Graphics card driver issue
pro-grammer
I have noticed that some new PC's come with software rendering on. For
example, 3 weeks ago, I bought a DELL optiplex 960 (quad core), with
Windows 7. When I ran my opengl app,
glGetString(GL_VENDOR) gave me Microsoft Corporation
and
glGetString(GL_RENDERER) gave me GDI Generic
and the version was 1.1.0
The opengl performance was terrible!
However, when I updated the drivers on Intel's site, everything worked
great, and the GL_VENDOR became intel and the GL_RENDERER intel's Q45
Express chipset, and the version 2.1.0. My opengl animations were as
smooth as ever.
Curiously, Intel's update tool said that I already had a fairly
reecent intel driver installed. Apparently, for some reason, my system
was not using it.
I have 3 questions:
(a) In such a scenario, is a driver update the only way to get
hardware rendering on? Is there a way to turn it on programatically?
(b) Would this problem happen only with systems using Intel graphics
chips?
(c) When software rendering is on, is there a way to programatically
(in windows) get the name of my graphics chip(GL_VENDOR will give me
Microsoft, that wont help)... Or even better the link to the vendors
graphics update page... Please dont make me use some directx api...
Ruud van Gaal
> Hi,
>
> I have noticed that some new PC's come with software rendering on. For
> example, 3 weeks ago, I bought a DELL optiplex 960 (quad core), with
> Windows 7. When I ran my opengl app,
>
> glGetString(GL_VENDOR) gave me Microsoft Corporation
> and
> glGetString(GL_RENDERER) gave me GDI Generic
> and the version was 1.1.0
>
> The opengl performance was terrible!
Out of the box most Windows versions I've seen all provide basic drivers
that just get the window drawing working. For 3D work, you need to
install a 'real' driver from the gfx vendor.
That said, Intel doesn't really have decent gfxcards that I've seen; all
the current cards s*ck compared to ATI/nVidia in their implementation
(not too many extensions implemented).
jbwest
"Ruud van Gaal" <ru...@racer.nl> wrote in message
news:hg94tt$2el5$1...@adenine.netfront.net...
I'm sorry that you just got a nice quad-core with a crippled graphics
subsystem.
Go stuff a $40 (or better) newer nVidia card in your PCI-Express x16 slot
(if you have one).
Systems with nVidia almost always have at least an accelerated driver -- not
the latest
(from nVidia) but at least usable.
jbw
fungus
> Hi,
>
> I have noticed that some new PC's come with software rendering on. For
> example, 3 weeks ago, I bought a DELL optiplex 960 (quad core), with
> Windows 7. When I ran my opengl app,
>
> glGetString(GL_VENDOR) gave me Microsoft Corporation
> and
> glGetString(GL_RENDERER) gave me GDI Generic
> and the version was 1.1.0
>
> The opengl performance was terrible!
Microsoft doesn't like OpenGL and doesn't include
any accelerated drivers with Windows.
OTOH I thought DELL might have installed some...
--
<\___/>
/ O O \
\_____/ FTB.
http://www.topaz3d.com/ - New 3D editor for real time simulation
pro-grammer
hardware just to run my animations.
Intel cards may not be great, but theyre hardware rendering is many
times better than software rendering...
My problem is, how can I either:
(a) turn on hardware rendering programatically if the necessary
drivers are already available
or
(b) Find out the vendor of the graphics chip, so that I can direct the
user to the appropriate website to update his drivers(somehow the
update process turns on hardware rendering).
Am I in the right forum for this question?... If not please direct me
to the right one!
Charles E Hardwidge
news:672df7a3-9149-45af...@r1g2000vbn.googlegroups.com...
I'm only guessing but there's a possibility the supplied Intel drivers don't
support OpenGL out of the box. OpenGL should "just work" if it's there so I
wouldn't go through hoops.
The standard response I've seen since whenever is for an app to barf if it
doesn't detect an IHV OpenGL driver and for support to direct customers to
installing the latest driver.
Maybe it's just me but I'm of the view that Microsoft should be prosecuted
for competition violations over Direct3D and OpenGL. They market D3D as a
standalone API and platform control to lock out the competition.
--
Charles E Hardwidge
pro-grammer
If software rendering is on, my animations make adjustments so that
they look presentable... BUT, the first time the user runs my mfc app,
I want to let the user know that for a good user experience, you need
to update your driver, and _________ is the site that can help you do
it. I just need to know who manufactured the graphics chip, so that I
can offer the correct link to the user...
I can use the dxdiag utility to get this information. But thats messy,
and too slow.
