looks like Catalyst 9.2beta leaked.

[energyman]

http://drivers.mydrivers.com/download/250-100232-AMD-Radeon-9.2-For-Linux/

I am contemplating to try it ....

edit found here with second link:
http://forums.guru3d.com/showthread.php?t=283978

http://www.ati-forum.de/files/Downloads/Driver/Linux_catalyst_9.2_CES_09_preview_driver.zip

[d2kx]

Lol those chinese people are insane. Great find :)

[energyman]

the people at guru3d found it - and I found them. Btw, the second download link points to a broken package.

and now:

glxinfo
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_OML_swap_method,
GLX_SGIS_multisample, GLX_SGIX_fbconfig, GLX_SGIX_visual_select_group
client glx vendor string: SGI
client glx version string: 1.4
client glx extensions:
GLX_ARB_create_context, 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_NV_swap_group, 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_swap_barrier, GLX_SGIX_swap_group,
GLX_SGIX_visual_select_group, GLX_EXT_texture_from_pixmap
GLX version: 1.2
GLX extensions:
GLX_ARB_create_context, GLX_ARB_get_proc_address, GLX_ARB_multisample,
GLX_EXT_import_context, GLX_EXT_visual_info, GLX_EXT_visual_rating,
GLX_MESA_swap_control, GLX_NV_swap_group, GLX_OML_swap_method,
GLX_SGI_video_sync, GLX_SGIS_multisample, GLX_SGIX_fbconfig,
GLX_SGIX_swap_barrier, GLX_SGIX_swap_group, GLX_SGIX_visual_select_group
OpenGL vendor string: ATI Technologies Inc.
OpenGL renderer string: ATI Radeon HD 3870
OpenGL version string: 2.1.8472 Release
OpenGL shading language version string: 1.20
OpenGL extensions:
GL_AMDX_vertex_shader_tessellator, GL_AMD_performance_monitor,
GL_AMD_texture_texture4, GL_ARB_color_buffer_float,
GL_ARB_depth_buffer_float, GL_ARB_depth_texture, GL_ARB_draw_buffers,
GL_ARB_draw_instanced, GL_ARB_fragment_program,
GL_ARB_fragment_program_shadow, GL_ARB_fragment_shader,
GL_ARB_framebuffer_object, GL_ARB_framebuffer_sRGB,
GL_ARB_half_float_pixel, GL_ARB_half_float_vertex,
GL_ARB_instanced_arrays, 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_shader_objects,
GL_ARB_shader_texture_lod, GL_ARB_shading_language_100, GL_ARB_shadow,
GL_ARB_shadow_ambient, GL_ARB_texture_border_clamp,
GL_ARB_texture_compression, GL_ARB_texture_compression_rgtc,
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_float,
GL_ARB_texture_mirrored_repeat, GL_ARB_texture_non_power_of_two,
GL_ARB_texture_rectangle, GL_ARB_texture_rg, GL_ARB_transpose_matrix,
GL_ARB_vertex_array_object, GL_ARB_vertex_buffer_object,
GL_ARB_vertex_program, GL_ARB_vertex_shader, GL_ARB_window_pos,
GL_ATI_draw_buffers, GL_ATI_envmap_bumpmap, GL_ATI_fragment_shader,
GL_ATI_meminfo, GL_ATI_separate_stencil, GL_ATI_texture_compression_3dc,
GL_ATI_texture_env_combine3, GL_ATI_texture_float, GL_EXT_abgr,
GL_EXT_bgra, GL_EXT_bindable_uniform, GL_EXT_blend_color,
GL_EXT_blend_equation_separate, GL_EXT_blend_func_separate,
GL_EXT_blend_minmax, GL_EXT_blend_subtract, GL_EXT_compiled_vertex_array,
GL_EXT_copy_texture, GL_EXT_draw_buffers2, GL_EXT_draw_range_elements,
GL_EXT_fog_coord, GL_EXT_framebuffer_blit, GL_EXT_framebuffer_multisample,
GL_EXT_framebuffer_object, GL_EXT_framebuffer_sRGB,
GL_EXT_gpu_program_parameters, GL_EXT_gpu_shader4,
GL_EXT_multi_draw_arrays, GL_EXT_packed_depth_stencil,
GL_EXT_packed_float, GL_EXT_packed_pixels, GL_EXT_point_parameters,
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_texgen_reflection, GL_EXT_texture3D,
GL_EXT_texture_array, GL_EXT_texture_compression_latc,
GL_EXT_texture_compression_rgtc, GL_EXT_texture_compression_s3tc,
GL_EXT_texture_cube_map, 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_integer, GL_EXT_texture_lod_bias,
GL_EXT_texture_mirror_clamp, GL_EXT_texture_object,
GL_EXT_texture_rectangle, GL_EXT_texture_sRGB,
GL_EXT_texture_shared_exponent, GL_EXT_transform_feedback,
GL_EXT_vertex_array, GL_KTX_buffer_region, GL_NV_blend_square,
GL_NV_conditional_render, GL_NV_copy_depth_to_color,
GL_NV_texgen_reflection, GL_SGIS_generate_mipmap,
GL_SGIS_texture_edge_clamp, GL_SGIS_texture_lod, GL_WIN_swap_hint,
WGL_EXT_swap_control

