Chromium Code Reviews
|
|
Chromium Code Reviews
Issue
1869793002:
Ozone GBM: support R_8 format to GpuMemoryBuffers (Closed)
DescriptionOzone GBM: support R_8 format to GpuMemoryBuffers
It makes zero-copy video playback on ChromeOS use native GpuMemoryBuffers.
Comparison of power consumption
- Use Pixel-2.
- Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc
- Use power_supply_info tool, which is software tool, so not very accurate.
- Measure for 1 min, and take sample every 1 sec.
1) software zero-copy video playback
energy rate (W): 16.92
stdev: 1.16
2) native zero-copy video playback
energy rate (W): 16.16
stdev: 1.21
Native zero-copy video playback seems to save power but stdev is bigger than
difference. Need to measure using power meter.
BUG=356871
TEST=amd64-generic on HSW/BDW/SKL
content_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests
gpu_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests
gl_unittests --gtest_filter=GLImage* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
Committed: https://crrev.com/eb34f43bb954faa9b61d6d94933549c3b392a9fd
Cr-Commit-Position: refs/heads/master@{#392022}
Patch Set 1 #Patch Set 2 : rebase to ToT #
Total comments: 4
Patch Set 3 : guard #ifndef #
Total comments: 2
Patch Set 4 : resolve nits #
Total comments: 3
Patch Set 5 : comment Intel, not i915 #
Total comments: 2
Messages
Total messages: 42 (18 generated)
Description was changed from ========== Ozone GBM: zero-copy video playback on ChromeOS using native GpuMemoryBuffers BUG=356871 TEST=amd64-generic on HSW/BDW/SKL ========== to ========== Ozone GBM: zero-copy video playback on ChromeOS using native GpuMemoryBuffers BUG=356871 TEST=amd64-generic on HSW/BDW/SKL ==========
The CQ bit was checked by dongseong.hwang@intel.com to run a CQ dry run
Dry run: CQ is trying da patch. Follow status at https://chromium-cq-status.appspot.com/patch-status/1869793002/1 View timeline at https://chromium-cq-status.appspot.com/patch-timeline/1869793002/1
dnicoara, could you review? This CL is blocked by https://chromium-review.googlesource.com/#/c/337720/1 https://chromium-review.googlesource.com/#/c/337730/1
dongseong.hwang@intel.com changed reviewers: + marcheu@chromium.org
The CQ bit was unchecked by commit-bot@chromium.org
Dry run: Try jobs failed on following builders: linux_chromium_compile_dbg_32_ng on tryserver.chromium.linux (JOB_FAILED, http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...) ios_dbg_simulator_ninja on tryserver.chromium.mac (JOB_FAILED, http://build.chromium.org/p/tryserver.chromium.mac/builders/ios_dbg_simulator...) ios_rel_device_gn on tryserver.chromium.mac (JOB_FAILED, http://build.chromium.org/p/tryserver.chromium.mac/builders/ios_rel_device_gn...) ios_rel_device_ninja on tryserver.chromium.mac (JOB_FAILED, http://build.chromium.org/p/tryserver.chromium.mac/builders/ios_rel_device_ni...) mac_chromium_compile_dbg_ng on tryserver.chromium.mac (JOB_FAILED, http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_comp...) mac_chromium_gn_rel on tryserver.chromium.mac (JOB_FAILED, http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_gn_r...) mac_chromium_rel_ng on tryserver.chromium.mac (JOB_FAILED, http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)
https://codereview.chromium.org/1869793002/diff/20001/ui/ozone/platform/drm/c... File ui/ozone/platform/drm/common/drm_util.cc (right): https://codereview.chromium.org/1869793002/diff/20001/ui/ozone/platform/drm/c... ui/ozone/platform/drm/common/drm_util.cc:21: #define DRM_FORMAT_R8 fourcc_code('R', '8', ' ', ' ') Surround it with #if-defs to avoid warnings when it is added to libdrm. https://codereview.chromium.org/1869793002/diff/20001/ui/ozone/platform/drm/c... ui/ozone/platform/drm/common/drm_util.cc:335: int GetFourCCFormatForFramebuffer(gfx::BufferFormat format) { Shouldn't you add the format in here as well?
