New issue
Advanced search Search tips

Issue 614704 link

Starred by 5 users

Issue metadata

Status: Fixed
Owner:
Closed: Jun 2016
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Windows
Pri: 2
Type: Compat



Sign in to add a comment

Chrome 50 - slow and laggy Farmville two after upgrade to Chrome 50

Reported by sarves...@gmail.com, May 25 2016

Issue description

UserAgent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_10_4) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/50.0.2661.94 Safari/537.36

Example URL:
http://apps.facebook.com/farmville-two

Steps to reproduce the problem:
1. Log to Farmville two on Chrome 50 on Windows 7 or Windows 10
2. The game is slower as compared to Chrome 49
3. Firefox does not have the issue

What is the expected behavior?
Game should have been as fast or similar to earlier Chrome versions, In Firefox the game has not slowed down

What went wrong?
Once the browser upgraded to Chrome 50, players started facing this issue. Our players migrated to Chrome on May 3rd 2016(even though it released on April 28th) and we have the spike on same date. However on Firefox the same spike is not present.  Attached images show the spike on May 3rd. 

The data is hourly sampled average Frame time for the players. More frame time indicates worse FPS. We have lot of players complaining of slow FPS and CS Tickets post May 3rd after the Chrome migration.

Does it occur on multiple sites: N/A

Is it a problem with a plugin? Yes Flash

Did this work before? N/A 

Does this work in other browsers? Yes 

Chrome version: 50.0.2661.94  Channel: stable
OS Version: OS X 10.10.4
Flash Version: Shockwave Flash 21.0 r0
 
Chrome Windows 7.png
224 KB View Download
Chrome Windows 10.png
290 KB View Download
Firefox.png
313 KB View Download

Comment 1 by ihf@chromium.org, May 25 2016

Can you tell us more about affected system configurations? In particular attaching the output of about:gpu form one of the slow systems.

Comment 2 Deleted

@sarvesh.n: Could you please respond to comment #1.

Thank you.

Comment 4 by sarves...@gmail.com, May 26 2016

We are reaching out to getting the details of about:gpu from some of the affected players.

We are still trying to replicate this in our environment, however our stats data from players is showing a spike in Chrome on Windows 10 and Windows 7 the same time our players migrated to Chrome version. The spike isn’t there in Firefox

We had a similar issue a year ago and this was resolved with a fix in PPAPI code base.
https://code.google.com/p/chromium/issues/detail?id=496699

We do see a lot of PPAPI changes from 49 to 50 which might have caused the issue. Do you think any of these would have affected Flash plugin only on windows?

Comment 5 Deleted

Comment 6 by ihf@chromium.org, May 27 2016

Cc: vmi...@chromium.org
Labels: M-50

Comment 7 Deleted

Comment 8 by sarves...@gmail.com, May 27 2016

Below are about:gpu details on Windows 10 for one of the affected users


-
 Canvas: Hardware accelerated 
- Flash: Hardware accelerated 
- Flash Stage3D: Hardware accelerated 
- Flash Stage3D Baseline profile: Hardware accelerated 
- Compositing: Hardware accelerated 
- Multiple Raster Threads: Disabled 
- Native GpuMemoryBuffers: Software only. Hardware acceleration disabled 
- Rasterization: Software only. Hardware acceleration disabled 
- Video Decode: Hardware accelerated 
- Video Encode: Hardware accelerated 
- WebGL: Hardware accelerated 
 
Driver Bug Workarounds 
 
- clear_uniforms_before_first_program_use 
- disable_d3d11 
- disable_discard_framebuffer 
- exit_on_context_lost 
- force_cube_complete 
- scalarize_vec_and_mat_constructor_args 
- texsubimage_faster_than_teximage 
 
