From: Ian Rogers <irogers@google.com>
To: Sandipan Das <sandipan.das@amd.com>,
Ravi Bangoria <ravi.bangoria@amd.com>,
Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@redhat.com>,
Arnaldo Carvalho de Melo <acme@kernel.org>,
Namhyung Kim <namhyung@kernel.org>,
Mark Rutland <mark.rutland@arm.com>,
Alexander Shishkin <alexander.shishkin@linux.intel.com>,
Jiri Olsa <jolsa@kernel.org>, Ian Rogers <irogers@google.com>,
Adrian Hunter <adrian.hunter@intel.com>,
John Garry <john.g.garry@oracle.com>,
Kan Liang <kan.liang@linux.intel.com>,
Jing Zhang <renyu.zj@linux.alibaba.com>,
Thomas Richter <tmricht@linux.ibm.com>,
James Clark <james.clark@arm.com>,
linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org,
Stephane Eranian <eranian@google.com>
Subject: [PATCH v1 06/13] perf jevents: Add hardware prefetch (hwpf) metric group for AMD
Date: Wed, 28 Feb 2024 16:15:30 -0800 [thread overview]
Message-ID: <20240229001537.4158049-7-irogers@google.com> (raw)
In-Reply-To: <20240229001537.4158049-1-irogers@google.com>
Add metrics that give the utility of hardware prefetches on zen2, zen3
and zen4.
Signed-off-by: Ian Rogers <irogers@google.com>
---
tools/perf/pmu-events/amd_metrics.py | 61 ++++++++++++++++++++++++++++
1 file changed, 61 insertions(+)
diff --git a/tools/perf/pmu-events/amd_metrics.py b/tools/perf/pmu-events/amd_metrics.py
index 52451bcb4bbf..dda904998269 100755
--- a/tools/perf/pmu-events/amd_metrics.py
+++ b/tools/perf/pmu-events/amd_metrics.py
@@ -125,6 +125,66 @@ def AmdBr():
description="breakdown of retired branch instructions")
+def AmdHwpf():
+ """Returns a MetricGroup representing AMD hardware prefetch metrics."""
+ if zen_model <= 1:
+ return None
+
+ hwp_ld = Event("ls_dispatch.ld_dispatch")
+ hwp_l2 = Event("ls_hw_pf_dc_fills.local_l2",
+ "ls_hw_pf_dc_fills.lcl_l2",
+ "ls_hw_pf_dc_fill.ls_mabresp_lcl_l2")
+ hwp_lc = Event("ls_hw_pf_dc_fills.local_ccx",
+ "ls_hw_pf_dc_fills.int_cache",
+ "ls_hw_pf_dc_fill.ls_mabresp_lcl_cache")
+ hwp_lm = Event("ls_hw_pf_dc_fills.dram_io_near",
+ "ls_hw_pf_dc_fills.mem_io_local",
+ "ls_hw_pf_dc_fill.ls_mabresp_lcl_dram")
+ hwp_rc = Event("ls_hw_pf_dc_fills.far_cache",
+ "ls_hw_pf_dc_fills.ext_cache_remote",
+ "ls_hw_pf_dc_fill.ls_mabresp_rmt_cache")
+ hwp_rm = Event("ls_hw_pf_dc_fills.dram_io_far",
+ "ls_hw_pf_dc_fills.mem_io_remote",
+ "ls_hw_pf_dc_fill.ls_mabresp_rmt_dram")
+
+ loc_pf = hwp_l2 + hwp_lc + hwp_lm
+ rem_pf = hwp_rc + hwp_rm
+ all_pf = loc_pf + rem_pf
+
+ r1 = d_ratio(ins, all_pf)
+ r2 = d_ratio(hwp_ld, all_pf)
+ r3 = d_ratio(all_pf, interval_sec)
+
+ overview = MetricGroup("hwpf_overview", [
+ Metric("hwpf_ov_insn_bt_hwpf", "Insn between HWPF", r1, "insns"),
+ Metric("hwpf_ov_loads_bt_hwpf", "Loads between HWPF", r2, "loads"),
+ Metric("hwpf_ov_rate", "HWPF per second", r3, "hwpf/s"),
+ ])
+ r1 = d_ratio(hwp_l2, all_pf)
+ r2 = d_ratio(hwp_lc, all_pf)
+ r3 = d_ratio(hwp_lm, all_pf)
+ data_src_local = MetricGroup("hwpf_data_src_local", [
+ Metric("hwpf_data_src_local_l2", "Data source local l2", r1, "100%"),