Description was changed from ========== Ozone GBM: zero-copy video playback on ChromeOS using native GpuMemoryBuffers BUG=356871 TEST=amd64-generic on HSW/BDW/SKL ========== to ========== Ozone GBM: zero-copy video playback on ChromeOS using native GpuMemoryBuffers Comparison of power consumption - Use Pixel-2. - Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc - Use power_supply_info tool, which is software tool, so not very accurate. - Measure for 1 min, and take sample every 1 sec. 1) software zero-copy video playback energy rate (W): 16.92 stdev: 1.16 2) native zero-copy video playback energy rate (W): 16.16 stdev: 1.21 Native zero-copy video playback seems to save power but stdev is bigger than difference. BUG=356871 TEST=amd64-generic on HSW/BDW/SKL ==========
Description was changed from ========== Ozone GBM: zero-copy video playback on ChromeOS using native GpuMemoryBuffers Comparison of power consumption - Use Pixel-2. - Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc - Use power_supply_info tool, which is software tool, so not very accurate. - Measure for 1 min, and take sample every 1 sec. 1) software zero-copy video playback energy rate (W): 16.92 stdev: 1.16 2) native zero-copy video playback energy rate (W): 16.16 stdev: 1.21 Native zero-copy video playback seems to save power but stdev is bigger than difference. BUG=356871 TEST=amd64-generic on HSW/BDW/SKL ========== to ========== Ozone GBM: zero-copy video playback on ChromeOS using native GpuMemoryBuffers Comparison of power consumption - Use Pixel-2. - Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc - Use power_supply_info tool, which is software tool, so not very accurate. - Measure for 1 min, and take sample every 1 sec. 1) software zero-copy video playback energy rate (W): 16.92 stdev: 1.16 2) native zero-copy video playback energy rate (W): 16.16 stdev: 1.21 Native zero-copy video playback seems to save power but stdev is bigger than difference. Need to measure using power meter. BUG=356871 TEST=amd64-generic on HSW/BDW/SKL ==========
Thanks for reviewing. I added benchmark data in the description. https://codereview.chromium.org/1869793002/diff/20001/ui/ozone/platform/drm/c... File ui/ozone/platform/drm/common/drm_util.cc (right): https://codereview.chromium.org/1869793002/diff/20001/ui/ozone/platform/drm/c... ui/ozone/platform/drm/common/drm_util.cc:21: #define DRM_FORMAT_R8 fourcc_code('R', '8', ' ', ' ') On 2016/04/07 15:57:16, dnicoara wrote: > Surround it with #if-defs to avoid warnings when it is added to libdrm. Done. guard by #ifndef DRM_FORMAT_R8 https://codereview.chromium.org/1869793002/diff/20001/ui/ozone/platform/drm/c... ui/ozone/platform/drm/common/drm_util.cc:335: int GetFourCCFormatForFramebuffer(gfx::BufferFormat format) { On 2016/04/07 15:57:16, dnicoara wrote: > Shouldn't you add the format in here as well? This function is for scanout buffer, which is used for hardware overlay. All vendor's DRM in kernel don't support hardware overlay for R8 and GR8.
The CQ bit was checked by dongseong.hwang@intel.com to run a CQ dry run
Description was changed from ========== Ozone GBM: zero-copy video playback on ChromeOS using native GpuMemoryBuffers Comparison of power consumption - Use Pixel-2. - Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc - Use power_supply_info tool, which is software tool, so not very accurate. - Measure for 1 min, and take sample every 1 sec. 1) software zero-copy video playback energy rate (W): 16.92 stdev: 1.16 2) native zero-copy video playback energy rate (W): 16.16 stdev: 1.21 Native zero-copy video playback seems to save power but stdev is bigger than difference. Need to measure using power meter. BUG=356871 TEST=amd64-generic on HSW/BDW/SKL ========== to ========== Ozone GBM: zero-copy video playback on ChromeOS using native GpuMemoryBuffers Comparison of power consumption - Use Pixel-2. - Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc - Use power_supply_info tool, which is software tool, so not very accurate. - Measure for 1 min, and take sample every 1 sec. 1) software zero-copy video playback energy rate (W): 16.92 stdev: 1.16 2) native zero-copy video playback energy rate (W): 16.16 stdev: 1.21 Native zero-copy video playback seems to save power but stdev is bigger than difference. Need to measure using power meter. BUG=356871 TEST=amd64-generic on HSW/BDW/SKL gpu_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests ==========
Dry run: CQ is trying da patch. Follow status at https://chromium-cq-status.appspot.com/patch-status/1869793002/40001 View timeline at https://chromium-cq-status.appspot.com/patch-timeline/1869793002/40001
The CQ bit was unchecked by commit-bot@chromium.org
Dry run: This issue passed the CQ dry run.