Is there another way/API? I wonder how windows does it in its
DisplaySettings->Settings dialog box..
pro-grammer
guess this will have to do until there is a better way.... Thanks for
your help..
Greg Couch
for OpenGL than the ones that came with a 2007 laptop. I regularly
tell people to go to <http://downloadcenter.intel.com/> to get a new
driver to run my software. For instance, in the newer drivers you can
share display lists and other objects between contexts!
Philipp Klaus Krause
> That said, Intel doesn't really have decent gfxcards that I've seen; all
> the current cards s*ck compared to ATI/nVidia in their implementation
> (not too many extensions implemented).
>
Doesn't look so bad to me:
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_import_context,
GLX_EXT_texture_from_pixmap,
GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_MESA_copy_sub_buffer,
GLX_OML_swap_method, GLX_SGI_swap_control, GLX_SGIS_multisample,
GLX_SGIX_fbconfig, GLX_SGIX_visual_select_group
client glx vendor string: Mesa Project and 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_copy_sub_buffer,
GLX_MESA_swap_frame_usage, GLX_OML_swap_method, GLX_SGI_video_sync,
GLX_SGIS_multisample, GLX_SGIX_fbconfig, GLX_SGIX_visual_select_group,
GLX_EXT_texture_from_pixmap
OpenGL vendor string: Tungsten Graphics, Inc
OpenGL renderer string: Mesa DRI Intel(R) 965GM GEM 20090712 2009Q2 RC3
OpenGL version string: 2.1 Mesa 7.7-rc2
OpenGL shading language version string: 1.20
OpenGL extensions:
GL_ARB_copy_buffer, GL_ARB_depth_texture, GL_ARB_depth_clamp,
GL_ARB_draw_buffers, GL_ARB_draw_elements_base_vertex,
GL_ARB_fragment_program, GL_ARB_fragment_program_shadow,
GL_ARB_fragment_shader, GL_ARB_framebuffer_object,
GL_ARB_half_float_pixel, GL_ARB_map_buffer_range, GL_ARB_multisample,
GL_ARB_multitexture, GL_ARB_occlusion_query,
GL_ARB_pixel_buffer_object,
GL_ARB_point_parameters, GL_ARB_point_sprite, GL_ARB_provoking_vertex,
GL_ARB_seamless_cube_map, GL_ARB_shader_objects,
GL_ARB_shading_language_100, GL_ARB_shading_language_120,
GL_ARB_shadow,
GL_ARB_sync, 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_crossbar,
GL_ARB_texture_env_dot3, GL_ARB_texture_mirrored_repeat,
GL_ARB_texture_non_power_of_two, GL_ARB_texture_rectangle,
GL_ARB_transpose_matrix, GL_ARB_vertex_array_bgra,
GL_ARB_vertex_array_object, GL_ARB_vertex_buffer_object,
GL_ARB_vertex_program, GL_ARB_vertex_shader, 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_logic_op, GL_EXT_blend_minmax, GL_EXT_blend_subtract,
GL_EXT_cull_vertex, GL_EXT_compiled_vertex_array, GL_EXT_copy_texture,
GL_EXT_draw_range_elements, GL_EXT_framebuffer_blit,
GL_EXT_framebuffer_object, GL_EXT_fog_coord,
GL_EXT_gpu_program_parameters, GL_EXT_multi_draw_arrays,
GL_EXT_packed_depth_stencil, GL_EXT_packed_pixels,
GL_EXT_pixel_buffer_object, GL_EXT_point_parameters,
GL_EXT_polygon_offset, GL_EXT_provoking_vertex, GL_EXT_rescale_normal,
GL_EXT_secondary_color, GL_EXT_separate_specular_color,
GL_EXT_shadow_funcs, GL_EXT_stencil_two_side, GL_EXT_stencil_wrap,
GL_EXT_subtexture, GL_EXT_texture, GL_EXT_texture3D,
GL_EXT_texture_compression_s3tc, 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_texture_sRGB, GL_EXT_texture_swizzle, GL_EXT_vertex_array,
GL_EXT_vertex_array_bgra, GL_3DFX_texture_compression_FXT1,
GL_APPLE_client_storage, GL_APPLE_packed_pixels,
GL_APPLE_vertex_array_object, GL_ATI_blend_equation_separate,
GL_ATI_envmap_bumpmap, GL_ATI_texture_env_combine3,
GL_ATI_separate_stencil, GL_IBM_multimode_draw_arrays,
GL_IBM_rasterpos_clip, GL_IBM_texture_mirrored_repeat,
GL_INGR_blend_func_separate, GL_MESA_pack_invert,
GL_MESA_texture_signed_rgba, GL_MESA_ycbcr_texture, GL_MESA_window_pos,
GL_NV_blend_square, GL_NV_depth_clamp, GL_NV_light_max_exponent,
GL_NV_packed_depth_stencil, GL_NV_texture_env_combine4,
GL_NV_texture_rectangle, GL_NV_texgen_reflection, GL_NV_vertex_program,
GL_NV_vertex_program1_1, GL_OES_read_format, GL_SGIS_generate_mipmap,
GL_SGIS_texture_border_clamp, GL_SGIS_texture_edge_clamp,
GL_SGIS_texture_lod, GL_SUN_multi_draw_arrays, GL_S3_s3tc
pro-grammer
Well I cant say you're wrong... I was pretty impressed with the
performance on my new dell optiplex.... Unfortunately we got to
support 2007 laptops as well... And a quad core dell optiplex, with a
great intel graphics chip, doesnt ship with drivers that support
opengl!!!! Thats frustrating... I still dont understand how and why
this can happen... Dell ought to know better..
