xorg7 and Intel 915GM problem [Update 7/12: power related?]

cbarnes913 · 2006-07-07 02:42:48

So I have a fresh installation of Arch and I have just installed the module xorg7 on it. Now I get the "I810: No matching device ID for instance (BusID PCI:0:2:1) found" warning that I used to get in the days of xorg6. I also don't get widescreen resolution like I should. I have installed the xf86-video-i810 package. I do see i810, i915, agpgart, drm and intel_agp in my lsmod. glxinfo reports "Direct Rendering: yes". glxgears gives me 876 FPS consistently.

Here are the important parts of xorg.conf:

...
Section "Module"
Load "bitmap"
Load "dbe"
Load "ddc"
Load "dri"
Load "extmod"
Load "freetype"
Load "glx"
Load "int10"
Load "record"
Load "type1"
Load "vbe"
Load "synaptics"
Load "xtrap"
EndSection
...
Section "Device"
Identifier "VGA Controller Intel 915GM"
Driver "i810"
BusID "PCI:0:2:0"
EndSection
Section "Device"
Identifier "Display Controller Intel 915GM"
Driver "i810"
BusID "PCI:0:2:1"
EndSection
Section "Monitor"
Identifier "Z71A LCD"
HorizSync 29-64
VertRefresh 0-60
EndSection
Section "Screen"
Identifier "Default Screen"
Device "VGA Controller Intel 915GM"
# Device "Display Controller Intel 915GM"
Monitor "Z71A LCD"
DefaultDepth 24
SubSection "Display"
Depth 24
Modes "1280x800"
EndSubSection
EndSection
Section "ServerLayout"
Identifier "Default Layout"
Screen "Default Screen"
InputDevice "Generic Keyboard" "CoreKeyboard"
InputDevice "Synaptics Touchpad" "AlwaysCore"
EndSection
Section "ServerFlags"
EndSection
Section "DRI"
Mode 0666
EndSection

Any ideas? I thought this thing was supposed to be fixed in 7.0. Am I going to have to cheat and use 855resolution?

Thanks,
CB

fk · 2006-07-07 09:59:29

The driver is to old you need a Version higer than 1.5, and a newer Version of the deiver is only supported by xorg 7.1

JGC · 2006-07-07 12:40:12

The fact that he has a working configuration means the driver works with his chipset (he even has DRI!)

The problem with the resolution is that the i810 driver uses the videobios to switch resolutions (the driver doesn't know how to set modes, it just abuses the BIOS for that). The windows driver can set modes from the driver, so many laptop manufacturers just ship a laptop with a broken VBIOS that doesn't contain the right resolutions (hey, it works in windows, so fuck you, we're not fixing it is the basic attitude from most manufacturers).

You should use 855resolution or 915resolution to get the full resolution. Results may vary on your laptop.

cbarnes913 · 2006-07-07 13:13:10

Well, I broke down last night after I posted and started using 855resolution in the interim until I figured out something better. I appreciate all the great information, JGC. Evidently, this is about as good as I can get? I didn't realize there was a 915resolution out there, so I went and got that since the author says there is a substantial difference in the way it's coded. It bumped my framerate in glxgears from 825fps (in 855resolution) to 828fps-- essentially no difference, but I'll take it. : )

I was a gentoo user for a while, and I ran into this brick wall while using xorg 6.8, but they had a hack to get that working with DRI. I still never got widescreen without hacking the BIOS, though. Is the BIOS hack going to hinder performance? Should I be satisfied with 828fps in glxgears on laptop graphics or should I be striving for more? Will the newer driver and xorg 7.1 fix the widescreen problem, and is that a viable solution for performance?

Thanks so much for the prompt replies.
CB

patroclo7 · 2006-07-07 13:24:36

I have your same card and I get about 1350 fps. Obviously, it depends ony from other factors, as the CPU (a 2 GHz centrino for me). I use 915 resolution (no impact on the performance, I think).

Anyway, it seems that a dev will put xorg7.1 in a specific repo in the next future.

cbarnes913 · 2006-07-07 14:53:38

I am on a laptop running at 1.73Ghz (verified by /proc/cpuinfo) and a gig of ram, but I am also running in 24 bit color. Are you running 16bit? We should be getting about the same. Have you tweaked it in any way?