Can you update ui/gl/gl_image_ozone_native_pixmap_unittest.cc to test R8 too? Something similar to what I was trying to do in crrev.com/1660413004
On 2016/04/07 18:46:31, Daniele Castagna wrote: > Can you update ui/gl/gl_image_ozone_native_pixmap_unittest.cc to test R8 too? > Something similar to what I was trying to do in crrev.com/1660413004 That's good idea! Ozone requires sorta big code change to support gl_image unittests. I'm working on https://codereview.chromium.org/1484473003 I added R_8 to the CL.
Description was changed from ========== Ozone GBM: zero-copy video playback on ChromeOS using native GpuMemoryBuffers Comparison of power consumption - Use Pixel-2. - Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc - Use power_supply_info tool, which is software tool, so not very accurate. - Measure for 1 min, and take sample every 1 sec. 1) software zero-copy video playback energy rate (W): 16.92 stdev: 1.16 2) native zero-copy video playback energy rate (W): 16.16 stdev: 1.21 Native zero-copy video playback seems to save power but stdev is bigger than difference. Need to measure using power meter. BUG=356871 TEST=amd64-generic on HSW/BDW/SKL gpu_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests ========== to ========== Ozone GBM: zero-copy video playback on ChromeOS using native GpuMemoryBuffers Comparison of power consumption - Use Pixel-2. - Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc - Use power_supply_info tool, which is software tool, so not very accurate. - Measure for 1 min, and take sample every 1 sec. 1) software zero-copy video playback energy rate (W): 16.92 stdev: 1.16 2) native zero-copy video playback energy rate (W): 16.16 stdev: 1.21 Native zero-copy video playback seems to save power but stdev is bigger than difference. Need to measure using power meter. BUG=356871 TEST=amd64-generic on HSW/BDW/SKL content_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gpu_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gl_unittests --gtest_filter=GLImage* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests ==========
Description was changed from ========== Ozone GBM: zero-copy video playback on ChromeOS using native GpuMemoryBuffers Comparison of power consumption - Use Pixel-2. - Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc - Use power_supply_info tool, which is software tool, so not very accurate. - Measure for 1 min, and take sample every 1 sec. 1) software zero-copy video playback energy rate (W): 16.92 stdev: 1.16 2) native zero-copy video playback energy rate (W): 16.16 stdev: 1.21 Native zero-copy video playback seems to save power but stdev is bigger than difference. Need to measure using power meter. BUG=356871 TEST=amd64-generic on HSW/BDW/SKL content_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gpu_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gl_unittests --gtest_filter=GLImage* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests ========== to ========== Ozone GBM: zero-copy video playback on ChromeOS using native GpuMemoryBuffers Comparison of power consumption - Use Pixel-2. - Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc - Use power_supply_info tool, which is software tool, so not very accurate. - Measure for 1 min, and take sample every 1 sec. 1) software zero-copy video playback energy rate (W): 16.92 stdev: 1.16 2) native zero-copy video playback energy rate (W): 16.16 stdev: 1.21 Native zero-copy video playback seems to save power but stdev is bigger than difference. Need to measure using power meter. BUG=356871 TEST=amd64-generic on HSW/BDW/SKL content_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gpu_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gl_unittests --gtest_filter=GLImage* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel ==========
Patchset #4 (id:60001) has been deleted
On 2016/04/08 at 13:23:47, dongseong.hwang wrote: > On 2016/04/07 18:46:31, Daniele Castagna wrote: > > Can you update ui/gl/gl_image_ozone_native_pixmap_unittest.cc to test R8 too? > > Something similar to what I was trying to do in crrev.com/1660413004 > > That's good idea! > Ozone requires sorta big code change to support gl_image unittests. I'm working on https://codereview.chromium.org/1484473003 > I added R_8 to the CL. Nice. I'd suggest to change the title of the CL. While enabling native R8 will be used by videos, this CL is not really video specific in any way. Let's wait for Stephane's opinion on https://chromium-review.googlesource.com/337730 before moving forward though.
