All of lore.kernel.org
 help / color / mirror / Atom feed
From: Katarzyna Dec <katarzyna.dec@intel.com>
To: Chris Wilson <chris@chris-wilson.co.uk>
Cc: igt-dev@lists.freedesktop.org
Subject: Re: [igt-dev] [PATCH i-g-t 05/13] lib: Add GPU power measurement
Date: Tue, 5 Feb 2019 13:27:57 +0100	[thread overview]
Message-ID: <20190205122757.GA20062@kdec5-desk.ger.corp.intel.com> (raw)
In-Reply-To: <20190204083614.2385-5-chris@chris-wilson.co.uk>

On Mon, Feb 04, 2019 at 08:36:06AM +0000, Chris Wilson wrote:
> Read the RAPL power metrics courtesy of perf. Or your local HW
> equivalent?
I think this commit message it too brief. Who should answet the question? :)
> 
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> ---
>  lib/Makefile.sources |   2 +
>  lib/igt_gpu_power.c  | 109 +++++++++++++++++++++++++++++++++++++++++++
>  lib/igt_gpu_power.h  |  59 +++++++++++++++++++++++
>  lib/meson.build      |   2 +
>  4 files changed, 172 insertions(+)
>  create mode 100644 lib/igt_gpu_power.c
>  create mode 100644 lib/igt_gpu_power.h
> 
> diff --git a/lib/Makefile.sources b/lib/Makefile.sources
> index 808b9617e..54d9e3e7b 100644
> --- a/lib/Makefile.sources
> +++ b/lib/Makefile.sources
> @@ -24,6 +24,8 @@ lib_source_list =	 	\
>  	igt_color_encoding.c	\
>  	igt_color_encoding.h	\
>  	igt_edid_template.h	\
> +	igt_gpu_power.c		\
> +	igt_gpu_power.h		\
>  	igt_gt.c		\
>  	igt_gt.h		\
>  	igt_gvt.c		\
> diff --git a/lib/igt_gpu_power.c b/lib/igt_gpu_power.c
> new file mode 100644
> index 000000000..d62bf5dc8
> --- /dev/null
> +++ b/lib/igt_gpu_power.c
> @@ -0,0 +1,109 @@
> +#include <ctype.h>
> +#include <errno.h>
> +#include <fcntl.h>
> +#include <locale.h>
> +#include <math.h>
> +#include <unistd.h>
> +
> +#include "igt_gpu_power.h"
> +#include "igt_perf.h"
> +
> +static int filename_to_buf(const char *filename, char *buf, unsigned int sz)
> +{
> +	int fd;
> +	ssize_t ret;
> +
> +	fd = open(filename, O_RDONLY);
> +	if (fd < 0)
> +		return -1;
> +
> +	ret = read(fd, buf, sz - 1);
> +	close(fd);
> +	if (ret < 1)
> +		return -1;
> +
> +	buf[ret] = '\0';
> +
> +	return 0;
> +}
> +
> +static uint64_t filename_to_u64(const char *filename, int base)
> +{
> +	char buf[64], *b;
> +
> +	if (filename_to_buf(filename, buf, sizeof(buf)))
> +		return 0;
> +
> +	/*
> +	 * Handle both single integer and key=value formats by skipping
> +	 * leading non-digits.
> +	 */
> +	b = buf;
> +	while (*b && !isdigit(*b))
> +		b++;
> +
> +	return strtoull(b, NULL, base);
> +}
> +
> +static double filename_to_double(const char *filename)
> +{
> +	char *oldlocale;
> +	char buf[80];
> +	double v;
> +
> +	if (filename_to_buf(filename, buf, sizeof(buf)))
> +		return 0;
> +
> +	oldlocale = setlocale(LC_ALL, "C");
> +	v = strtod(buf, NULL);
> +	setlocale(LC_ALL, oldlocale);
> +
> +	return v;
> +}
> +
> +static uint64_t rapl_type_id(void)
> +{
> +	return filename_to_u64("/sys/devices/power/type", 10);
> +}
> +
> +static uint64_t rapl_gpu_power(void)
> +{
> +	return filename_to_u64("/sys/devices/power/events/energy-gpu", 0);
> +}
> +
> +static double rapl_gpu_power_scale(void)
> +{
> +	return filename_to_double("/sys/devices/power/events/energy-gpu.scale");
> +}
> +
> +int gpu_power_open(struct gpu_power *power)
> +{
> +	power->fd = igt_perf_open(rapl_type_id(), rapl_gpu_power());
> +	if (power->fd < 0) {
> +		power->fd = -errno;
> +		goto err;
> +	}
> +
> +	power->scale = rapl_gpu_power_scale();
> +	if (isnan(power->scale) || !power->scale) {
> +		close(power->fd);
> +		goto err;
> +	}
> +	power->scale *= 1e9;
Where this scale value is coming from? I looked at this parameter on my desktop
and it is e-10.
