selftests/bpf: add multi-uprobe benchmarks
Subsystems (3)
Andrii Nakryiko
maintainerPATCH 1/1 selftests/bpf: add multi-uprobe benchmarks
Add multi-uprobe and multi-uretprobe benchmarks to bench tool. Multi- and classic uprobes/uretprobes have different low-level triggering code paths, so it's sometimes important to be able to benchmark both flavors of uprobes/uretprobes.
Sample examples from my dev machine below. Single-threaded peformance almost doesn't differ, but with more parallel CPUs triggering the same uprobe/uretprobe the difference grows. This might be due to [0], but given the code is slightly different, there could be other sources of slowdown.
Note, all these numbers will change due to ongoing work to improve uprobe/uretprobe scalability (e.g., [1]), but having benchmark like this is useful for measurements and debugging nevertheless.
uprobe-nop ( 1 cpus): 1.020 ± 0.005M/s ( 1.020M/s/cpu) uretprobe-nop ( 1 cpus): 0.515 ± 0.009M/s ( 0.515M/s/cpu) uprobe-multi-nop ( 1 cpus): 1.036 ± 0.004M/s ( 1.036M/s/cpu) uretprobe-multi-nop ( 1 cpus): 0.512 ± 0.005M/s ( 0.512M/s/cpu)
uprobe-nop ( 8 cpus): 3.481 ± 0.030M/s ( 0.435M/s/cpu) uretprobe-nop ( 8 cpus): 2.222 ± 0.008M/s ( 0.278M/s/cpu) uprobe-multi-nop ( 8 cpus): 3.769 ± 0.094M/s ( 0.471M/s/cpu) uretprobe-multi-nop ( 8 cpus): 2.482 ± 0.007M/s ( 0.310M/s/cpu)
uprobe-nop (16 cpus): 2.968 ± 0.011M/s ( 0.185M/s/cpu) uretprobe-nop (16 cpus): 1.870 ± 0.002M/s ( 0.117M/s/cpu) uprobe-multi-nop (16 cpus): 3.541 ± 0.037M/s ( 0.221M/s/cpu) uretprobe-multi-nop (16 cpus): 2.123 ± 0.026M/s ( 0.133M/s/cpu)
uprobe-nop (32 cpus): 2.524 ± 0.026M/s ( 0.079M/s/cpu) uretprobe-nop (32 cpus): 1.572 ± 0.003M/s ( 0.049M/s/cpu) uprobe-multi-nop (32 cpus): 2.717 ± 0.003M/s ( 0.085M/s/cpu) uretprobe-multi-nop (32 cpus): 1.687 ± 0.007M/s ( 0.053M/s/cpu)
[0] https://lore.kernel.org/linux-trace-kernel/[email protected]/
[1] https://lore.kernel.org/linux-trace-kernel/[email protected]/
--- tools/testing/selftests/bpf/bench.c | 12 +++ .../selftests/bpf/benchs/bench_trigger.c | 81 +++++++++++++++---- .../selftests/bpf/progs/trigger_bench.c | 7 ++ 3 files changed, 85 insertions(+), 15 deletions(-)
tools/testing/selftests/bpf/bench.c +12 -0
@@ -520,6 +520,12 @@ extern const struct bench bench_trig_uprobe_push;extern const struct bench bench_trig_uretprobe_push;extern const struct bench bench_trig_uprobe_ret;extern const struct bench bench_trig_uretprobe_ret; +extern const struct bench bench_trig_uprobe_multi_nop; +extern const struct bench bench_trig_uretprobe_multi_nop; +extern const struct bench bench_trig_uprobe_multi_push; +extern const struct bench bench_trig_uretprobe_multi_push; +extern const struct bench bench_trig_uprobe_multi_ret; +extern const struct bench bench_trig_uretprobe_multi_ret;extern const struct bench bench_rb_libbpf;extern const struct bench bench_rb_custom;
@@ -574,6 +580,12 @@ static const struct bench *benchs[] = { &bench_trig_uretprobe_push, &bench_trig_uprobe_ret, &bench_trig_uretprobe_ret, + &bench_trig_uprobe_multi_nop, + &bench_trig_uretprobe_multi_nop, + &bench_trig_uprobe_multi_push, + &bench_trig_uretprobe_multi_push, + &bench_trig_uprobe_multi_ret, + &bench_trig_uretprobe_multi_ret,/* ringbuf/perfbuf benchmarks */ &bench_rb_libbpf, &bench_rb_custom,
tools/testing/selftests/bpf/benchs/bench_trigger.c +66 -15
131 lines not on this page · raw message
tools/testing/selftests/bpf/progs/trigger_bench.c +7 -0
@@ -32,6 +32,13 @@ int bench_trigger_uprobe(void *ctx)return 0 ; } +SEC("?uprobe.multi") +int bench_trigger_uprobe_multi (void *ctx) +{ +inc_counter (); +return 0 ; +} +const volatile int batch_iters =0 ;SEC ("?raw_tp" )
Signature · 1 line
2.43.5
Quoted 30 lines · Andrii Nakryiko
Add multi-uprobe and multi-uretprobe benchmarks to bench tool. Multi- and classic uprobes/uretprobes have different low-level triggering code paths, so it's sometimes important to be able to benchmark both flavors of uprobes/uretprobes.