Description was changed from ========== Ozone GBM: zero-copy video playback on ChromeOS using native GpuMemoryBuffers Comparison of power consumption - Use Pixel-2. - Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc - Use power_supply_info tool, which is software tool, so not very accurate. - Measure for 1 min, and take sample every 1 sec. 1) software zero-copy video playback energy rate (W): 16.92 stdev: 1.16 2) native zero-copy video playback energy rate (W): 16.16 stdev: 1.21 Native zero-copy video playback seems to save power but stdev is bigger than difference. Need to measure using power meter. BUG=356871 TEST=amd64-generic on HSW/BDW/SKL content_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gpu_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gl_unittests --gtest_filter=GLImage* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel ========== to ========== Ozone GBM: support R_8 format to GpuMemoryBuffers It makes zero-copy video playback on ChromeOS use native GpuMemoryBuffers. Comparison of power consumption - Use Pixel-2. - Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc - Use power_supply_info tool, which is software tool, so not very accurate. - Measure for 1 min, and take sample every 1 sec. 1) software zero-copy video playback energy rate (W): 16.92 stdev: 1.16 2) native zero-copy video playback energy rate (W): 16.16 stdev: 1.21 Native zero-copy video playback seems to save power but stdev is bigger than difference. Need to measure using power meter. BUG=356871 TEST=amd64-generic on HSW/BDW/SKL content_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gpu_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gl_unittests --gtest_filter=GLImage* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel ==========
On 2016/04/08 18:47:04, Daniele Castagna wrote: > On 2016/04/08 at 13:23:47, dongseong.hwang wrote: > I'd suggest to change the title of the CL. While enabling native R8 will be used > by videos, this CL is not really video specific in any way. Done. > Let's wait for Stephane's opinion on > https://chromium-review.googlesource.com/337730 before moving forward though. Yes, let's wait :)
The minigbm patch will be landed soon https://chromium-review.googlesource.com/#/c/337730/ dnicoara, could you review again?
lgtm https://codereview.chromium.org/1869793002/diff/40001/ui/ozone/platform/drm/c... File ui/ozone/platform/drm/common/drm_util.cc (right): https://codereview.chromium.org/1869793002/diff/40001/ui/ozone/platform/drm/c... ui/ozone/platform/drm/common/drm_util.cc:20: #ifndef DRM_FORMAT_R8 nit: #if !defined(DRM_FORMAT_R8)
thx for reviewing. I narrow R_8 support to only Intel architecture. dnicoara, could you check again? https://codereview.chromium.org/1869793002/diff/40001/ui/ozone/platform/drm/c... File ui/ozone/platform/drm/common/drm_util.cc (right): https://codereview.chromium.org/1869793002/diff/40001/ui/ozone/platform/drm/c... ui/ozone/platform/drm/common/drm_util.cc:20: #ifndef DRM_FORMAT_R8 On 2016/05/03 14:05:50, dnicoara wrote: > nit: #if !defined(DRM_FORMAT_R8) Done. https://codereview.chromium.org/1869793002/diff/80001/ui/ozone/platform/drm/c... File ui/ozone/platform/drm/client_native_pixmap_factory_gbm.cc (right): https://codereview.chromium.org/1869793002/diff/80001/ui/ozone/platform/drm/c... ui/ozone/platform/drm/client_native_pixmap_factory_gbm.cc:56: format == gfx::BufferFormat::R_8 || narrow R_8 support to only Intel architecture, because only Intel driver supports R_8 unlike BGRA_8888, which all drivers support.
lgtm
https://codereview.chromium.org/1869793002/diff/80001/ui/ozone/platform/drm/c... File ui/ozone/platform/drm/client_native_pixmap_factory_gbm.cc (right): https://codereview.chromium.org/1869793002/diff/80001/ui/ozone/platform/drm/c... ui/ozone/platform/drm/client_native_pixmap_factory_gbm.cc:56: format == gfx::BufferFormat::R_8 || On 2016/05/04 at 07:28:36, dshwang wrote: > narrow R_8 support to only Intel architecture, because only Intel driver supports R_8 unlike BGRA_8888, which all drivers support. Is it true that if the processor architecture is X86 we're always using i915 drivers? If that is true on ChromeOS, then LGTM.
