On Mon, Oct 21, 2024 at 5:22 AM Puranjay Mohan <puranjay@xxxxxxxxxx> wrote:
>
> Add a microbenchmark for bpf_csum_diff() helper. This benchmark works by
> filling a 4KB buffer with random data and calculating the internet
> checksum on different parts of this buffer using bpf_csum_diff().
>
> Example run using ./benchs/run_bench_csum_diff.sh on x86_64:
>
> [bpf]$ ./benchs/run_bench_csum_diff.sh
> 4 2.296 ± 0.066M/s (drops 0.000 ± 0.000M/s)
> 8 2.320 ± 0.003M/s (drops 0.000 ± 0.000M/s)
> 16 2.315 ± 0.001M/s (drops 0.000 ± 0.000M/s)
> 20 2.318 ± 0.001M/s (drops 0.000 ± 0.000M/s)
> 32 2.308 ± 0.003M/s (drops 0.000 ± 0.000M/s)
> 40 2.300 ± 0.029M/s (drops 0.000 ± 0.000M/s)
> 64 2.286 ± 0.001M/s (drops 0.000 ± 0.000M/s)
> 128 2.250 ± 0.001M/s (drops 0.000 ± 0.000M/s)
> 256 2.173 ± 0.001M/s (drops 0.000 ± 0.000M/s)
> 512 2.023 ± 0.055M/s (drops 0.000 ± 0.000M/s)
you are not benchmarking bpf_csum_diff(), you are benchmarking how
often you can call bpf_prog_test_run(). Add some batching on the BPF
side, these numbers tell you that there is no difference between
calculating checksum for 4 bytes and for 512, that didn't seem strange
to you?
pw-bot: cr
>
> Signed-off-by: Puranjay Mohan <puranjay@xxxxxxxxxx>
> ---
> tools/testing/selftests/bpf/Makefile | 2 +
> tools/testing/selftests/bpf/bench.c | 4 +
> .../selftests/bpf/benchs/bench_csum_diff.c | 164 ++++++++++++++++++
> .../bpf/benchs/run_bench_csum_diff.sh | 10 ++
> .../selftests/bpf/progs/csum_diff_bench.c | 25 +++
> 5 files changed, 205 insertions(+)
> create mode 100644 tools/testing/selftests/bpf/benchs/bench_csum_diff.c
> create mode 100755 tools/testing/selftests/bpf/benchs/run_bench_csum_diff.sh
> create mode 100644 tools/testing/selftests/bpf/progs/csum_diff_bench.c
>
[...]
> +
> +static void csum_diff_setup(void)
> +{
> + int err;
> + char *buff;
> + size_t i, sz;
> +
> + sz = sizeof(ctx.skel->rodata->buff);
> +
> + setup_libbpf();
> +
> + ctx.skel = csum_diff_bench__open();
> + if (!ctx.skel) {
> + fprintf(stderr, "failed to open skeleton\n");
> + exit(1);
> + }
> +
> + srandom(time(NULL));
> + buff = ctx.skel->rodata->buff;
> +
> + /*
> + * Set first 8 bytes of buffer to 0xdeadbeefdeadbeef, this is later used to verify the
> + * correctness of the helper by comparing the checksum result for 0xdeadbeefdeadbeef that
> + * should be 0x3b3b
> + */
> +
> + *(u64 *)buff = 0xdeadbeefdeadbeef;
> +
> + for (i = 8; i < sz; i++)
> + buff[i] = '1' + random() % 9;
so, you only generate 9 different values for bytes, why? Why not full
byte range?
> +
> + ctx.skel->rodata->buff_len = args.buff_len;
> +
> + err = csum_diff_bench__load(ctx.skel);
> + if (err) {
> + fprintf(stderr, "failed to load skeleton\n");
> + csum_diff_bench__destroy(ctx.skel);
> + exit(1);
> + }
> +}
> +
[...]
