
 	return hpp_list__parse(&c2c_hists->list, output, sort);
 }
 
-#define DISPLAY_LINE_LIMIT  0.0005
+#define DISPLAY_LINE_LIMIT  0.001
 
 static bool he__display(struct hist_entry *he, struct c2c_stats *stats)
 {
maintainer Applied
#4 lore raw
In reply to Jiri Olsa

Em Fri, Dec 28, 2018 at 11:18:19AM +0100, Jiri Olsa escreveu:

Quoted 7 lines

Joe suggested to have the coalesce default set just to 'iaddr', because it's easier to read on the (default) c2c report run.

By removing pid from -c option, the c2c report will group all the relevant pids under instruction address bucket. User can always run -c pid,iaddr for more grained output on particular pids.

Thanks, applied.

- Arnaldo

#5 lore raw
In reply to Jiri Olsa
perf c2c: Change the default coalesce setup
Commit-ID:  423701a0c8d754d9a39547f6c48904347e4f1eca
Gitweb:     https://git.kernel.org/tip/423701a0c8d754d9a39547f6c48904347e4f1eca
Author:     Jiri Olsa <[email protected]>
AuthorDate: Fri, 28 Dec 2018 11:18:19 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Fri, 28 Dec 2018 16:33:06 -0300

perf c2c: Change the default coalesce setup

Joe suggested to have the coalesce default set just to 'iaddr', because it's easier to read on the default 'perf c2c report' output.

By removing the "pid" field from the default -c/--coalesce option, the 'perf c2c' report will group all the relevant PIDs under the instruction address ('iaddr') bucket. User can always run "-c pid,iaddr" for a more fine grained output on particular PIDs.

---
 tools/perf/builtin-c2c.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/builtin-c2c.c +1 -1
index f3aa9d02a5ab..964f3f5ef420 100644
@@ -68,7 +68,7 @@ struct c2c_hist_entry {
 	struct hist_entry	he;
 };
 
-static char const *coalesce_default = "pid,iaddr";
+static char const *coalesce_default = "iaddr";
 
 struct perf_c2c {
 	struct perf_tool	tool;