[ 72.272894] fglrx: module license 'Proprietary. (C) 2002 - ATI Technologies, Starnberg, GERMANY' taints kernel.
[ 72.290446] [fglrx] Maximum main memory to use for locked dma buffers: 3787 MBytes.
[ 72.290671] [fglrx] vendor: 1002 device: 9501 count: 1
[ 72.291049] [fglrx] ioport: bar 4, base 0xc000, size: 0x100
[ 72.291060] pci 0000:02:00.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
[ 72.291064] pci 0000:02:00.0: setting latency timer to 64
[ 72.291246] [fglrx] Kernel PAT support detected, disabling driver built-in PAT support
[ 72.291264] [fglrx] module loaded - fglrx 8.58.2 [Dec 30 2008] with 1 minors
[ 73.000842] Clocksource tsc unstable (delta = -200176421 ns)
[ 76.261016] [fglrx] Firegl kernel thread PID: 3795
[ 76.282255] [fglrx] Gart cacheable size:1970 M.
[ 76.282261] [fglrx] Reserved FB block: Shared offset:0, size:1000000
[ 76.282263] [fglrx] Reserved FB block: Unshared offset:fc1b000, size:3e5000
[ 76.282264] [fglrx] Reserved FB block: Unshared offset:1fffc000, size:4000

composite works ok, no flickering with xv, but the picture bleeds into the decoration and is stretched.

Scrolling is fast, window movement is instantious (aka, no lag), manual resizing is still slow.

[energyman]

so, tried ut2004 and vegastrike. 3d performance is great. Seems to be better (but I may be mistaken). In vegastrike there is a 'colour' problem. The Atlantis planet in Cephid17 isreddish - while the planet was alwas blue in the past. Also with composite turned on the colours are way too dark of the ship. Strange, but not a problem.

[energyman]

hm, strange problem:
aticonfig --odgt
ERROR - Could not find library: libatiadlxx.so

but:
ls -lh /usr/lib64/libatiadlxx.so
-rwxr-xr-x 1 root root 147K 9. Jan 15:37 /usr/lib64/libatiadlxx.so

[conholster]

2 months in a row...might it be we will see public betas in the future?

[Linuxhippy]

has anybody run gtkperf or jxrendermark on the 9.2 (or 9.1) drivers with composite disabled?

I would be really interested in 2D hw accaleration, and if its still as bad as it used to be in previous releases.
Unfourtunatly I am currently not at home, so can't test my Radeon3850 :-/

Thank you in advance, Clemens

[Kano]

It would be more easy then all betas are announced in a similar way that nvidia does. Most easy would be a standard position + latest.txt file which would point to latest stable + beta. I see no real reason why those have not been published before. This would decrease the frustration at launch day of the next official stable driver - when you know what you can expect.

[grantek]

Wow, is 9.1 final even out?