Sample examples from my dev machine below. Single-threaded peformance almost doesn't differ, but with more parallel CPUs triggering the same uprobe/uretprobe the difference grows. This might be due to [0], but given the code is slightly different, there could be other sources of slowdown.
Note, all these numbers will change due to ongoing work to improve uprobe/uretprobe scalability (e.g., [1]), but having benchmark like this is useful for measurements and debugging nevertheless.
uprobe-nop ( 1 cpus): 1.020 ± 0.005M/s ( 1.020M/s/cpu) uretprobe-nop ( 1 cpus): 0.515 ± 0.009M/s ( 0.515M/s/cpu) uprobe-multi-nop ( 1 cpus): 1.036 ± 0.004M/s ( 1.036M/s/cpu) uretprobe-multi-nop ( 1 cpus): 0.512 ± 0.005M/s ( 0.512M/s/cpu)uprobe-nop ( 8 cpus): 3.481 ± 0.030M/s ( 0.435M/s/cpu) uretprobe-nop ( 8 cpus): 2.222 ± 0.008M/s ( 0.278M/s/cpu) uprobe-multi-nop ( 8 cpus): 3.769 ± 0.094M/s ( 0.471M/s/cpu) uretprobe-multi-nop ( 8 cpus): 2.482 ± 0.007M/s ( 0.310M/s/cpu)uprobe-nop (16 cpus): 2.968 ± 0.011M/s ( 0.185M/s/cpu) uretprobe-nop (16 cpus): 1.870 ± 0.002M/s ( 0.117M/s/cpu) uprobe-multi-nop (16 cpus): 3.541 ± 0.037M/s ( 0.221M/s/cpu) uretprobe-multi-nop (16 cpus): 2.123 ± 0.026M/s ( 0.133M/s/cpu)
Quoted 4 lines
uprobe-nop (32 cpus): 2.524 ± 0.026M/s ( 0.079M/s/cpu) uretprobe-nop (32 cpus): 1.572 ± 0.003M/s ( 0.049M/s/cpu) uprobe-multi-nop (32 cpus): 2.717 ± 0.003M/s ( 0.085M/s/cpu) uretprobe-multi-nop (32 cpus): 1.687 ± 0.007M/s ( 0.053M/s/cpu)
nice, do you have script for this output?
we could add it to benchs/run_bench_uprobes.sh
lgtm
jirka
Quoted 25 lines
[0] https://lore.kernel.org/linux-trace-kernel/[email protected]/
[1] https://lore.kernel.org/linux-trace-kernel/[email protected]/Signed-off-by Andrii Nakryiko <[email protected]>--- tools/testing/selftests/bpf/bench.c | 12 +++ .../selftests/bpf/benchs/bench_trigger.c | 81 +++++++++++++++---- .../selftests/bpf/progs/trigger_bench.c | 7 ++ 3 files changed, 85 insertions(+), 15 deletions(-)tools/testing/selftests/bpf/bench.c +6
-0@@ -520,6 +520,12 @@ extern const struct bench bench_trig_uprobe_push;extern const struct bench bench_trig_uretprobe_push;extern const struct bench bench_trig_uprobe_ret;extern const struct bench bench_trig_uretprobe_ret; +extern const struct bench bench_trig_uprobe_multi_nop; +extern const struct bench bench_trig_uretprobe_multi_nop; +extern const struct bench bench_trig_uprobe_multi_push; +extern const struct bench bench_trig_uretprobe_multi_push; +extern const struct bench bench_trig_uprobe_multi_ret; +extern const struct bench bench_trig_uretprobe_multi_ret;
SNIP
Quoted 32 lines · Jiri Olsa
Quoted 30 lines · Andrii Nakryiko
Add multi-uprobe and multi-uretprobe benchmarks to bench tool. Multi- and classic uprobes/uretprobes have different low-level triggering code paths, so it's sometimes important to be able to benchmark both flavors of uprobes/uretprobes.