Problems Detected 
 
- Some drivers are unable to reset the D3D device in the GPU process sandbox 
Applied Workarounds: exit_on_context_lost 
- TexSubImage is faster for full uploads on ANGLE 
Applied Workarounds: texsubimage_faster_than_teximage 
- Clear uniforms before first program use on all platforms: 124764, 349137 
Applied Workarounds: clear_uniforms_before_first_program_use 
- Always rewrite vec/mat constructors to be consistent: 398694 
Applied Workarounds: scalarize_vec_and_mat_constructor_args 
- ANGLE crash on glReadPixels from incomplete cube map texture: 518889 
Applied Workarounds: force_cube_complete 
- Framebuffer discarding can hurt performance on non-tilers: 570897 
Applied Workarounds: disable_discard_framebuffer 
- Crashes in D3D11 on specific AMD drivers: 517040 
Applied Workarounds: disable_d3d11 
- Accelerated rasterization has been disabled, either via about:flags or command line. 
Disabled Features: rasterization 
- Raster is using a single thread. 
Disabled Features: multiple_raster_threads 
- Native GpuMemoryBuffers have been disabled, either via about:flags or command line. 
Disabled Features: native_gpu_memory_buffers 
 
Version Information 
 
| Data exported | 5/26/2016, 5:23:53 PM | 
| Chrome version | Chrome/50.0.2661.102 | 
| Operating system | Windows NT 10.0 | 
| Software rendering list version | 10.18 | 
| Driver bug list version | 8.59 | 
| ANGLE commit id | c46018b8598d | 
| 2D graphics backend | Skia | 
| Command Line Args | Files (x86)\Google\Chrome\Application\chrome.exe" --flag-switches-begin --flag-switches-end | 
 
 
Driver Information 
 
| Initialization time | 197 | 
| In-process GPU | false | 
| Sandboxed | false | 
| GPU0 | VENDOR = 0x1002, DEVICE= 0x9802 | 
| Optimus | false | 
| AMD switchable | false | 
| Desktop compositing | Aero Glass | 
| Diagonal Monitor Size of \\.\DISPLAY1 | 15.5" | 
| Driver vendor | Advanced Micro Devices, Inc. | 
| Driver version | 15.201.1151.0 | 
| Driver date | 8-21-2015 | 
| Pixel shader version | 3.0 | 
| Vertex shader version | 3.0 | 
| Max. MSAA samples | 4 | 
| Machine model name | 
| 
| Machine model version | 
| 
| GL_VENDOR | Google Inc. | 
| GL_RENDERER | ANGLE (AMD Radeon HD 6310 Graphics Direct3D9Ex vs_3_0 ps_3_0) | 
| GL_VERSION | OpenGL ES 2.0 (ANGLE 2.1.0.c46018b8598d) | 
| GL_EXTENSIONS | GL_OES_element_index_uint GL_OES_packed_depth_stencil GL_OES_get_program_binary GL_OES_rgb8_rgba8 GL_EXT_texture_format_BGRA8888 GL_EXT_read_format_bgra GL_EXT_color_buffer_half_float GL_OES_texture_half_float GL_OES_texture_half_float_linear GL_OES_texture_float GL_OES_texture_float_linear GL_EXT_texture_compression_dxt1 GL_ANGLE_texture_compression_dxt3 GL_ANGLE_texture_compression_dxt5 GL_OES_depth32 GL_EXT_texture_storage GL_OES_texture_npot GL_EXT_texture_filter_anisotropic GL_EXT_occlusion_query_boolean GL_NV_fence GL_EXT_robustness GL_EXT_blend_minmax GL_ANGLE_framebuffer_blit GL_ANGLE_framebuffer_multisample GL_ANGLE_instanced_arrays GL_ANGLE_pack_reverse_row_order GL_OES_standard_derivatives GL_EXT_shader_texture_lod GL_EXT_frag_depth GL_ANGLE_texture_usage GL_ANGLE_translated_shader_source GL_EXT_debug_marker GL_OES_EGL_image GL_EXT_unpack_subimage GL_NV_pack_subimage GL_OES_vertex_array_object GL_KHR_debug | 
| Disabled Extensions | 
| 
| Window system binding vendor | Google Inc. (adapter LUID: 000000000000d264) | 
| Window system binding version | 1.4 (ANGLE 2.1.0.c46018b8598d) | 
| Window system binding extensions | EGL_EXT_create_context_robustness EGL_ANGLE_d3d_share_handle_client_buffer EGL_ANGLE_surface_d3d_texture_2d_share_handle EGL_ANGLE_query_surface_pointer EGL_ANGLE_window_fixed_size EGL_NV_post_sub_buffer EGL_KHR_create_context EGL_EXT_device_query EGL_KHR_image EGL_KHR_image_base EGL_KHR_gl_texture_2D_image EGL_KHR_gl_renderbuffer_image EGL_KHR_get_all_proc_addresses EGL_ANGLE_flexible_surface_compatibility | 
| Direct rendering | Yes | 
| Reset notification strategy | 0x8252 | 
| GPU process crash count | 0 | 
 