> +
> +	return 0;
> +
> +err:
> +	errno = 0;
> +	return power->fd;
> +}
> +
> +bool gpu_power_read(struct gpu_power *power, struct gpu_power_sample *s)
> +{
> +	return read(power->fd, s, sizeof(*s)) == sizeof(*s);
> +}
> +
> +void gpu_power_close(struct gpu_power *power)
> +{
> +	close(power->fd);
> +}
> diff --git a/lib/igt_gpu_power.h b/lib/igt_gpu_power.h
> new file mode 100644
> index 000000000..4e1b747c0
> --- /dev/null
> +++ b/lib/igt_gpu_power.h
> @@ -0,0 +1,59 @@
> +/*
> + * Copyright © 2019 Intel Corporation
> + *
> + * Permission is hereby granted, free of charge, to any person obtaining a
> + * copy of this software and associated documentation files (the "Software"),
> + * to deal in the Software without restriction, including without limitation
> + * the rights to use, copy, modify, merge, publish, distribute, sublicense,
> + * and/or sell copies of the Software, and to permit persons to whom the
> + * Software is furnished to do so, subject to the following conditions:
> + *
> + * The above copyright notice and this permission notice (including the next
> + * paragraph) shall be included in all copies or substantial portions of the
> + * Software.
> + *
> + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
> + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
> + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
> + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
> + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
> + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
> + * IN THE SOFTWARE.
> + *
> + */
> +
> +#ifndef IGT_GPU_POWER_H
> +#define IGT_GPU_POWER_H
> +
> +#include <stdbool.h>
> +#include <stdint.h>
> +
> +struct gpu_power {
> +	int fd;
> +	double scale;
> +};
> +
> +struct gpu_power_sample {
> +	uint64_t energy;
> +	uint64_t time;
> +};
> +
> +int gpu_power_open(struct gpu_power *power);
> +bool gpu_power_read(struct gpu_power *power, struct gpu_power_sample *s);
> +void gpu_power_close(struct gpu_power *power);
> +
> +static inline double gpu_power_J(const struct gpu_power *p,
> +				 const struct gpu_power_sample *t0,
> +				 const struct gpu_power_sample *t1)
> +{
> +	return (t1->energy - t0->energy) * p->scale * 1e-9;
Same a above - where this value comes from?

Kasia :)
> +}
> +
> +static inline double gpu_power_W(const struct gpu_power *p,
> +				 const struct gpu_power_sample *t0,
> +				 const struct gpu_power_sample *t1)
> +{
> +	return (t1->energy - t0->energy) * p->scale / (t1->time - t0->time);
> +}
> +
> +#endif /* IGT_GPU_POWER_H */
> diff --git a/lib/meson.build b/lib/meson.build
> index dd36f8180..63279ae30 100644
> --- a/lib/meson.build
> +++ b/lib/meson.build
> @@ -8,9 +8,11 @@ lib_sources = [
>  	'igt_debugfs.c',
>  	'igt_device.c',
>  	'igt_aux.c',
> +	'igt_gpu_power.c',
>  	'igt_gt.c',
>  	'igt_gvt.c',
>  	'igt_matrix.c',
> +	'igt_perf.c',
>  	'igt_primes.c',
>  	'igt_rand.c',
>  	'igt_stats.c',
> -- 
> 2.20.1
> 
> _______________________________________________
> igt-dev mailing list
> igt-dev@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/igt-dev
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev

  parent reply	other threads:[~2019-02-05 12:28 UTC|newest]