CB

patroclo7 · 2006-07-07 21:26:12

No, I am running 24 as my default depth. The only interesting difference in my xorg.conf (I always write it by hand) is the VideoRam option in the Device Section. This is what 'man i810' says:
"By default 8 Megabytes of system memory are used for graphics. For the
830M and later, the default is 8 Megabytes when DRI is not enabled and
32 Megabytes with DRI is enabled. This amount may be changed with the
VideoRam entry in the config file Device section. It may be set to any
reasonable value up to 64MB for older chipsets or 128MB for newer
chipets. It is advisable to check the Xorg log file to check if any
features have been disabled because of insufficient video memory. In
particular, DRI support or tiling mode may be disabled with insuffi-
cient video memory. Either of these being disabled will reduce perfor-
mance for 3D applications. Note however, that increasing this value
too much will reduce the amount of system memory available for other
applications."

The following is my device section (I think you can ignore the other options:

Section "Device"
        ChipSet     "915GM"
        Identifier  "Intel Graphics Media Accelerator 900"
        Driver      "i810"
        VendorName  "Intel Corp."
        VideoRam    131072
        BusID       "PCI:00:02:0"
        #Option      "VBERestore" "Enable"
EndSection

I advice to read the full output of 'man i810'.

cbarnes913 · 2006-07-08 03:30:22

Well, I went through my xorg log and corrected some problems in my xorg.conf file, and I also changed the VideoRam in mine to what you have yours set at. The xorg log reflects the change (no memory problems, no auto-set to a different value). But, I get no change in performance. I still get in the 820s with glxgears.

Now that I have cleaned up my xorg.conf as per the output I got in the logs, I get very few warnings, however I do get a couple. I do get a warning for bad V_BIOS checksum, and I am assuming that is normal for someone using either the 855resolution or 915resolution BIOS hacks, and that is exactly what I should see, correct? I do still have the following error, which I now hold no hopes of getting rid of:

(WW) I810: No matching Device section for instance (BusID PCI:0:2:1) found

I have this in my lspci, which you no doubt also have:
00:02.0 VGA compatible controller: Intel Corporation Mobile 915GM/GMS/910GML Express Graphics Controller (rev 03)
00:02.1 Display controller: Intel Corporation Mobile 915GM/GMS/910GML Express Graphics Controller (rev 03)

Right now I specify my graphics as 0:2:0, hence the warning. Is there a way in xorg.conf to compensate for both items? Do you get that warning?

CB

patroclo7 · 2006-07-08 11:36:52

Yes, I get your same warnings. I agree about the v_bios stuff. Concerning the other warning, I tried different thinks to get rid of it, but without success (it always complains about one or the other entry).
I read somewhere that the new release in xorg 7.1 solves this second "problem" (not a real problem, I think: a harmless warning).

May be that the different cpu justifies the difference in performance. A last thing: look in the bios for the amount of ram for the video card.

patroclo7 · 2006-07-08 21:29:25

Another idea: do you use any framebuffer device? Intelfb or vesafb? I run intelfb with the noaccel option, otherwise there are occasional lockups when switching from X to console.

I do not know with this videocard, but it happens that the fb influences the dri performance

cbarnes913 · 2006-07-09 20:07:12

I'll give that intel framebuffer thing a look. I came across some more interesting information.

Xorg just flat out runs slugishly. I run gnome, and even the cruddy animations lag. I got xscreensaver installed and even the matrix screensaver was very laggy. Then I was running glxgears while I was screwing around with the screensavers stuff and noticed that if glxgears was running, the screensaver ran fine without hickups. We're talking between 25-35 fps on the GLmatrix screensaver with glxgears running as opposed to 2 fps without it running. Exiting glxgears makes performance go back to the way it was. I thought it might have been something that was getting loaded by glxgears that was causing the performance increase, but I also noticed that simply moving my mouse pointer over the screensaver helped increase the framerate. I can get up to 12 fps in GLmatrix while moving the mouse around as opposed to the 1 or 2 without. I am stumped.

In the meantime, I have taken some time experimenting with the modules loaded in xorg.conf. I started by disabling all of them, and then checking for errors in the log. As it happens, xorg must autoload what it detects, because many of the needed modules were loaded, despite the fact that they were not present in my xorg.conf file, including my synaptics touchpad module. The only things that didn't get autoloaded were the dri and glx modules and the dbe module, so I have paired it down to three from the 10 or 15 that were there. It doesn't make any difference in performance, but it cleans up the file. (And dbe does not affect performance in anyway I can notice. I did test that.)

CB

JGC · 2006-07-10 06:51:35

DBE is the double buffering extension, you only need that to run things like XGL, as it needs a double buffered desktop surface to do compositing.

cbarnes913 · 2006-07-10 18:44:29

Thanks for the info, JCG.