 
Compositor Information 
 
| Tile Update Mode | One-copy | 
| Partial Raster | Disabled | 
 
 
GpuMemoryBuffers Status 
 
| ATC | Software only | 
| ATCIA | Software only | 
| DXT1 | Software only | 
| DXT5 | Software only | 
| ETC1 | Software only | 
| R_8 | Software only | 
| RGBA_4444 | Software only | 
| RGBX_8888 | Software only | 
| RGBA_8888 | Software only | 
| BGRX_8888 | Software only | 
| BGRA_8888 | Software only | 
| YUV_420 | Software only | 
| YUV_420_BIPLANAR | Software only | 
| UYVY_422 | Software only | 
 
 
Diagnostics 
 
| 0 
| b3DAccelerationEnabled | true | 
| b3DAccelerationExists | true | 
| bAGPEnabled | true | 
| bAGPExistenceValid | true | 
| bAGPExists | true | 
| bCanRenderWindow | true | 
| bDDAccelerationEnabled | true | 
| bDriverBeta | false | 
| bDriverDebug | false | 
| bDriverSigned | false | 
| bDriverSignedValid | false | 
| bNoHardware | false | 
| dwBpp | 32 | 
| dwDDIVersion | 11 | 
| dwHeight | 768 | 
| dwRefreshRate | 60 | 
| dwWHQLLevel | 0 | 
| dwWidth | 1366 | 
| iAdapter | 0 | 
| lDriverSize | 1479808 | 
| lMiniVddSize | 0 | 
| szAGPStatusEnglish | Enabled | 
| szAGPStatusLocalized | Enabled | 
| szChipType | AMD Radeon Graphics Processor (0x9802) | 
| szD3DStatusEnglish | Enabled | 
| szD3DStatusLocalized | Enabled | 
| szDACType | Internal DAC(400MHz) | 
| szDDIVersionEnglish | 11.1 | 
| szDDIVersionLocalized | 11.1 | 
| szDDStatusEnglish | Enabled | 
| szDDStatusLocalized | Enabled | 
| szDXVAHDEnglish | Not Supported | 
| szDXVAModes | ModeMPEG2_A ModeMPEG2_C | 
| szDescription | AMD Radeon HD 6310 Graphics | 
| szDeviceId | 0x9802 | 
| szDeviceIdentifier | {D7B71EE2-DB42-11CF-F076-8138BEC2C535} | 
| szDeviceName | \\.\DISPLAY1 | 
| szDisplayMemoryEnglish | 2037 MB | 
| szDisplayMemoryLocalized | 2037 MB | 
| szDisplayModeEnglish | 1366 x 768 (32 bit) (60Hz) | 
| szDisplayModeLocalized | 1366 x 768 (32 bit) (60Hz) | 
| szDriverAssemblyVersion | 15.201.1151.0 | 
| szDriverAttributes | Final Retail | 
| szDriverDateEnglish | 8/21/2015 12:00:00 AM | 
| szDriverDateLocalized | 8/21/2015 00:00:00 | 
| szDriverLanguageEnglish | English | 
| szDriverLanguageLocalized | English | 
| szDriverModelEnglish | WDDM 1.2 | 
| szDriverModelLocalized | WDDM 1.2 | 
| szDriverName | aticfx64.dll,aticfx64.dll,aticfx64.dll,aticfx32,aticfx32,aticfx32,atiumd64.dll,atidxx64.dll,atidxx64.dll,atiumdag,atidxx32,atidxx32,atiumdva,atiumd6a.cap,atitmm64.dll | 
| szDriverNodeStrongName | oem0.inf:cb0ae4141d512e0e:ati2mtag_Wrestler:15.201.1151.0:pci\ven_1002&dev_9802 | 
| szDriverSignDate | 
| 
| szDriverVersion | 8.17.0010.1404 | 
| szKeyDeviceID | Enum\PCI\VEN_1002&DEV_9802&SUBSYS_188B103C&REV_00 | 
| szKeyDeviceKey | \Registry\Machine\System\CurrentControlSet\Control\Video\{2832F83F-06F0-496F-95E4-4525B4A8174C}\0000 | 
| szManufacturer | Advanced Micro Devices, Inc. | 
| szMiniVdd | n/a | 
| szMiniVddDateEnglish | n/a | 
| szMiniVddDateLocalized | n/a | 
| szMonitorMaxRes | 
| 
| szMonitorName | Generic PnP Monitor | 
| szNotesEnglish | No problems found. | 
| szNotesLocalized | No problems found. | 
| szOverlayEnglish | Not Supported | 
| szRankOfInstalledDriver | 00DA2001 | 
| szRegHelpText | 
| 
| szRevision | 
| 
| szRevisionId | 0x0000 | 
| szSubSysId | 0x188B103C | 
| szTestResultD3D7English | Not run | 
| szTestResultD3D7Localized | Not run | 
| szTestResultD3D8English | Not run | 
| szTestResultD3D8Localized | Not run | 
| szTestResultD3D9English | Not run | 
| szTestResultD3D9Localized | Not run | 
| szTestResultDDEnglish | Not run | 
| szTestResultDDLocalized | Not run | 
| szVdd | n/a | 
| szVendorId | 0x1002 | 
 
