Hide the initial setup cost of a new batch by performing it before we start the clock for measuring the execute-wait latency. Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@xxxxxxxxxxxxxxx> --- tests/gem_sync.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/tests/gem_sync.c b/tests/gem_sync.c index 8ed9760d..c9e2f014 100644 --- a/tests/gem_sync.c +++ b/tests/gem_sync.c @@ -138,6 +138,7 @@ sync_ring(int fd, unsigned ring, int num_children, int timeout) execbuf.buffer_count = 1; execbuf.flags = engines[child % num_engines]; gem_execbuf(fd, &execbuf); + gem_sync(fd, object.handle); start = gettime(); cycles = 0; @@ -553,6 +554,7 @@ sync_all(int fd, int num_children, int timeout) execbuf.buffers_ptr = to_user_pointer(&object); execbuf.buffer_count = 1; gem_execbuf(fd, &execbuf); + gem_sync(fd, object.handle); start = gettime(); cycles = 0; @@ -773,6 +775,7 @@ preempt(int fd, unsigned ring, int num_children, int timeout) execbuf.flags = engines[child % num_engines]; execbuf.rsvd1 = ctx[1]; gem_execbuf(fd, &execbuf); + gem_sync(fd, object.handle); start = gettime(); cycles = 0; -- 2.14.2 _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