As it turns out, I am running vesafb for my console. I don't know if the intelfb is gonna help, but I will try it at some point.

I installed xorg 7.1 and the i810 1.6 to see if that made a difference. It didn't help frame rates in glxgears and I still have the problem where things only run smoothly when glxgears is running. I have no idea how to go about figuring that out. Installing 7.1 gave me some new output for glxgears and glxinfo. Here is my glxinfo:

libGL warning: 3D driver claims to not support visual 0x23
libGL warning: 3D driver claims to not support visual 0x25
libGL warning: 3D driver claims to not support visual 0x27
libGL warning: 3D driver claims to not support visual 0x29
libGL warning: 3D driver claims to not support visual 0x2b
libGL warning: 3D driver claims to not support visual 0x2d
libGL warning: 3D driver claims to not support visual 0x2f
libGL warning: 3D driver claims to not support visual 0x31
name of display: :0.0
display: :0  screen: 0
direct rendering: Yes
server glx vendor string: SGI
server glx version string: 1.2
server glx extensions:
    GLX_ARB_multisample, GLX_EXT_visual_info, GLX_EXT_visual_rating, 
    GLX_EXT_import_context, GLX_EXT_texture_from_pixmap, GLX_OML_swap_method, 
    GLX_SGI_make_current_read, GLX_SGIS_multisample, GLX_SGIX_hyperpipe, 
    GLX_SGIX_swap_barrier, GLX_SGIX_fbconfig
client glx vendor string: SGI
client glx version string: 1.4
client glx extensions:
    GLX_ARB_get_proc_address, GLX_ARB_multisample, GLX_EXT_import_context, 
    GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_MESA_allocate_memory, 
    GLX_MESA_copy_sub_buffer, GLX_MESA_swap_control, 
    GLX_MESA_swap_frame_usage, GLX_OML_swap_method, GLX_OML_sync_control, 
    GLX_SGI_make_current_read, GLX_SGI_swap_control, GLX_SGI_video_sync, 
    GLX_SGIS_multisample, GLX_SGIX_fbconfig, GLX_SGIX_pbuffer, 
    GLX_SGIX_visual_select_group, GLX_EXT_texture_from_pixmap
GLX version: 1.2
GLX extensions:
    GLX_ARB_get_proc_address, GLX_ARB_multisample, GLX_EXT_import_context, 
    GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_MESA_allocate_memory, 
    GLX_MESA_swap_control, GLX_MESA_swap_frame_usage, GLX_OML_swap_method, 
    GLX_SGI_make_current_read, GLX_SGI_video_sync, GLX_SGIS_multisample, 
    GLX_SGIX_fbconfig, GLX_SGIX_visual_select_group
OpenGL vendor string: Tungsten Graphics, Inc
OpenGL renderer string: Mesa DRI Intel(R) 915GM 20050225 x86/MMX/SSE2
OpenGL version string: 1.3 Mesa 6.5
OpenGL extensions:
    GL_ARB_depth_texture, GL_ARB_fragment_program, GL_ARB_imaging, 
    GL_ARB_multisample, GL_ARB_multitexture, GL_ARB_point_parameters, 
    GL_ARB_shadow, GL_ARB_texture_border_clamp, GL_ARB_texture_compression, 
    GL_ARB_texture_cube_map, GL_ARB_texture_env_add, 
    GL_ARB_texture_env_combine, GL_ARB_texture_env_dot3, 
    GL_ARB_texture_mirrored_repeat, GL_ARB_texture_rectangle, 
    GL_ARB_transpose_matrix, GL_ARB_vertex_buffer_object, 
    GL_ARB_vertex_program, GL_ARB_window_pos, GL_EXT_abgr, GL_EXT_bgra, 
    GL_EXT_blend_color, GL_EXT_blend_equation_separate, 
    GL_EXT_blend_func_separate, GL_EXT_blend_minmax, GL_EXT_blend_subtract, 
    GL_EXT_clip_volume_hint, GL_EXT_cull_vertex, GL_EXT_compiled_vertex_array, 
    GL_EXT_convolution, GL_EXT_copy_texture, GL_EXT_draw_range_elements, 
    GL_EXT_fog_coord, GL_EXT_histogram, GL_EXT_multi_draw_arrays, 
    GL_EXT_packed_pixels, GL_EXT_point_parameters, GL_EXT_polygon_offset, 
    GL_EXT_rescale_normal, GL_EXT_secondary_color, 
    GL_EXT_separate_specular_color, GL_EXT_shadow_funcs, GL_EXT_stencil_wrap, 
    GL_EXT_subtexture, GL_EXT_texture, GL_EXT_texture3D, 
    GL_EXT_texture_edge_clamp, GL_EXT_texture_env_add, 
    GL_EXT_texture_env_combine, GL_EXT_texture_env_dot3, 
    GL_EXT_texture_filter_anisotropic, GL_EXT_texture_lod_bias, 
    GL_EXT_texture_object, GL_EXT_texture_rectangle, GL_EXT_vertex_array, 
    GL_3DFX_texture_compression_FXT1, GL_APPLE_client_storage, 
    GL_APPLE_packed_pixels, GL_ATI_blend_equation_separate, 
    GL_IBM_rasterpos_clip, GL_IBM_texture_mirrored_repeat, 
    GL_INGR_blend_func_separate, GL_MESA_pack_invert, GL_MESA_ycbcr_texture, 
    GL_MESA_window_pos, GL_NV_blend_square, GL_NV_light_max_exponent, 
    GL_NV_texture_rectangle, GL_NV_texgen_reflection, GL_NV_vertex_program, 
    GL_NV_vertex_program1_1, GL_OES_read_format, GL_SGI_color_matrix, 
    GL_SGI_color_table, GL_SGIS_generate_mipmap, GL_SGIS_texture_border_clamp, 
    GL_SGIS_texture_edge_clamp, GL_SGIS_texture_lod, GL_SGIX_depth_texture, 
    GL_SUN_multi_draw_arrays
glu version: 1.3
glu extensions:
    GLU_EXT_nurbs_tessellator, GLU_EXT_object_space_tess