| 
 
 
Log Messages 
 
- GpuProcessHostUIShim: The GPU process exited normally. Everything is okay.

Comment 9 by sarves...@gmail.com, May 27 2016

Below are the about:gpu details for one of the affected users on Windows 7


* Canvas: Hardware accelerated 
* Flash: Hardware accelerated 
* Flash Stage3D: Hardware accelerated 
* Flash Stage3D Baseline profile: Hardware accelerated 
* Compositing: Hardware accelerated 
* Multiple Raster Threads: Enabled 
* Rasterization: Software only. Hardware acceleration disabled 
* Video Decode: Software only, hardware acceleration unavailable 
* Video Encode: Hardware accelerated 
* WebGL: Hardware accelerated 
Driver Bug Workarounds 
* clear_uniforms_before_first_program_use 
* disable_d3d11 
* disable_discard_framebuffer 
* disable_msaa_on_non_webgl_contexts 
* exit_on_context_lost 
* force_cube_complete 
* scalarize_vec_and_mat_constructor_args 
* texsubimage_faster_than_teximage 
Problems Detected 
* Accelerated video decode interferes with GPU sandbox on older Intel drivers: 180695 < http://crbug.com/180695 > , 298968 < http://crbug.com/298968 > , 436968 < http://crbug.com/436968 > 
Disabled Features: accelerated_video_decode 
* Some drivers are unable to reset the D3D device in the GPU process sandbox 
Applied Workarounds: exit_on_context_lost 
* TexSubImage is faster for full uploads on ANGLE 
Applied Workarounds: texsubimage_faster_than_teximage 
* Clear uniforms before first program use on all platforms: 124764 < http://crbug.com/124764 > , 349137 < http://crbug.com/349137 > 
Applied Workarounds: clear_uniforms_before_first_program_use 
* Always rewrite vec/mat constructors to be consistent: 398694 < http://crbug.com/398694 > 
Applied Workarounds: scalarize_vec_and_mat_constructor_args 
* Old Intel drivers cannot reliably support D3D11: 363721 < http://crbug.com/363721 > 
Applied Workarounds: disable_d3d11 
* ANGLE crash on glReadPixels from incomplete cube map texture: 518889 <http://crbug.com/518889> 
Applied Workarounds: force_cube_complete 
* On Intel GPUs MSAA performance is not acceptable for GPU rasterization: 527565 < http://crbug.com/527565 > 
Applied Workarounds: disable_msaa_on_non_webgl_contexts 
* Framebuffer discarding can hurt performance on non-tilers: 570897 < http://crbug.com/570897 > 
Applied Workarounds: disable_discard_framebuffer 
* Accelerated rasterization has been disabled, either via about:flags or command line. 
Disabled Features: rasterization 
GpuMemoryBuffer Status 
ATC 
Software only 
ATCIA 
Software only 
DXT1 
Software only 
DXT5 
Software only 
ETC1 
Software only 
R_8 
Software only 
RGBA_4444 
Software only 
RGBX_8888 
Software only 
RGBA_8888 
Software only 
BGRX_8888 