https://codereview.chromium.org/1869793002/diff/80001/ui/ozone/platform/drm/c... File ui/ozone/platform/drm/client_native_pixmap_factory_gbm.cc (right): https://codereview.chromium.org/1869793002/diff/80001/ui/ozone/platform/drm/c... ui/ozone/platform/drm/client_native_pixmap_factory_gbm.cc:56: format == gfx::BufferFormat::R_8 || On 2016/05/04 19:44:02, Daniele Castagna wrote: > On 2016/05/04 at 07:28:36, dshwang wrote: > > narrow R_8 support to only Intel architecture, because only Intel driver > supports R_8 unlike BGRA_8888, which all drivers support. > > Is it true that if the processor architecture is X86 we're always using i915 > drivers? > > If that is true on ChromeOS, then LGTM. Yes, All Intel architecture uses i915 in minigbm. There is some naming confusion between i965 and i915. There are two user level libraries; i965_dri.so and i915_dri.so However, in the kernel, there is only one implementation, i915.ko Unfortunately, i965_dri.so talks with i915.ko. This confusing naming causes many misleading. i915 in minigbm deals with both i915 and i965. FYI, since chromeos was introduced, all Intel architecture has used i965. i915 in comment is a bit wrong, in the sense. I'll change it to "Intel driver".
https://codereview.chromium.org/1869793002/diff/100001/ui/ozone/platform/drm/... File ui/ozone/platform/drm/client_native_pixmap_factory_gbm.cc (right): https://codereview.chromium.org/1869793002/diff/100001/ui/ozone/platform/drm/... ui/ozone/platform/drm/client_native_pixmap_factory_gbm.cc:53: #if defined(ARCH_CPU_X86_FAMILY) what does it take to not have this ifdef? ie. what's missing to support this on all devices that support native GMBs.
https://codereview.chromium.org/1869793002/diff/100001/ui/ozone/platform/drm/... File ui/ozone/platform/drm/client_native_pixmap_factory_gbm.cc (right): https://codereview.chromium.org/1869793002/diff/100001/ui/ozone/platform/drm/... ui/ozone/platform/drm/client_native_pixmap_factory_gbm.cc:53: #if defined(ARCH_CPU_X86_FAMILY) On 2016/05/05 09:29:24, reveman wrote: > what does it take to not have this ifdef? ie. what's missing to support this on > all devices that support native GMBs. gbm bo creation fails on ARM device. ARM has to support R8 format in the 3D driver. e.g. Mesa https://lists.freedesktop.org/archives/mesa-commit/2015-July/057676.html
On 2016/05/05 at 12:22:13, dongseong.hwang wrote: > https://codereview.chromium.org/1869793002/diff/100001/ui/ozone/platform/drm/... > File ui/ozone/platform/drm/client_native_pixmap_factory_gbm.cc (right): > > https://codereview.chromium.org/1869793002/diff/100001/ui/ozone/platform/drm/... > ui/ozone/platform/drm/client_native_pixmap_factory_gbm.cc:53: #if defined(ARCH_CPU_X86_FAMILY) > On 2016/05/05 09:29:24, reveman wrote: > > what does it take to not have this ifdef? ie. what's missing to support this on > > all devices that support native GMBs. > > gbm bo creation fails on ARM device. > > ARM has to support R8 format in the 3D driver. e.g. Mesa https://lists.freedesktop.org/archives/mesa-commit/2015-July/057676.html Is anyone working on this? How much work is it to get this done?
On 2016/05/05 16:34:20, reveman wrote: > On 2016/05/05 at 12:22:13, dongseong.hwang wrote: > > > https://codereview.chromium.org/1869793002/diff/100001/ui/ozone/platform/drm/... > > File ui/ozone/platform/drm/client_native_pixmap_factory_gbm.cc (right): > > > > > https://codereview.chromium.org/1869793002/diff/100001/ui/ozone/platform/drm/... > > ui/ozone/platform/drm/client_native_pixmap_factory_gbm.cc:53: #if > defined(ARCH_CPU_X86_FAMILY) > > On 2016/05/05 09:29:24, reveman wrote: > > > what does it take to not have this ifdef? ie. what's missing to support this > on > > > all devices that support native GMBs. > > > > gbm bo creation fails on ARM device. > > > > ARM has to support R8 format in the 3D driver. e.g. Mesa > https://lists.freedesktop.org/archives/mesa-commit/2015-July/057676.html > > Is anyone working on this? How much work is it to get this done? As far as I know, nobody is working on. I filed mesa bug https://bugs.freedesktop.org/show_bug.cgi?id=95294 I don't know whether DRI or Gallium need to be touched and if Mesa change is sufficient for ARM ChromeOS. So I cannot say how much work is needed.
