From: Lucas De Marchi <lucas.demarchi@intel.com>
To: igt-dev@lists.freedesktop.org
Subject: [CI 6/6] gputop: Add support to drm-cycles/drm-total-cycles
Date: Tue, 21 May 2024 09:15:25 -0700 [thread overview]
Message-ID: <20240521161525.159720-6-lucas.demarchi@intel.com> (raw)
In-Reply-To: <20240521161525.159720-1-lucas.demarchi@intel.com>
Add support for using only GPU timestamps to calculate the utilization.
It uses drm-cycles and drm-total-cycles read from fdinfo. For any
2 samples, the utilization is calculated as:
u[i+1] - u[i]
pct = -----------------
Gt[i+1] - Gt[i]
with u == drm-cycles and Gt == drm-total-cycles.
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Link: https://lore.kernel.org/r/20240504064643.25863-14-lucas.demarchi@intel.com
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
---
tools/gputop.c | 41 +++++++++++++++++++++++++++++++++++++----
1 file changed, 37 insertions(+), 4 deletions(-)
diff --git a/tools/gputop.c b/tools/gputop.c
index c06cbe738..6c7cf6927 100644
--- a/tools/gputop.c
+++ b/tools/gputop.c
@@ -31,6 +31,11 @@
#include "igt_drm_fdinfo.h"
#include "drmtest.h"
+enum utilization_type {
+ UTILIZATION_TYPE_ENGINE_TIME,
+ UTILIZATION_TYPE_TOTAL_CYCLES,
+};
+
static const char *bars[] = { " ", "▏", "▎", "▍", "▌", "▋", "▊", "▉", "█" };
static void n_spaces(const unsigned int n)
@@ -174,13 +179,33 @@ print_client(struct igt_drm_client *c, struct igt_drm_client **prevc,
double t, int lines, int con_w, int con_h,
unsigned int period_us, int *engine_w)
{
+ enum utilization_type utilization_type;
unsigned int i;
uint64_t sz;
int len;
+ if (c->utilization_mask & IGT_DRM_CLIENT_UTILIZATION_TOTAL_CYCLES &&
+ c->utilization_mask & IGT_DRM_CLIENT_UTILIZATION_CYCLES)
+ utilization_type = UTILIZATION_TYPE_TOTAL_CYCLES;
+ else if (c->utilization_mask & IGT_DRM_CLIENT_UTILIZATION_ENGINE_TIME)
+ utilization_type = UTILIZATION_TYPE_ENGINE_TIME;
+ else
+ return 0;
+
+ if (c->samples < 2)
+ return 0;
+
/* Filter out idle clients. */
- if (!c->total_engine_time || c->samples < 2)
- return lines;
+ switch (utilization_type) {
+ case UTILIZATION_TYPE_ENGINE_TIME:
+ if (!c->total_engine_time)
+ return 0;
+ break;
+ case UTILIZATION_TYPE_TOTAL_CYCLES:
+ if (!c->total_total_cycles)
+ return 0;
+ break;
+ }
/* Print header when moving to a different DRM card. */
if (newheader(c, *prevc)) {
@@ -211,8 +236,16 @@ print_client(struct igt_drm_client *c, struct igt_drm_client **prevc,
if (!c->engines->capacity[i])
continue;
- pct = (double)c->utilization[i].delta_engine_time / period_us / 1e3 * 100 /
- c->engines->capacity[i];
+ switch (utilization_type) {
+ case UTILIZATION_TYPE_ENGINE_TIME:
+ pct = (double)c->utilization[i].delta_engine_time / period_us / 1e3 * 100 /
+ c->engines->capacity[i];
+ break;
+ case UTILIZATION_TYPE_TOTAL_CYCLES:
+ pct = (double)c->utilization[i].delta_cycles / c->utilization[i].delta_total_cycles * 100 /
+ c->engines->capacity[i];
+ break;
+ }
/*
* Guard against fluctuations between our scanning period and
--
2.43.0
next prev parent reply other threads:[~2024-05-21 16:15 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-05-21 16:15 [CI 1/6] treewide: Rename engine busyness variables Lucas De Marchi
2024-05-21 16:15 ` [CI 2/6] lib/igt_drm_clients: Move engine fields to substruct Lucas De Marchi
2024-05-21 16:15 ` [CI 3/6] lib/igt_drm_clients: Record drm-cycles Lucas De Marchi
2024-05-21 16:15 ` [CI 4/6] lib/igt_drm_fdinfo: Parse drm-total-cycles Lucas De Marchi
2024-05-21 16:15 ` [CI 5/6] lib/igt_drm_clients: Record total cycles Lucas De Marchi
2024-05-21 16:15 ` Lucas De Marchi [this message]
2024-05-21 19:00 ` ✓ Fi.CI.BAT: success for series starting with [CI,1/6] treewide: Rename engine busyness variables Patchwork
2024-05-21 19:41 ` ✗ CI.xeBAT: failure " Patchwork
2024-05-22 2:15 ` ✗ CI.xeFULL: " Patchwork
2024-05-22 8:01 ` ✓ Fi.CI.IGT: success " Patchwork
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=20240521161525.159720-6-lucas.demarchi@intel.com \
--to=lucas.demarchi@intel.com \
--cc=igt-dev@lists.freedesktop.org \
/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