From 8c64183a461a422b0443f84622af48858d5fe6ed Mon Sep 17 00:00:00 2001 From: Eric Anholt Date: Thu, 26 Mar 2009 17:15:11 -0700 Subject: Initial import of intel-graphics-tools with some microbenchmarks. --- benchmarks/intel_upload_blit_large.c | 157 +++++++++++++++++++++++++++++++++++ 1 file changed, 157 insertions(+) create mode 100644 benchmarks/intel_upload_blit_large.c (limited to 'benchmarks/intel_upload_blit_large.c') diff --git a/benchmarks/intel_upload_blit_large.c b/benchmarks/intel_upload_blit_large.c new file mode 100644 index 00000000..9fc2ee1a --- /dev/null +++ b/benchmarks/intel_upload_blit_large.c @@ -0,0 +1,157 @@ +/* + * Copyright © 2009 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. + * + * Authors: + * Eric Anholt + * + */ + +/** + * Roughly simulates repeatedly uploading frames of images, by uploading + * the data all at once with pwrite, and then blitting it to another buffer. + * + * You might think of this like a movie player, but that wouldn't be entirely + * accurate, since the access patterns of the memory would be different + * (generally, smaller source image, upscaled, an thus different memory access + * pattern in both texel fetch for the stretching and the destination writes). + * However, some things like swfdec would be doing something like this since + * they compute their data in host memory and upload the full sw rendered + * frame. + * + * Additionally, those applications should be rendering at the screen refresh + * rate, while this test has no limits, and so can get itself into the + * working set larger than aperture size performance disaster. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include "drm.h" +#include "i915_drm.h" +#include "drmtest.h" +#include "intel_bufmgr.h" +#include "intel_batchbuffer.h" + +#define OBJECT_WIDTH 1280 +#define OBJECT_HEIGHT 720 + +static double +get_time_in_secs(void) +{ + struct timeval tv; + + gettimeofday(&tv, NULL); + + return (double)tv.tv_sec + tv.tv_usec / 1000000.0; +} + +static void +do_render(drm_intel_bufmgr *bufmgr, struct intel_batchbuffer *batch, + drm_intel_bo *dst_bo, int width, int height) +{ + uint32_t data[width * height]; + drm_intel_bo *src_bo; + int i; + + /* Generate some junk. Real workloads would be doing a lot more + * work to generate the junk. + */ + for (i = 0; i < width * height; i++) { + data[i] = (uint32_t)random(); + } + + /* Upload the junk. */ + src_bo = drm_intel_bo_alloc(bufmgr, "src", sizeof(data), 4096); + drm_intel_bo_subdata(src_bo, 0, sizeof(data), data); + + /* Render the junk to the dst. */ + BEGIN_BATCH(8); + OUT_BATCH(XY_SRC_COPY_BLT_CMD | + XY_SRC_COPY_BLT_WRITE_ALPHA | + XY_SRC_COPY_BLT_WRITE_RGB); + OUT_BATCH((3 << 24) | /* 32 bits */ + (0xcc << 16) | /* copy ROP */ + (width * 4) /* dst pitch */); + OUT_BATCH(0); /* dst x1,y1 */ + OUT_BATCH((height << 16) | width); /* dst x2,y2 */ + OUT_RELOC(dst_bo, I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER, 0); + OUT_BATCH(0); /* src x1,y1 */ + OUT_BATCH(width * 4); /* src pitch */ + OUT_RELOC(src_bo, I915_GEM_DOMAIN_RENDER, 0, 0); + ADVANCE_BATCH(); + + intel_batchbuffer_flush(batch); + + drm_intel_bo_unreference(src_bo); +} + +int main(int argc, char **argv) +{ + int fd; + int object_size = OBJECT_WIDTH * OBJECT_HEIGHT * 4; + double start_time, end_time; + drm_intel_bo *dst_bo; + drm_intel_bufmgr *bufmgr; + struct intel_batchbuffer *batch; + int i; + + fd = drm_open_any(); + + bufmgr = drm_intel_bufmgr_gem_init(fd, 4096); + drm_intel_bufmgr_gem_enable_reuse(bufmgr); + + batch = intel_batchbuffer_alloc(bufmgr); + + dst_bo = drm_intel_bo_alloc(bufmgr, "dst", object_size, 4096); + + /* Prep loop to get us warmed up. */ + for (i = 0; i < 60; i++) { + do_render(bufmgr, batch, dst_bo, OBJECT_WIDTH, OBJECT_HEIGHT); + } + drm_intel_bo_wait_rendering(dst_bo); + + /* Do the actual timing. */ + start_time = get_time_in_secs(); + for (i = 0; i < 200; i++) { + do_render(bufmgr, batch, dst_bo, OBJECT_WIDTH, OBJECT_HEIGHT); + } + drm_intel_bo_wait_rendering(dst_bo); + end_time = get_time_in_secs(); + + printf("%d iterations in %.03f secs: %.01f MB/sec\n", i, + end_time - start_time, + (double)OBJECT_WIDTH * OBJECT_HEIGHT * 4 / + (end_time - start_time)); + + intel_batchbuffer_free(batch); + drm_intel_bufmgr_destroy(bufmgr); + + close(fd); + + return 0; +} -- cgit v1.2.3