The CQ bit was checked by dongseong.hwang@intel.com
The patchset sent to the CQ was uploaded after l-g-t-m from dnicoara@chromium.org, dcastagna@chromium.org Link to the patchset: https://codereview.chromium.org/1869793002/#ps100001 (title: "comment Intel, not i915")
CQ is trying da patch. Follow status at https://chromium-cq-status.appspot.com/patch-status/1869793002/100001 View timeline at https://chromium-cq-status.appspot.com/patch-timeline/1869793002/100001
Message was sent while issue was closed.
Description was changed from ========== Ozone GBM: support R_8 format to GpuMemoryBuffers It makes zero-copy video playback on ChromeOS use native GpuMemoryBuffers. Comparison of power consumption - Use Pixel-2. - Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc - Use power_supply_info tool, which is software tool, so not very accurate. - Measure for 1 min, and take sample every 1 sec. 1) software zero-copy video playback energy rate (W): 16.92 stdev: 1.16 2) native zero-copy video playback energy rate (W): 16.16 stdev: 1.21 Native zero-copy video playback seems to save power but stdev is bigger than difference. Need to measure using power meter. BUG=356871 TEST=amd64-generic on HSW/BDW/SKL content_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gpu_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gl_unittests --gtest_filter=GLImage* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel ========== to ========== Ozone GBM: support R_8 format to GpuMemoryBuffers It makes zero-copy video playback on ChromeOS use native GpuMemoryBuffers. Comparison of power consumption - Use Pixel-2. - Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc - Use power_supply_info tool, which is software tool, so not very accurate. - Measure for 1 min, and take sample every 1 sec. 1) software zero-copy video playback energy rate (W): 16.92 stdev: 1.16 2) native zero-copy video playback energy rate (W): 16.16 stdev: 1.21 Native zero-copy video playback seems to save power but stdev is bigger than difference. Need to measure using power meter. BUG=356871 TEST=amd64-generic on HSW/BDW/SKL content_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gpu_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gl_unittests --gtest_filter=GLImage* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel ==========
Message was sent while issue was closed.
Committed patchset #5 (id:100001)
Message was sent while issue was closed.
Description was changed from ========== Ozone GBM: support R_8 format to GpuMemoryBuffers It makes zero-copy video playback on ChromeOS use native GpuMemoryBuffers. Comparison of power consumption - Use Pixel-2. - Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc - Use power_supply_info tool, which is software tool, so not very accurate. - Measure for 1 min, and take sample every 1 sec. 1) software zero-copy video playback energy rate (W): 16.92 stdev: 1.16 2) native zero-copy video playback energy rate (W): 16.16 stdev: 1.21 Native zero-copy video playback seems to save power but stdev is bigger than difference. Need to measure using power meter. BUG=356871 TEST=amd64-generic on HSW/BDW/SKL content_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gpu_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gl_unittests --gtest_filter=GLImage* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel ========== to ========== Ozone GBM: support R_8 format to GpuMemoryBuffers It makes zero-copy video playback on ChromeOS use native GpuMemoryBuffers. Comparison of power consumption - Use Pixel-2. - Use 1440p60 HD video; https://www.youtube.com/embed/UceRgEyfSsc - Use power_supply_info tool, which is software tool, so not very accurate. - Measure for 1 min, and take sample every 1 sec. 1) software zero-copy video playback energy rate (W): 16.92 stdev: 1.16 2) native zero-copy video playback energy rate (W): 16.16 stdev: 1.21 Native zero-copy video playback seems to save power but stdev is bigger than difference. Need to measure using power meter. BUG=356871 TEST=amd64-generic on HSW/BDW/SKL content_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gpu_unittests --gtest_filter=GpuMemoryBuffer* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests gl_unittests --gtest_filter=GLImage* --ozone-platform=gbm --ozone-use-surfaceless --single-process-tests CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel Committed: https://crrev.com/eb34f43bb954faa9b61d6d94933549c3b392a9fd Cr-Commit-Position: refs/heads/master@{#392022} ==========
Message was sent while issue was closed.
Patchset 5 (id:??) landed as https://crrev.com/eb34f43bb954faa9b61d6d94933549c3b392a9fd Cr-Commit-Position: refs/heads/master@{#392022} |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