Sample examples from my dev machine below. Single-threaded peformance almost doesn't differ, but with more parallel CPUs triggering the same uprobe/uretprobe the difference grows. This might be due to [0], but given the code is slightly different, there could be other sources of slowdown.
Note, all these numbers will change due to ongoing work to improve uprobe/uretprobe scalability (e.g., [1]), but having benchmark like this is useful for measurements and debugging nevertheless.
uprobe-nop ( 1 cpus): 1.020 ± 0.005M/s ( 1.020M/s/cpu) uretprobe-nop ( 1 cpus): 0.515 ± 0.009M/s ( 0.515M/s/cpu) uprobe-multi-nop ( 1 cpus): 1.036 ± 0.004M/s ( 1.036M/s/cpu) uretprobe-multi-nop ( 1 cpus): 0.512 ± 0.005M/s ( 0.512M/s/cpu)uprobe-nop ( 8 cpus): 3.481 ± 0.030M/s ( 0.435M/s/cpu) uretprobe-nop ( 8 cpus): 2.222 ± 0.008M/s ( 0.278M/s/cpu) uprobe-multi-nop ( 8 cpus): 3.769 ± 0.094M/s ( 0.471M/s/cpu) uretprobe-multi-nop ( 8 cpus): 2.482 ± 0.007M/s ( 0.310M/s/cpu)uprobe-nop (16 cpus): 2.968 ± 0.011M/s ( 0.185M/s/cpu) uretprobe-nop (16 cpus): 1.870 ± 0.002M/s ( 0.117M/s/cpu) uprobe-multi-nop (16 cpus): 3.541 ± 0.037M/s ( 0.221M/s/cpu) uretprobe-multi-nop (16 cpus): 2.123 ± 0.026M/s ( 0.133M/s/cpu)
Quoted 8 lines
Quoted 4 lines
uprobe-nop (32 cpus): 2.524 ± 0.026M/s ( 0.079M/s/cpu) uretprobe-nop (32 cpus): 1.572 ± 0.003M/s ( 0.049M/s/cpu) uprobe-multi-nop (32 cpus): 2.717 ± 0.003M/s ( 0.085M/s/cpu) uretprobe-multi-nop (32 cpus): 1.687 ± 0.007M/s ( 0.053M/s/cpu)nice, do you have script for this output?
we could add it tobenchs/run_bench_uprobes.sh
I keep tuning those scripts to my own needs, so I'm not sure if it's worth adding all of them to selftests. It's very similar to what we already have, but see the exact script below:
#!/bin/bash
set -eufo pipefail
for p in 1 8 16 32; do
for i in uprobe-nop uretprobe-nop uprobe-multi-nop uretprobe-multi-nop; do
summary=$(sudo ./bench -w1 -d3 -p$p -a trig-$i | tail -n1)
total=$(echo "$summary" | cut -d'(' -f1 | cut -d' ' -f3-)
percpu=$(echo "$summary" | cut -d'(' -f2 | cut -d')' -f1 | cut
-d'/' -f1)
printf "%-21s (%2d cpus): %s (%s/s/cpu)\n" $i $p "$total" "$percpu"
done
echodone
Quoted 33 lines
lgtm
Acked-by Jiri Olsa <[email protected]>jirka
Quoted 25 lines
[0] https://lore.kernel.org/linux-trace-kernel/[email protected]/
[1] https://lore.kernel.org/linux-trace-kernel/[email protected]/Signed-off-by Andrii Nakryiko <[email protected]>--- tools/testing/selftests/bpf/bench.c | 12 +++ .../selftests/bpf/benchs/bench_trigger.c | 81 +++++++++++++++---- .../selftests/bpf/progs/trigger_bench.c | 7 ++ 3 files changed, 85 insertions(+), 15 deletions(-)tools/testing/selftests/bpf/bench.c +6
-0@@ -520,6 +520,12 @@ extern const struct bench bench_trig_uprobe_push;extern const struct bench bench_trig_uretprobe_push;extern const struct bench bench_trig_uprobe_ret;extern const struct bench bench_trig_uretprobe_ret; +extern const struct bench bench_trig_uprobe_multi_nop; +extern const struct bench bench_trig_uretprobe_multi_nop; +extern const struct bench bench_trig_uprobe_multi_push; +extern const struct bench bench_trig_uretprobe_multi_push; +extern const struct bench bench_trig_uprobe_multi_ret; +extern const struct bench bench_trig_uretprobe_multi_ret;SNIP
Quoted 60 lines · Andrii Nakryiko
Quoted 32 lines · Jiri Olsa
Quoted 30 lines · Andrii Nakryiko
Add multi-uprobe and multi-uretprobe benchmarks to bench tool. Multi- and classic uprobes/uretprobes have different low-level triggering code paths, so it's sometimes important to be able to benchmark both flavors of uprobes/uretprobes.