   visual  x  bf lv rg d st colorbuffer ax dp st accumbuffer  ms  cav
 id dep cl sp sz l  ci b ro  r  g  b  a bf th cl  r  g  b  a ns b eat
----------------------------------------------------------------------
0x23 24 tc  0 32  0 r  y  .  8  8  8  8  0  0  0  0  0  0  0  0 0 None
0x24 24 tc  0 32  0 r  .  .  8  8  8  8  0  0  0  0  0  0  0  0 0 None
0x25 24 tc  0 32  0 r  y  .  8  8  8  8  0 24  8  0  0  0  0  0 0 None
0x26 24 tc  0 32  0 r  .  .  8  8  8  8  0 24  8  0  0  0  0  0 0 None
0x27 24 tc  0 32  0 r  y  .  8  8  8  8  0  0  0 16 16 16 16  0 0 Slow
0x28 24 tc  0 32  0 r  .  .  8  8  8  8  0  0  0 16 16 16 16  0 0 Slow
0x29 24 tc  0 32  0 r  y  .  8  8  8  8  0 24  8 16 16 16 16  0 0 Slow
0x2a 24 tc  0 32  0 r  .  .  8  8  8  8  0 24  8 16 16 16 16  0 0 Slow
0x2b 24 dc  0 32  0 r  y  .  8  8  8  8  0  0  0  0  0  0  0  0 0 None
0x2c 24 dc  0 32  0 r  .  .  8  8  8  8  0  0  0  0  0  0  0  0 0 None
0x2d 24 dc  0 32  0 r  y  .  8  8  8  8  0 24  8  0  0  0  0  0 0 None
0x2e 24 dc  0 32  0 r  .  .  8  8  8  8  0 24  8  0  0  0  0  0 0 None
0x2f 24 dc  0 32  0 r  y  .  8  8  8  8  0  0  0 16 16 16 16  0 0 Slow
0x30 24 dc  0 32  0 r  .  .  8  8  8  8  0  0  0 16 16 16 16  0 0 Slow
0x31 24 dc  0 32  0 r  y  .  8  8  8  8  0 24  8 16 16 16 16  0 0 Slow
0x32 24 dc  0 32  0 r  .  .  8  8  8  8  0 24  8 16 16 16 16  0 0 Slow

Anyone have any ideas?

CB

patroclo7 · 2006-07-11 16:39:05

Well, my output (I have now installed xorg 7.1 too) is very similar (same warnings). The only difference:
there is another last line in the table:
0x4b 32 tc 0 32 o r . . 8 8 8 8 0 0 0 0 0 0 0 0 0 Ncon

The glxgears performance is not better than before about 1050 fps.

However, as you probably have read elsewhere, glxgears is not a reliable dri test The tests in globs (in the AUR) should be better: with them, the results are different with the new drivers (some of them better, some others worse...)

cbarnes913 · 2006-07-12 22:03:23

I have determined that the problem is caused by my power system somehow. When my laptop is unplugged, the system performs as it should (same slightly less fps in glxgears, but performance is smooth). If I run a screensaver without the laptop plugged in, I can watch the fps drop to almost nothing when the power is plugged in. However, though I have traced the problem, I don't know how to fix it. Any ideas?