Anyways, I plan to do the same thing you do... tell my customer to
update his driver from the appropriate web site... I wish there was an
easier automatic way of updating the drivers, instead of making the
possibly naive user go to a web site and do it manually....
Leclerc
> performance on my new dell optiplex.... Unfortunately we got to
> support 2007 laptops as well... And a quad core dell optiplex, with a
> great intel graphics chip, doesnt ship with drivers that support
> opengl!!!! Thats frustrating... I still dont understand how and why
> this can happen... Dell ought to know better..
(I also have to support laptops with older intel graphics card)
We experienced stupid problem that laptop vendors didn't upgrade video
drivers, and original intel's drivers refused to install over
proprietary drivers. The sollution was to deinstall proprietary video
driver from laptop, reboot (using plain vanilla VGA driver), and than
install new intel drivers. Needles to say, this is easy procedure for
any sysop, but we couldn't expect ordinary user to follow this procedure.
cheers,
Gordan
fungus
> Guys, I agree with y'all. BUT, I cant make my customers change their
> hardware just to run my animations.
>
> Intel cards may not be great, but theyre hardware rendering is many
> times better than software rendering...
>
Yep. They're not much good for games but pretty
good for everything else. They're also the market
leader (there's more Intel GPUs out there than
ATI and NVIDIA combined) so the "ignore Intel
they're crap" attitude isn't helpful.
> My problem is, how can I either:
> (a) turn on hardware rendering programatically if the necessary
> drivers are already available
It's automatic, there's nothing you can do.
fungus
>
> We experienced stupid problem that laptop vendors didn't upgrade video
> drivers, and original intel's drivers refused to install over
> proprietary drivers.
Yes, this is a real problem. Many many laptops
are stuck with the supplied driver unless the
user is good at hacking.
OTOH, Intel Drivers *should* support OpenGL
(unless something changed in Windows 7)
Charles E Hardwidge
> Well, im not sure if I want to "barf it"..
>
> If software rendering is on, my animations make adjustments so that
> they look presentable... BUT, the first time the user runs my mfc app,
> I want to let the user know that for a good user experience, you need
> to update your driver, and _________ is the site that can help you do
> it. I just need to know who manufactured the graphics chip, so that I
> can offer the correct link to the user...
The games industry solved this issue years ago. Most companies typically
insist you upgrade to the latest driver and provide a readme along those
lines with the most obvious questions and solutions. Other products (like
3DS Max) default to software and give you a whole bunch of handwaving if you
switch rendering to OpenGL.
This sort of issue was so bad at one point companies insisted that IE5 was
installed. In fact, some games wouldn't run unless it was. The reason was
IE5 bundled a whole bunch of updates with it that provided a single support
base much like, say, demanding Windows XP SP3.
If you're really paranoid you'll supply a directory on an install disc or
online repository option for drivers. But, that depends on the distribution
license on a per driver basis and whether you want to subsidise the IHV's
bandwidth. If it's obscure stuff you might want a backup just in case.
--
Charles E Hardwidge
Ruud van Gaal
> Ruud van Gaal schrieb:
>
>> That said, Intel doesn't really have decent gfxcards that I've seen; all
>> the current cards s*ck compared to ATI/nVidia in their implementation
>> (not too many extensions implemented).
>>
>
> Doesn't look so bad to me:
>
> 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_import_context,
Indeed, that seems quite a few. Perhaps the Intel state is catching up,
that's good news. :) I've done a game for Intel laptops recently, and we
got around 20-30x less performance on those cards, and also I had to
turn off Cg for example. Quite a big setback.