Software only 
BGRA_8888 
Software only 
YUV_420 
Software only 
YUV_420_BIPLANAR 
Software only 
UYVY_422 
Software only 
Version Information 
Data exported 
5/27/2016, 6:53:33 AM 
Chrome version 
Chrome/49.0.2623.75 
Operating system 
Windows NT 6.1 SP1 
Software rendering list version 
10.17 
Driver bug list version 
8.45 
ANGLE commit id 
5fc4e56dfa3f 
2D graphics backend 
Skia 
Command Line Args 
Files (x86)\Google\Chrome\Application\chrome.exe" --flag-switches-begin --flag-switches-end 
Driver Information 
Initialization time 
72 
In-process GPU 
false 
Sandboxed 
false 
GPU0 
VENDOR = 0x8086, DEVICE= 0x0116 
Optimus 
true 
AMD switchable 
false 
Desktop compositing 
Aero Glass 
Driver vendor 
Intel Corporation 
Driver version 
8.15.10.2253 
Driver date 
11-28-2010 
Pixel shader version 
3.0 
Vertex shader version 
3.0 
Max. MSAA samples 
4 
Machine model name 
Machine model version 
GL_VENDOR 
Google Inc. 
GL_RENDERER 
ANGLE (Intel(R) HD Graphics Family Direct3D9Ex vs_3_0 ps_3_0) 
GL_VERSION 
OpenGL ES 2.0 (ANGLE 2.1.0.5fc4e56dfa3f) 
GL_EXTENSIONS 
GL_OES_element_index_uint GL_OES_packed_depth_stencil GL_OES_get_program_binary GL_OES_rgb8_rgba8 GL_EXT_texture_format_BGRA8888 GL_EXT_read_format_bgra GL_EXT_color_buffer_half_float GL_OES_texture_half_float GL_OES_texture_half_float_linear GL_OES_texture_float GL_OES_texture_float_linear GL_EXT_texture_compression_dxt1 GL_ANGLE_texture_compression_dxt3 GL_ANGLE_texture_compression_dxt5 GL_ANGLE_depth_texture GL_OES_depth32 GL_EXT_texture_storage GL_OES_texture_npot GL_EXT_texture_filter_anisotropic GL_EXT_occlusion_query_boolean GL_NV_fence GL_EXT_robustness GL_EXT_blend_minmax GL_ANGLE_framebuffer_blit GL_ANGLE_framebuffer_multisample GL_ANGLE_instanced_arrays GL_ANGLE_pack_reverse_row_order GL_OES_standard_derivatives GL_EXT_shader_texture_lod GL_EXT_frag_depth GL_ANGLE_texture_usage GL_ANGLE_translated_shader_source GL_EXT_debug_marker GL_OES_EGL_image GL_EXT_unpack_subimage GL_NV_pack_subimage GL_OES_vertex_array_object GL_KHR_debug 
Disabled Extensions 
Window system binding vendor 
Google Inc. (adapter LUID: 000000000000a3db) 
Window system binding version 
1.4 (ANGLE 2.1.0.5fc4e56dfa3f) 
Window system binding extensions 
EGL_EXT_create_context_robustness EGL_ANGLE_d3d_share_handle_client_buffer EGL_ANGLE_surface_d3d_texture_2d_share_handle EGL_ANGLE_query_surface_pointer EGL_ANGLE_window_fixed_size EGL_NV_post_sub_buffer EGL_KHR_create_context EGL_EXT_device_query EGL_KHR_image EGL_KHR_image_base EGL_KHR_gl_texture_2D_image EGL_KHR_gl_renderbuffer_image EGL_KHR_get_all_proc_addresses EGL_ANGLE_flexible_surface_compatibility 
Direct rendering 
Yes 
Reset notification strategy 