[energyman]

nope ;) 8.12 is the latest release

[energyman]

jxrendermark, 9.2pre composite turned on:
/JXRenderMark-0.6
482323.553055 Ops/s; rects (!); 15x15
135638.297872 Ops/s; rects (!); 75x75
31211.538462 Ops/s; rects (!); 250x250
123327.479339 Ops/s; rects composition (!); 15x15
36528.497409 Ops/s; rects composition (!); 75x75
5268.120805 Ops/s; rects composition (!); 250x250
170763.750000 Ops/s; put composition (!); 15x15
24544.763514 Ops/s; put composition (!); 75x75
3065.954774 Ops/s; put composition (!); 250x250
93616.111952 Ops/s; mixed composition (!); 15x15
18726.923077 Ops/s; mixed composition (!); 75x75
2594.178082 Ops/s; mixed composition (!); 250x250
156801.136364 Ops/s; lines composition (!); 15x15
23537.564767 Ops/s; lines composition (!); 75x75
2568.897638 Ops/s; lines composition (!); 250x250
340029.296875 Ops/s; Simple Blit (!); 15x15
107912.707838 Ops/s; Simple Blit (!); 75x75
15428.440926 Ops/s; Simple Blit (!); 250x250
139749.658936 Ops/s; Transformed Blit Linear (!); 15x15
13771.484375 Ops/s; Transformed Blit Linear (!); 75x75
1275.000000 Ops/s; Transformed Blit Linear (!); 250x250
1691.815857 Ops/s; Transformed Blit Billinear sharp edges (!); 15x15
69.982394 Ops/s; Transformed Blit Billinear sharp edges (!); 75x75
31.830239 Ops/s; Transformed Blit Billinear sharp edges (!); 250x250
34019.417476 Ops/s; transformed texture paint (!); 15x15
1756.377551 Ops/s; transformed texture paint (!); 75x75
154.137931 Ops/s; transformed texture paint (!); 250x250
63631.266491 Ops/s; gradient+temp texture(!); 15x15
8494.058501 Ops/s; gradient+temp texture(!); 75x75
813.829787 Ops/s; gradient+temp texture(!); 250x250
65578.378378 Ops/s; 12pt text grayscale; 5x5
22457.547170 Ops/s; 12pt text grayscale; 15x15
65002.000000 Ops/s; 12pt text lcd; 5x5
22643.324607 Ops/s; 12pt text lcd; 15x15

composite off:
474491.228070 Ops/s; rects (!); 15x15
127869.064748 Ops/s; rects (!); 75x75
31360.150376 Ops/s; rects (!); 250x250
144641.176471 Ops/s; rects composition (!); 15x15
35478.179551 Ops/s; rects composition (!); 75x75
5417.621777 Ops/s; rects composition (!); 250x250
166658.188586 Ops/s; put composition (!); 15x15
21911.943320 Ops/s; put composition (!); 75x75
3054.511278 Ops/s; put composition (!); 250x250
80539.370079 Ops/s; mixed composition (!); 15x15
17467.318059 Ops/s; mixed composition (!); 75x75
2265.686275 Ops/s; mixed composition (!); 250x250
159600.000000 Ops/s; lines composition (!); 15x15
22020.776567 Ops/s; lines composition (!); 75x75
2503.125000 Ops/s; lines composition (!); 250x250
341328.750000 Ops/s; Simple Blit (!); 15x15
114164.062500 Ops/s; Simple Blit (!); 75x75
17973.750000 Ops/s; Simple Blit (!); 250x250
129646.706587 Ops/s; Transformed Blit Linear (!); 15x15
13471.945137 Ops/s; Transformed Blit Linear (!); 75x75
1211.081794 Ops/s; Transformed Blit Linear (!); 250x250
1653.750000 Ops/s; Transformed Blit Billinear sharp edges (!); 15x15
65.428110 Ops/s; Transformed Blit Billinear sharp edges (!); 75x75
30.848329 Ops/s; Transformed Blit Billinear sharp edges (!); 250x250
31164.148352 Ops/s; transformed texture paint (!); 15x15
1718.827930 Ops/s; transformed texture paint (!); 75x75
158.718331 Ops/s; transformed texture paint (!); 250x250
60126.068376 Ops/s; gradient+temp texture(!); 15x15
7665.413534 Ops/s; gradient+temp texture(!); 75x75
741.071429 Ops/s; gradient+temp texture(!); 250x250
61177.480916 Ops/s; 12pt text grayscale; 5x5
21150.375940 Ops/s; 12pt text grayscale; 15x15
58362.328767 Ops/s; 12pt text lcd; 5x5
21347.014925 Ops/s; 12pt text lcd; 15x15