CB

karsten · 2006-07-14 02:43:29

**EDIT nevermind i'm dumb, i was looking in the base repo thanks! where does one find this 915 resolution package? i couldn't find it in AUR anywhere. thanks

patroclo7 wrote:

I have your same card and I get about 1350 fps. Obviously, it depends ony from other factors, as the CPU (a 2 GHz centrino for me). I use 915 resolution (no impact on the performance, I think).
Anyway, it seems that a dev will put xorg7.1 in a specific repo in the next future.

Snowman · 2006-07-14 07:03:57

karsten wrote:

**EDIT nevermind i'm dumb, i was looking in the base repo thanks! where does one find this 915 resolution package? i couldn't find it in AUR anywhere.

The PKGBUILD is here: http://aur.archlinux.org/packages.php?d … =1&ID=4489
You'll need to build the package yourself.

patroclo7 · 2006-07-14 11:11:34

cbarnes913 wrote:

I have determined that the problem is caused by my power system somehow. When my laptop is unplugged, the system performs as it should (same slightly less fps in glxgears, but performance is smooth). If I run a screensaver without the laptop plugged in, I can watch the fps drop to almost nothing when the power is plugged in. However, though I have traced the problem, I don't know how to fix it. Any ideas?
CB

Well, and is it dri specific? Or the general performance of your laptop is worst when it is plugged in?
If it is specific, it is really awkward. If general, there could be three kinds of problems: acpi, bios, hardware. If you happen to run also a non-linux system, you should first of all check if analogous problems (also acceleration related) happen on these other Oss.

cbarnes913 · 2006-07-14 14:19:54

Alright, to be more specific:
The laptop performs worse when it is plugged in and this problem is not specific to X, but seems system wide. The problem is only noticible, however, during bootup (much longer boot when plugged in, 1:20 as opposed to 30 seconds when unplugged) and in X by observing framerates in applications other than glxgears. When I run glxgears, however, the problem seems to be temporarily resolved with other application framerates (like the xscreensaver performance).

I have run other distributions of Linux as well as Windows on this hardware without any issues like this. It seems to be Arch-specific.

Because I had not had this system installed for very long, I tested to see if this was just some sort of fluke problem by installing Arch fresh. I did a simple base install from the 0.7.2 CDROM packages. The slow bootup occurs before installing any other packages and continues to occur after a pacman -Syu.

I will post the outputs of bootchartd for both the powered and unpowered bootups later. Maybe this post should be moved since we have discovered it is not related to X (although the fact that glxgears fixes screensaver framerates is interesting).

CB

cbarnes913 · 2006-07-18 04:02:44

I copped out.

I was running Windows for a while before I decided to get back to Linux and installed Arch. I went with Arch to start out with because I wanted Linux fast. I love Arch and all, but Gentoo is my bread and butter, so I took the plunge and I have been compiling for days. BUT I have DRI and there appears to be no issues slowing down performance.

I still don't know what was causing the performance problem in Arch, and it appears that I am the only one who had the problem, so it looks like this thread will get burried.

Thanks for all the help.

CB out.

QuimaxW · 2007-01-03 06:12:33

I ran across this post looking for information on the i915. I'm having the same thing with the power unplugged, performance is better...

Any ideas?

Arch Linux

#1 2006-07-07 02:42:48

xorg7 and Intel 915GM problem [Update 7/12: power related?]

#2 2006-07-07 09:59:29

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#3 2006-07-07 12:40:12

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#4 2006-07-07 13:13:10

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#5 2006-07-07 13:24:36

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#6 2006-07-07 14:53:38

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#7 2006-07-07 21:26:12

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#8 2006-07-08 03:30:22

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#9 2006-07-08 11:36:52

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#10 2006-07-08 21:29:25

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#11 2006-07-09 20:07:12

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#12 2006-07-10 06:51:35

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#13 2006-07-10 18:44:29

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#14 2006-07-11 16:39:05

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#15 2006-07-12 22:03:23

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#16 2006-07-14 02:43:29

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#17 2006-07-14 07:03:57

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#18 2006-07-14 11:11:34

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#19 2006-07-14 14:19:54

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#20 2006-07-18 04:02:44

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

#21 2007-01-03 06:12:33

Re: xorg7 and Intel 915GM problem [Update 7/12: power related?]

Board footer