0x8252 
GPU process crash count 
0 
Diagnostics 
0 
b3DAccelerationEnabled 
true 
b3DAccelerationExists 
true 
bAGPEnabled 
true 
bAGPExistenceValid 
true 
bAGPExists 
true 
bCanRenderWindow 
true 
bDDAccelerationEnabled 
true 
bDriverBeta 
false 
bDriverDebug 
false 
bDriverSigned 
false 
bDriverSignedValid 
false 
bNoHardware 
false 
dwBpp 
32 
dwDDIVersion 
10 
dwHeight 
1050 
dwRefreshRate 
60 
dwWHQLLevel 
0 
dwWidth 
1680 
iAdapter 
0 
lDriverSize 
7428608 
lMiniVddSize 
0 
szAGPStatusEnglish 
Enabled 
szAGPStatusLocalized 
Enabled 
szChipType 
Intel(R) HD Graphics Family 
szD3DStatusEnglish 
Enabled 
szD3DStatusLocalized 
Enabled 
szDACType 
Internal 
szDDIVersionEnglish 
10.1 
szDDIVersionLocalized 
10.1 
szDDStatusEnglish 
Enabled 
szDDStatusLocalized 
Enabled 
szDXVAHDEnglish 
Supported 
szDXVAModes 
ModeMPEG2_A ModeMPEG2_C ModeWMV9_C ModeVC1_C 
szDescription 
Intel(R) HD Graphics Family 
szDeviceId 
0x0116 
szDeviceIdentifier 
{D7B78E66-4256-11CF-4666-B524B4C2C735} 
szDeviceName 
\\.\DISPLAY1 
szDisplayMemoryEnglish 
1696 MB 
szDisplayMemoryLocalized 
1696 MB 
szDisplayModeEnglish 
1680 x 1050 (32 bit) (60Hz) 
szDisplayModeLocalized 
1680 x 1050 (32 bit) (60Hz) 
szDriverAssemblyVersion 
8.15.10.2253 
szDriverAttributes 
Final Retail 
szDriverDateEnglish 
11/29/2010 06:23:16 
szDriverDateLocalized 
11/29/2010 6:23:16 AM 
szDriverLanguageEnglish 
English 
szDriverLanguageLocalized 
English 
szDriverModelEnglish 
WDDM 1.1 
szDriverModelLocalized 
WDDM 1.1 
szDriverName 
igdumd64.dll,igd10umd64.dll,igd10umd64.dll,igdumdx32,igd10umd32,igd10umd32 
szDriverNodeStrongName 
oem16.inf:Intel.Mfg.NTamd64:iSNBM0:8.15.10.2253:pci\ven_8086&dev_0116&subsys_04b81028 
szDriverSignDate 
szDriverVersion 
8.15.0010.2253 
szKeyDeviceID 
Enum\PCI\VEN_8086&DEV_0116&SUBSYS_04B81028&REV_09 
szKeyDeviceKey 
\Registry\Machine\System\CurrentControlSet\Control\Video\{2E5FF97C-1000-4A29-86F9-7E69DB2953A4}\0000 
szManufacturer 
Intel Corporation 
szMiniVdd 
n/a 
szMiniVddDateEnglish 
n/a 
szMiniVddDateLocalized 
n/a 
szMonitorMaxRes 
szMonitorName 
Generic PnP Monitor 
szNotesEnglish 
No problems found. 
szNotesLocalized 
No problems found. 
szOverlayEnglish 
Supported 
szRankOfInstalledDriver 
00E60001 
szRegHelpText 
szRevision 
szRevisionId 
0x0009 
szSubSysId 
0x04B81028 
szTestResultD3D7English 
Not run 
szTestResultD3D7Localized 
Not run 
szTestResultD3D8English 
Not run 
szTestResultD3D8Localized 
Not run 
szTestResultD3D9English 
Not run 
szTestResultD3D9Localized 
Not run 
szTestResultDDEnglish 
Not run 
szTestResultDDLocalized 
Not run 
szVdd 
n/a 
szVendorId 
0x8086 
Log Messages 
* GpuProcessHostUIShim: The GPU process exited normally. Everything is okay.