+ Metric("hwpf_data_src_local_ccx_l3_loc_ccx",
+ "Data source local ccx l3 loc ccx", r2, "100%"),
+ Metric("hwpf_data_src_local_memory_or_io",
+ "Data source local memory or IO", r3, "100%"),
+ ])
+
+ r1 = d_ratio(hwp_rc, all_pf)
+ r2 = d_ratio(hwp_rm, all_pf)
+ data_src_remote = MetricGroup("hwpf_data_src_remote", [
+ Metric("hwpf_data_src_remote_cache", "Data source remote cache", r1,
+ "100%"),
+ Metric("hwpf_data_src_remote_memory_or_io",
+ "Data source remote memory or IO", r2, "100%"),
+ ])
+
+ data_src = MetricGroup("hwpf_data_src", [data_src_local, data_src_remote])
+ return MetricGroup("hwpf", [overview, data_src],
+ description="Hardware prefetch breakdown (CCX L3 = L3 of current thread, Loc CCX = CCX cache on some socket)")
+
+
def AmdSwpf() -> Optional[MetricGroup]:
"""Returns a MetricGroup representing AMD software prefetch metrics."""
if zen_model <= 1:
@@ -256,6 +316,7 @@ def Rapl() -> MetricGroup:
all_metrics = MetricGroup("", [
AmdBr(),
+ AmdHwpf(),
AmdSwpf(),
AmdUpc(),
Idle(),
--
2.44.0.278.ge034bb2e1d-goog
next prev parent reply other threads:[~2024-02-29 0:15 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-02-29 0:15 [PATCH v1 00/13] Python generated AMD Zen metrics Ian Rogers
2024-02-29 0:15 ` [PATCH v1 01/13] perf jevents: Add RAPL event metric for AMD zen models Ian Rogers
2024-02-29 0:15 ` [PATCH v1 02/13] perf jevents: Add idle " Ian Rogers
2024-02-29 0:15 ` [PATCH v1 03/13] perf jevents: Add upc metric for uops per cycle for AMD Ian Rogers
2024-02-29 0:15 ` [PATCH v1 04/13] perf jevents: Add br metric group for branch statistics on AMD Ian Rogers
2024-02-29 0:15 ` [PATCH v1 05/13] perf jevents: Add software prefetch (swpf) metric group for AMD Ian Rogers
2024-02-29 0:15 ` Ian Rogers [this message]
2024-02-29 0:15 ` [PATCH v1 07/13] perf jevents: Add itlb " Ian Rogers
2024-02-29 0:15 ` [PATCH v1 08/13] perf jevents: Add dtlb " Ian Rogers
2024-02-29 0:15 ` [PATCH v1 09/13] perf jevents: Add uncore l3 " Ian Rogers
2024-02-29 0:15 ` [PATCH v1 10/13] perf jevents: Add load store breakdown metrics ldst " Ian Rogers
2024-02-29 0:15 ` [PATCH v1 11/13] perf jevents: Add ILP metrics " Ian Rogers
2024-02-29 0:15 ` [PATCH v1 12/13] perf jevents: Add context switch " Ian Rogers
2024-02-29 0:15 ` [PATCH v1 13/13] perf jevents: Add cycles breakdown metric " Ian Rogers
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240229001537.4158049-7-irogers@google.com \
--to=irogers@google.com \
--cc=acme@kernel.org \
--cc=adrian.hunter@intel.com \
--cc=alexander.shishkin@linux.intel.com \
--cc=eranian@google.com \
--cc=james.clark@arm.com \
--cc=john.g.garry@oracle.com \
--cc=jolsa@kernel.org \
--cc=kan.liang@linux.intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mark.rutland@arm.com \
--cc=mingo@redhat.com \
--cc=namhyung@kernel.org \
--cc=peterz@infradead.org \
--cc=ravi.bangoria@amd.com \
--cc=renyu.zj@linux.alibaba.com \
--cc=sandipan.das@amd.com \
--cc=tmricht@linux.ibm.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).