From: Chris Wilson <chris@chris-wilson.co.uk> To: igt-dev@lists.freedesktop.org Cc: intel-gfx@lists.freedesktop.org Subject: [PATCH i-g-t 08/17] igt/gem_sync: Show the baseline poll latency for wakeups Date: Mon, 2 Jul 2018 10:07:18 +0100 [thread overview] Message-ID: <20180702090727.7721-8-chris@chris-wilson.co.uk> (raw) In-Reply-To: <20180702090727.7721-1-chris@chris-wilson.co.uk> Distinguish between the latency required to switch away from the pollable spinner into the target nops from the client wakeup of synchronisation on the last nop. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> --- tests/gem_sync.c | 35 +++++++++++++++++++++++++++++++---- 1 file changed, 31 insertions(+), 4 deletions(-) diff --git a/tests/gem_sync.c b/tests/gem_sync.c index 60d61a02e..493ae61df 100644 --- a/tests/gem_sync.c +++ b/tests/gem_sync.c @@ -207,7 +207,7 @@ wakeup_ring(int fd, unsigned ring, int timeout, int wlen) const uint32_t bbe = MI_BATCH_BUFFER_END; struct drm_i915_gem_exec_object2 object; struct drm_i915_gem_execbuffer2 execbuf; - double end, this, elapsed, now; + double end, this, elapsed, now, baseline; unsigned long cycles; uint32_t cmd; igt_spin_t *spin; @@ -233,6 +233,32 @@ wakeup_ring(int fd, unsigned ring, int timeout, int wlen) igt_spin_batch_end(spin); gem_sync(fd, object.handle); + for (int warmup = 0; warmup <= 1; warmup++) { + end = gettime() + timeout/10.; + elapsed = 0; + cycles = 0; + do { + *spin->batch = cmd; + *spin->running = 0; + gem_execbuf(fd, &spin->execbuf); + while (!READ_ONCE(*spin->running)) + ; + + this = gettime(); + igt_spin_batch_end(spin); + gem_sync(fd, spin->handle); + now = gettime(); + + elapsed += now - this; + cycles++; + } while (now < end); + baseline = elapsed / cycles; + } + igt_info("%s%saseline %ld cycles: %.3f us\n", + names[child % num_engines] ?: "", + names[child % num_engines] ? " b" : "B", + cycles, elapsed*1e6/cycles); + end = gettime() + timeout; elapsed = 0; cycles = 0; @@ -254,16 +280,17 @@ wakeup_ring(int fd, unsigned ring, int timeout, int wlen) elapsed += now - this; cycles++; } while (now < end); + elapsed -= cycles * baseline; - igt_info("%s%sompleted %ld cycles: %.3f us\n", + igt_info("%s%sompleted %ld cycles: %.3f + %.3f us\n", names[child % num_engines] ?: "", names[child % num_engines] ? " c" : "C", - cycles, elapsed*1e6/cycles); + cycles, 1e6*baseline, elapsed*1e6/cycles); igt_spin_batch_free(fd, spin); gem_close(fd, object.handle); } - igt_waitchildren_timeout(timeout+10, NULL); + igt_waitchildren_timeout(2*timeout, NULL); igt_assert_eq(intel_detect_and_clear_missed_interrupts(fd), 0); } -- 2.18.0 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx
WARNING: multiple messages have this Message-ID (diff)
From: Chris Wilson <chris@chris-wilson.co.uk> To: igt-dev@lists.freedesktop.org Cc: intel-gfx@lists.freedesktop.org Subject: [igt-dev] [PATCH i-g-t 08/17] igt/gem_sync: Show the baseline poll latency for wakeups Date: Mon, 2 Jul 2018 10:07:18 +0100 [thread overview] Message-ID: <20180702090727.7721-8-chris@chris-wilson.co.uk> (raw) In-Reply-To: <20180702090727.7721-1-chris@chris-wilson.co.uk> Distinguish between the latency required to switch away from the pollable spinner into the target nops from the client wakeup of synchronisation on the last nop. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> --- tests/gem_sync.c | 35 +++++++++++++++++++++++++++++++---- 1 file changed, 31 insertions(+), 4 deletions(-) diff --git a/tests/gem_sync.c b/tests/gem_sync.c index 60d61a02e..493ae61df 100644 --- a/tests/gem_sync.c +++ b/tests/gem_sync.c @@ -207,7 +207,7 @@ wakeup_ring(int fd, unsigned ring, int timeout, int wlen) const uint32_t bbe = MI_BATCH_BUFFER_END; struct drm_i915_gem_exec_object2 object; struct drm_i915_gem_execbuffer2 execbuf; - double end, this, elapsed, now; + double end, this, elapsed, now, baseline; unsigned long cycles; uint32_t cmd; igt_spin_t *spin; @@ -233,6 +233,32 @@ wakeup_ring(int fd, unsigned ring, int timeout, int wlen) igt_spin_batch_end(spin); gem_sync(fd, object.handle); + for (int warmup = 0; warmup <= 1; warmup++) { + end = gettime() + timeout/10.; + elapsed = 0; + cycles = 0; + do { + *spin->batch = cmd; + *spin->running = 0; + gem_execbuf(fd, &spin->execbuf); + while (!READ_ONCE(*spin->running)) + ; + + this = gettime(); + igt_spin_batch_end(spin); + gem_sync(fd, spin->handle); + now = gettime(); + + elapsed += now - this; + cycles++; + } while (now < end); + baseline = elapsed / cycles; + } + igt_info("%s%saseline %ld cycles: %.3f us\n", + names[child % num_engines] ?: "", + names[child % num_engines] ? " b" : "B", + cycles, elapsed*1e6/cycles); + end = gettime() + timeout; elapsed = 0; cycles = 0; @@ -254,16 +280,17 @@ wakeup_ring(int fd, unsigned ring, int timeout, int wlen) elapsed += now - this; cycles++; } while (now < end); + elapsed -= cycles * baseline; - igt_info("%s%sompleted %ld cycles: %.3f us\n", + igt_info("%s%sompleted %ld cycles: %.3f + %.3f us\n", names[child % num_engines] ?: "", names[child % num_engines] ? " c" : "C", - cycles, elapsed*1e6/cycles); + cycles, 1e6*baseline, elapsed*1e6/cycles); igt_spin_batch_free(fd, spin); gem_close(fd, object.handle); } - igt_waitchildren_timeout(timeout+10, NULL); + igt_waitchildren_timeout(2*timeout, NULL); igt_assert_eq(intel_detect_and_clear_missed_interrupts(fd), 0); } -- 2.18.0 _______________________________________________ igt-dev mailing list igt-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/igt-dev
next prev parent reply other threads:[~2018-07-02 9:07 UTC|newest] Thread overview: 68+ messages / expand[flat|nested] mbox.gz Atom feed top 2018-07-02 9:07 [PATCH i-g-t 01/17] lib: Report file cache as available system memory Chris Wilson 2018-07-02 9:07 ` [Intel-gfx] " Chris Wilson 2018-07-02 9:07 ` [PATCH i-g-t 02/17] igt/gem_tiled_partial_pwrite_pread: Check for known swizzling Chris Wilson 2018-07-02 9:07 ` [igt-dev] " Chris Wilson 2018-07-02 12:00 ` Tvrtko Ursulin 2018-07-02 12:00 ` Tvrtko Ursulin 2018-07-05 11:14 ` Chris Wilson 2018-07-05 11:14 ` [Intel-gfx] " Chris Wilson 2018-07-05 12:30 ` Tvrtko Ursulin 2018-07-05 12:30 ` [Intel-gfx] " Tvrtko Ursulin 2018-07-05 12:35 ` Chris Wilson 2018-07-05 12:35 ` [Intel-gfx] " Chris Wilson 2018-07-05 15:26 ` Tvrtko Ursulin 2018-07-05 15:26 ` [Intel-gfx] " Tvrtko Ursulin 2018-07-05 15:55 ` Chris Wilson 2018-07-05 15:55 ` Chris Wilson 2018-07-02 9:07 ` [PATCH i-g-t 03/17] igt/gem_set_tiling_vs_pwrite: Show the erroneous value Chris Wilson 2018-07-02 9:07 ` [Intel-gfx] " Chris Wilson 2018-07-02 12:00 ` [igt-dev] " Tvrtko Ursulin 2018-07-02 12:00 ` Tvrtko Ursulin 2018-07-02 9:07 ` [PATCH i-g-t 04/17] lib: Convert spin batch constructor to a factory Chris Wilson 2018-07-02 9:07 ` [igt-dev] " Chris Wilson 2018-07-02 15:34 ` Tvrtko Ursulin 2018-07-02 15:34 ` Tvrtko Ursulin 2018-07-02 9:07 ` [PATCH i-g-t 05/17] lib: Spin fast, retire early Chris Wilson 2018-07-02 9:07 ` [igt-dev] " Chris Wilson 2018-07-02 15:36 ` Tvrtko Ursulin 2018-07-02 15:36 ` [Intel-gfx] " Tvrtko Ursulin 2018-07-05 11:23 ` Chris Wilson 2018-07-05 11:23 ` [igt-dev] [Intel-gfx] " Chris Wilson 2018-07-05 12:33 ` Tvrtko Ursulin 2018-07-05 12:33 ` [Intel-gfx] " Tvrtko Ursulin 2018-07-05 12:42 ` Chris Wilson 2018-07-05 12:42 ` [igt-dev] [Intel-gfx] " Chris Wilson 2018-07-05 15:29 ` Tvrtko Ursulin 2018-07-05 15:29 ` [igt-dev] [Intel-gfx] " Tvrtko Ursulin 2018-07-05 15:52 ` Chris Wilson 2018-07-05 15:52 ` [igt-dev] [Intel-gfx] " Chris Wilson 2018-07-02 9:07 ` [PATCH i-g-t 06/17] igt/gem_sync: Alternate stress for nop+sync Chris Wilson 2018-07-02 9:07 ` [Intel-gfx] " Chris Wilson 2018-07-02 9:07 ` [PATCH i-g-t 07/17] igt/gem_sync: Double the wakeups, twice the pain Chris Wilson 2018-07-02 9:07 ` [igt-dev] " Chris Wilson 2018-07-02 9:07 ` Chris Wilson [this message] 2018-07-02 9:07 ` [igt-dev] [PATCH i-g-t 08/17] igt/gem_sync: Show the baseline poll latency for wakeups Chris Wilson 2018-07-02 9:07 ` [PATCH i-g-t 09/17] igt/gem_userptr: Check read-only mappings Chris Wilson 2018-07-02 9:07 ` [igt-dev] " Chris Wilson 2018-07-02 9:07 ` [PATCH i-g-t 10/17] igt: Exercise creating context with shared GTT Chris Wilson 2018-07-02 9:07 ` [Intel-gfx] " Chris Wilson 2018-07-02 9:07 ` [PATCH i-g-t 11/17] igt/gem_ctx_switch: Exercise queues Chris Wilson 2018-07-02 9:07 ` [Intel-gfx] " Chris Wilson 2018-07-02 9:07 ` [PATCH i-g-t 12/17] igt/gem_exec_whisper: Fork all-engine tests one-per-engine Chris Wilson 2018-07-02 9:07 ` [igt-dev] " Chris Wilson 2018-07-02 9:07 ` [PATCH i-g-t 13/17] igt: Add gem_ctx_engines Chris Wilson 2018-07-02 9:07 ` [igt-dev] " Chris Wilson 2018-07-02 9:07 ` [PATCH i-g-t 14/17] igt: Add gem_exec_balancer Chris Wilson 2018-07-02 9:07 ` [igt-dev] " Chris Wilson 2018-07-02 9:07 ` [PATCH i-g-t 15/17] benchmarks/wsim: Simulate and interpret .wsim Chris Wilson 2018-07-02 9:07 ` [igt-dev] " Chris Wilson 2018-07-02 9:07 ` [PATCH i-g-t 16/17] tools: capture execution pathways Chris Wilson 2018-07-02 9:07 ` [igt-dev] " Chris Wilson 2018-07-02 9:07 ` [PATCH i-g-t 17/17] igt/gem_exec_latency: Robustify measurements Chris Wilson 2018-07-02 9:07 ` [igt-dev] " Chris Wilson 2018-07-02 11:54 ` [igt-dev] [PATCH i-g-t 01/17] lib: Report file cache as available system memory Tvrtko Ursulin 2018-07-02 11:54 ` Tvrtko Ursulin 2018-07-02 12:08 ` Chris Wilson 2018-07-02 12:08 ` Chris Wilson 2018-07-02 13:09 ` [igt-dev] ✓ Fi.CI.BAT: success for series starting with [i-g-t,01/17] " Patchwork 2018-07-02 14:16 ` [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=20180702090727.7721-8-chris@chris-wilson.co.uk \ --to=chris@chris-wilson.co.uk \ --cc=igt-dev@lists.freedesktop.org \ --cc=intel-gfx@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: linkBe 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.