Thread overview: 46+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-02-04  8:36 [PATCH i-g-t 01/13] i915/gem_mocs_settings: Allow hangs around reset tests Chris Wilson
2019-02-04  8:36 ` [igt-dev] " Chris Wilson
2019-02-04  8:36 ` [PATCH i-g-t 02/13] lib/i915: Pretty print busy stats support Chris Wilson
2019-02-04  8:36   ` [igt-dev] " Chris Wilson
2019-02-04  8:36 ` [PATCH i-g-t 03/13] i915/perf_pmu: Query for enhanced busyness tracking Chris Wilson
2019-02-04  8:36   ` [Intel-gfx] " Chris Wilson
2019-02-04  8:36 ` [PATCH i-g-t 04/13] lib/i915: Pretty print HW semaphores Chris Wilson
2019-02-04  8:36   ` [igt-dev] " Chris Wilson
2019-02-04  8:36 ` [PATCH i-g-t 05/13] lib: Add GPU power measurement Chris Wilson
2019-02-04  8:36   ` [igt-dev] " Chris Wilson
2019-02-04 10:05   ` Katarzyna Dec
2019-02-04 10:05     ` Katarzyna Dec
2019-02-04 10:25     ` Chris Wilson
2019-02-04 10:25       ` Chris Wilson
2019-02-05 14:41       ` Ville Syrjälä
2019-02-05 14:41         ` Ville Syrjälä
2019-02-05 15:09         ` Chris Wilson
2019-02-05 15:09           ` Chris Wilson
2019-02-05 12:27   ` Katarzyna Dec [this message]
2019-02-05 12:31     ` Chris Wilson
2019-02-05 13:44       ` Katarzyna Dec
2019-02-04  8:36 ` [PATCH i-g-t 06/13] i915/gem_exec_schedule: Measure semaphore power consumption Chris Wilson
2019-02-04  8:36   ` [igt-dev] " Chris Wilson
2019-02-05 12:50   ` Katarzyna Dec
2019-02-05 12:53     ` Chris Wilson
2019-02-05 13:57       ` Katarzyna Dec
2019-02-04  8:36 ` [PATCH i-g-t 07/13] i915/gem_exec_whisper: Measure total power consumed Chris Wilson
2019-02-04  8:36   ` [Intel-gfx] " Chris Wilson
2019-02-04  8:36 ` [PATCH i-g-t 08/13] i915/gem_exec_latency: Normalize results into ns Chris Wilson
2019-02-04  8:36   ` [igt-dev] " Chris Wilson
2019-02-04  8:36 ` [PATCH i-g-t 09/13] i915/gem_exec_latency: Eliminate the wakeup penalty Chris Wilson
2019-02-04  8:36   ` [igt-dev] " Chris Wilson
2019-02-04  8:36 ` [PATCH i-g-t 10/13] i915/gem_exec_latency: Robustify measurements Chris Wilson
2019-02-04  8:36   ` [igt-dev] " Chris Wilson
2019-02-04  8:36 ` [PATCH i-g-t 11/13] i915/gem_exec_schedule: Verify that using HW semaphores doesn't block Chris Wilson
2019-02-04  8:36   ` [igt-dev] " Chris Wilson
2019-02-04  8:36 ` [PATCH i-g-t 12/13] i915/gem_exec_nop: poll-sequential requires ordering between rings Chris Wilson
2019-02-04  8:36   ` [igt-dev] " Chris Wilson
2019-02-04  8:36 ` [PATCH i-g-t 13/13] i915/gem_sync: Make switch-default asymmetric Chris Wilson
2019-02-04  8:36   ` [igt-dev] " Chris Wilson
2019-02-04  9:16 ` [PATCH i-g-t 01/13] i915/gem_mocs_settings: Allow hangs around reset tests Tvrtko Ursulin
2019-02-04  9:16   ` [igt-dev] [Intel-gfx] " Tvrtko Ursulin
2019-02-04  9:31 ` [igt-dev] " Katarzyna Dec
2019-02-04  9:31   ` Katarzyna Dec
2019-02-04  9:55 ` [igt-dev] ✓ Fi.CI.BAT: success for series starting with [i-g-t,01/13] " Patchwork
2019-02-04 10:45 ` [igt-dev] ✓ Fi.CI.IGT: " 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=20190205122757.GA20062@kdec5-desk.ger.corp.intel.com \
    --to=katarzyna.dec@intel.com \
    --cc=chris@chris-wilson.co.uk \
    --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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.