Comment 10 by ihf@chromium.org, May 28 2016

Owner: pucchakayala@chromium.org
The Win7 machine has a Sandybridge integrated GPU and an outdated driver.
The Win10 machine has a more recent integrated AMD GPU and more recent driver.
Nevertheless graphics acceleration is enabled for Flash.

Now all we need is a repro. Geetha, could you give this a try? Use a Windows machine with a slow (Intel) GPU, no NVIDIA.
I can actually reproduce a drop on browserstack.com (no hardware
acceleration) and locally using some devices with good GPUs (Nvidia)


I logged into Browser Stack, Use the  sample application at

http://files.jacksondunstan.com/articles/1797/Stage3DStateChanging.html


Log into Chrome v49 browser on Windows 7 and note down the FPS with the
boxes unchecked below - 4.8 FPS

Log into Chrome v49 browser on Windows 7 and note down the FPS with the
boxes checked below - 4.2 FPS


Now

Log into Chrome v50 browser on Windows 7 and note down the FPS with the
boxes unchecked below - 3.8 FPS

Log into Chrome v50 browser on Windows 7 and note down the FPS with the
boxes unchecked below - 3.2 FPS


Issue persists with v51 on BrowserStack

The drop is also there around 2 FPS with hardware acceleration (Nvidia)
when we test with some devices locally with good GPUs.

I would assume the drop would be substantially more with weaker GPUs
Going by our past experience in previous Chrome issue(
In https://code.google.com/p/chromium/issues/detail?id=496699 it was a flush command being inserted into a swapbuffer call.), 

we would "suspect" this to be a ppapi related commit amongst those below
https://chromium.googlesource.com/chromium/src/+log/49.0.2623.112..50.0.2661.75?pretty=fuller&n=10000

And knowing that this issue appears more in Windows maybe we could narrow down to Windows specific ppapi changes in the commits above (PPAPI commits are around 40-50, and platform specific would be fewer?)


Comment 13 by ihf@chromium.org, May 30 2016

Could you consider bisecting on a machine that reproduces the drop in fps? Scroll down or search for "Flash" for instructions (you have to find the path to a pepflashplayer.dll on your system):
https://www.chromium.org/developers/bisect-builds-py
If I go through he documentation at 
https://www.chromium.org/developers/bisect-builds-py it mentions

""You can also infer the revision number from the version number. The third set of numbers in a version number (e.g. the 835 in 14.0.835.202) is the branch number. You can look up http://src.chromium.org/viewvc/chrome/branches/<branch_number>/src/ and see find messages of the form "Branching for <branch_number> @<revision_number>".""

Now I go to Google Chrome 49.0.2623.112

Then according to the above 2623 should be the branch number, so I go to
http://src.chromium.org/viewvc/chrome/branches/2623/src
but it gives a 404 not found.

Same applies when I go to the bad version --> Google Chrome 50.0.2661.94
Then the branch should be 2661
which again breaks
http://src.chromium.org/viewvc/chrome/branches/2661/src

Can you give me the revision numbers for the below two versions of Chrome:-
Google Chrome 49.0.2623.112 (Good)
Google Chrome 50.0.2661.94 (Bad)



Can you confirm the below?

Branch 2623 —> Chrome revision 369907
Branch 2661 —> Chrome revision 378081

Comment 16 by ihf@chromium.org, Jun 1 2016

This is correct. Chrome M49 was branched at 369907 and M50 at 378081.
python bisect-builds.py -g 369907 -b 378081 -a win -f "C:\Program Files (x86)\Google\Chrome\...\PepperFlash\pepflashplayer.dll"
We have isolated the commit as between the following revisions

375908 - 24-25 FPS (First known good)
375918 - 22-23 FPS (First known bad)