Sample examples from my dev machine below. Single-threaded peformance almost doesn't differ, but with more parallel CPUs triggering the same uprobe/uretprobe the difference grows. This might be due to [0], but given the code is slightly different, there could be other sources of slowdown.
Note, all these numbers will change due to ongoing work to improve uprobe/uretprobe scalability (e.g., [1]), but having benchmark like this is useful for measurements and debugging nevertheless.
uprobe-nop ( 1 cpus): 1.020 ± 0.005M/s ( 1.020M/s/cpu) uretprobe-nop ( 1 cpus): 0.515 ± 0.009M/s ( 0.515M/s/cpu) uprobe-multi-nop ( 1 cpus): 1.036 ± 0.004M/s ( 1.036M/s/cpu) uretprobe-multi-nop ( 1 cpus): 0.512 ± 0.005M/s ( 0.512M/s/cpu)uprobe-nop ( 8 cpus): 3.481 ± 0.030M/s ( 0.435M/s/cpu) uretprobe-nop ( 8 cpus): 2.222 ± 0.008M/s ( 0.278M/s/cpu) uprobe-multi-nop ( 8 cpus): 3.769 ± 0.094M/s ( 0.471M/s/cpu) uretprobe-multi-nop ( 8 cpus): 2.482 ± 0.007M/s ( 0.310M/s/cpu)uprobe-nop (16 cpus): 2.968 ± 0.011M/s ( 0.185M/s/cpu) uretprobe-nop (16 cpus): 1.870 ± 0.002M/s ( 0.117M/s/cpu) uprobe-multi-nop (16 cpus): 3.541 ± 0.037M/s ( 0.221M/s/cpu) uretprobe-multi-nop (16 cpus): 2.123 ± 0.026M/s ( 0.133M/s/cpu)Quoted 8 lines
Quoted 4 lines
uprobe-nop (32 cpus): 2.524 ± 0.026M/s ( 0.079M/s/cpu) uretprobe-nop (32 cpus): 1.572 ± 0.003M/s ( 0.049M/s/cpu) uprobe-multi-nop (32 cpus): 2.717 ± 0.003M/s ( 0.085M/s/cpu) uretprobe-multi-nop (32 cpus): 1.687 ± 0.007M/s ( 0.053M/s/cpu)nice, do you have script for this output?
we could add it tobenchs/run_bench_uprobes.shI keep tuning those scripts to my own needs, so I'm not sure if it's worth adding all of them to selftests. It's very similar to what we already have, but see the exact script below:
#!/bin/bash
set -eufo pipefail
for p in 1 8 16 32; do
for i in uprobe-nop uretprobe-nop uprobe-multi-nop uretprobe-multi-nop; do summary=$(sudo ./bench -w1 -d3 -p$p -a trig-$i | tail -n1) total=$(echo "$summary" | cut -d'(' -f1 | cut -d' ' -f3-) percpu=$(echo "$summary" | cut -d'(' -f2 | cut -d')' -f1 | cut -d'/' -f1) printf "%-21s (%2d cpus): %s (%s/s/cpu)\n" $i $p "$total" "$percpu" done echo
Quoted 1 line
done
Added this script to commit log while applying.
This patch was applied to bpf/bpf-next.git (master) by Alexei Starovoitov <[email protected]>:
Hello:
This patch was applied to bpf/bpf-next.git (master) by Alexei Starovoitov <[email protected]>:
Quoted 5 lines · On Mon, 5 Aug 2024 21:29:35 -0700 you wrote:
Add multi-uprobe and multi-uretprobe benchmarks to bench tool. Multi- and classic uprobes/uretprobes have different low-level triggering code paths, so it's sometimes important to be able to benchmark both flavors of uprobes/uretprobes.
Quoted 7 lines
Sample examples from my dev machine below. Single-threaded peformance almost doesn't differ, but with more parallel CPUs triggering the same uprobe/uretprobe the difference grows. This might be due to [0], but given the code is slightly different, there could be other sources of slowdown.
[...]
Here is the summary with links:
- [bpf-next] selftests/bpf: add multi-uprobe benchmarks
https://git.kernel.org/bpf/bpf-next/c/f727b13dbea1
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html