Ruud
Vladimir Jovic
> Philipp Klaus Krause wrote:
>> Ruud van Gaal schrieb:
>>
>>> That said, Intel doesn't really have decent gfxcards that I've seen; all
>>> the current cards s*ck compared to ATI/nVidia in their implementation
>>> (not too many extensions implemented).
>>>
>>
>> Doesn't look so bad to me:
>>
>> 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_import_context,
> ...
>
> Indeed, that seems quite a few. Perhaps the Intel state is catching up,
> that's good news. :) I've done a game for Intel laptops recently, and we
What is catching up - graphical chips or drivers? Because both are much
worse then nvidia's (not sure how the ATI compare to nvidia lately).
> got around 20-30x less performance on those cards, and also I had to
> turn off Cg for example. Quite a big setback.
Why did you have to turn off Cg for those cards? I am using an intel
chip and Cg works. It has some quirks, but I found a workaround for most
problems.
Philipp Klaus Krause
> Ruud van Gaal wrote:
>> Philipp Klaus Krause wrote:
>>> Ruud van Gaal schrieb:
>>>
>>>> That said, Intel doesn't really have decent gfxcards that I've seen;
>>>> all
>>>> the current cards s*ck compared to ATI/nVidia in their implementation
>>>> (not too many extensions implemented).
>>>>
>>>
>>> Doesn't look so bad to me:
>>>
>>> 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_import_context,
>> ...
>>
>> Indeed, that seems quite a few. Perhaps the Intel state is catching
>> up, that's good news. :) I've done a game for Intel laptops recently,
>> and we
>
> What is catching up - graphical chips or drivers? Because both are much
> worse then nvidia's (not sure how the ATI compare to nvidia lately).
I don't know about Windows or MacOS, but as far as I can see drivers for
Intel chips seem to be the most stable, bug-free and feature-complete
(in terms of exposing the features the hardware can support), better
than ATI or Nvidia drivers at the moment.
Philipp
jbwest
"Philipp Klaus Krause" <p...@spth.de> wrote in message
news:4b3093d7$0$9752$6e1e...@read.cnntp.org...
Are you speaking of Linux drivers for nVidia? Do you mean open-source or
vendor-supplied?
The nvidia-supplied nVidia driver is by far the most comprehensive &
bug-free implementation of OpenGL:.
ATI Linux drivers have been known in the past to have problems, although
I've heard that they are better
now.
jbw
Philipp Klaus Krause
>
> Are you speaking of Linux drivers for nVidia? Do you mean open-source or
> vendor-supplied?
> The nvidia-supplied nVidia driver is by far the most comprehensive &
> bug-free implementation of OpenGL:.
> ATI Linux drivers have been known in the past to have problems, although
> I've heard that they are better
> now.
The free drivers for Nvidia still have a very long way to go. The
non-free ones are OK quality-wise (and have lots of features since the
hardware supports them). Nevetheless, IMO the free Intel drivers are at
least equal to them in quality (although limited by the capabilities of
the hardware, which are in turn limited by economic considerations).
ATI's drivers are getting better (both free and non-free), but they
(both) are still significantly buggier than free Intel and non-free
Nvidia drivers.
Philipp
jbwest
Well, it is easier to have stable code when the hardware is limited in
capability and types supported...and you don't even try to support lots of
features...
Yes, the open source NV drivers aren't very good. I have never found any
bugs with recent proprietary Quadro drivers on Linux, and I've seen some
excellent performance characteristics. My experience is with much larger
datasets than one could ever conceivably try on Intel.
NB: The nVidia corporate drivers are completely "free" -- but proprietary
and not open-source.
jbw
Ruud van Gaal
> Ruud van Gaal wrote:
>> Philipp Klaus Krause wrote:
>>> Ruud van Gaal schrieb:
>> up, that's good news. :) I've done a game for Intel laptops recently,
>> and we
>
> What is catching up - graphical chips or drivers? Because both are much
> worse then nvidia's (not sure how the ATI compare to nvidia lately).
Drivers perhaps, although I don't assume too many high-end features to
be present on Intel hardware.
>> got around 20-30x less performance on those cards, and also I had to
>> turn off Cg for example. Quite a big setback.
>
> Why did you have to turn off Cg for those cards? I am using an intel
> chip and Cg works. It has some quirks, but I found a workaround for most
> problems.
The got errors when trying to do Cg, and updating drivers was not a
possibility. Also, the project had to be done in a few weeks, so better
to shift towards the safe side.
Ruud
--- news://freenews.netfront.net/ - complaints: ne...@netfront.net ---