CHANGELOG URL:
  https://chromium.googlesource.com/chromium/src/+log/20310f501d9498723bcf669b40
49d38bd38f5a9c..f0d89485a1584f87de303bd35a4bed7ef75ce777

We can use the following command to launch a narrow range of revisions and see the FPS difference on a Windows 7 machine

python bisect-builds.py -f "C:\Users\...\AppData\Local\Google\Chrome\User Data\PepperFlash\21.0.0.21
6\pepflashplayer.dll" -b 375924 -g 375903 -a win

Use the sample app
http://files.jacksondunstan.com/articles/1797/Stage3DStateChanging.html

Check all the checkboxes at the bottom left and also click on the 15*15*15 cubes, note the FPS before and after

The commit does have the following changes which are specific to Chrome on Windows (Direct3D changes)
https://chromium.googlesource.com/chromium/src/+/916870ee9a800d625dd2bcf31c226b0bb48fd262

pucchakayala@chromium.org, ihf@chromium.org - Is it possible for you to analyze and see for a fix for this drop?

Comment 18 by ihf@chromium.org, Jun 2 2016

Cc: pucchakayala@chromium.org
Components: Internals>GPU>ANGLE
Owner: senorblanco@chromium.org
Status: Assigned (was: Unconfirmed)
Assigning to Stephen to take a further look at the Angle roll:
https://chromium.googlesource.com/angle/angle.git/+log/8613f49..bce743a

Fixed link for bisection changelog:
https://chromium.googlesource.com/chromium/src/+log/20310f501d9498723bcf669b4049d38bd38f5a9c..f0d89485a1584f87de303bd35a4bed7ef75ce777


Owner: geoffl...@chromium.org
Geoff, could you take a look?
Will try bisecting ANGLE to see if there are any suspect CLs.

Otherwise, I think this may be due to us blacklisting some AMD drivers from using D3D11 for M50.  I found that a particular legacy driver (15.2XX) were very crashy.  I'll see if the blacklist can be narrowed.
Any updates on this?, also is there some tool through which we can bisect at a commit level (rather than a revision level) to isolate the commit?.

Also for your info the Driver I used above to reproduce the drop above was GL_RENDERER : NVIDIA NVS 5200M Direct3D11 vs_5_0 PS_5_0


I haven't had a chance to take a look yet, it will probably be a few days because I'm currently traveling.  A commit-level bisect will require building Chrome yourself, the bisect-builds.py tool uses builds from our continuous integration which doesn't do builds for every commit.

Thanks for pointing out the system using D3D11.

Comment 23 by msa...@zynga.com, Jun 11 2016

Hi Geoff, have you had a chance to take a look at this yet?  Thanks!
Status: Fixed (was: Assigned)
Bisected myself and discovered the patch causing the regression is in that roll.  Luckily, it was fixed a few commits later so this regression only happened for about 10 commits (https://chromium.googlesource.com/angle/angle.git/+log/618bebc..6253415f).  You can probably verify that performance is better with a recent chrome canary or dev build but it's not going to be possible to merge this fix into a M50 at this point.
Hi Geoff, Thanks for the update, has this fix gone out with Chrome 51.0.2704.84 or is it only visible in the dev version? We are actually seeing some degree of recovery in  Chrome 51 in overall stats, though some players are still having the slow gameplay issue, so am not sure it is result of this fix or some other performance improvement in Chrome 51?
It should be in M51.  It was unfortunate that chrome happened to branch for M50 in the brief window of this regression.
I'll see if there is anything obvious in more recent versions.

Could you let us know which drivers were Blacklisted so that we can try replicating with those?

Which file can I lookup to see in future as to which drivers become blacklisted?

Also can you confirm the fix went out with 51 and not with any minor version of 50?
This was fixed within M50 (50.2651) but may not have been pushed out until M51.

The blacklisting I mentioned was only from our D3D11 driver to D3D9, it is unlikely that caused the performance issues.

If you're interested, driver blackists are handled here: https://cs.chromium.org/chromium/src/gpu/config/gpu_driver_bug_list_json.cc .  The specific blacklist was for AMD drivers on the 15.200 and 15.201 version.
 Issue 630839  has been merged into this issue.

Sign in to add a comment