[energyman]

composite on 100rounds:
GtkPerf 0.40 - Starting testing: Fri Jan 9 22:58:53 2009

GtkEntry - time: 0,01
GtkComboBox - time: 2,90
GtkComboBoxEntry - time: 2,60
GtkSpinButton - time: 0,08
GtkProgressBar - time: 0,05
GtkToggleButton - time: 0,08
GtkCheckButton - time: 0,07
GtkRadioButton - time: 0,08
GtkTextView - Add text - time: 0,21
GtkTextView - Scroll - time: 0,11
GtkDrawingArea - Lines - time: 3,13
GtkDrawingArea - Circles - time: 3,82
GtkDrawingArea - Text - time: 1,53
GtkDrawingArea - Pixbufs - time: 0,03
---
Total time: 14,70


composite off, 100 rounds:
GtkPerf 0.40 - Starting testing: Fri Jan 9 22:57:44 2009

GtkEntry - time: 0,01
GtkComboBox - time: 1,00
GtkComboBoxEntry - time: 1,16
GtkSpinButton - time: 0,04
GtkProgressBar - time: 0,02
GtkToggleButton - time: 0,06
GtkCheckButton - time: 0,03
GtkRadioButton - time: 0,08
GtkTextView - Add text - time: 0,17
GtkTextView - Scroll - time: 0,10
GtkDrawingArea - Lines - time: 0,26
GtkDrawingArea - Circles - time: 0,45
GtkDrawingArea - Text - time: 6,84
GtkDrawingArea - Pixbufs - time: 0,06
---
Total time: 10,29

[energyman]

8.12, composite on:
GtkPerf 0.40 - Starting testing: Fri Jan 9 23:13:46 2009

GtkEntry - time: 0,02
GtkComboBox - time: 3,12
GtkComboBoxEntry - time: 3,11
GtkSpinButton - time: 0,09
GtkProgressBar - time: 0,04
GtkToggleButton - time: 0,09
GtkCheckButton - time: 0,06
GtkRadioButton - time: 0,14
GtkTextView - Add text - time: 0,23
GtkTextView - Scroll - time: 0,16
GtkDrawingArea - Lines - time: 3,03
GtkDrawingArea - Circles - time: 3,63
GtkDrawingArea - Text - time: 1,42
GtkDrawingArea - Pixbufs - time: 0,02
---
Total time: 15,15

composite off:
GtkEntry - time: 0,02
GtkComboBox - time: 0,99
GtkComboBoxEntry - time: 1,15
GtkSpinButton - time: 0,04
GtkProgressBar - time: 0,02
GtkToggleButton - time: 0,06
GtkCheckButton - time: 0,03
GtkRadioButton - time: 0,09
GtkTextView - Add text - time: 0,17
GtkTextView - Scroll - time: 0,11
GtkDrawingArea - Lines - time: 0,27
GtkDrawingArea - Circles - time: 3,78
GtkDrawingArea - Text - time: 7,04
GtkDrawingArea - Pixbufs - time: 0,05
---
Total time: 13,83

[energyman]

8.12, jxrendermark, composite on:
/JXRenderMark-0.6
439612.548512 Ops/s; rects (!); 15x15
150031.390135 Ops/s; rects (!); 75x75
31613.281250 Ops/s; rects (!); 250x250
133231.671554 Ops/s; rects composition (!); 15x15
37636.904762 Ops/s; rects composition (!); 75x75
5937.290970 Ops/s; rects composition (!); 250x250
178172.456576 Ops/s; put composition (!); 15x15
26022.769517 Ops/s; put composition (!); 75x75
3058.564232 Ops/s; put composition (!); 250x250
98818.181818 Ops/s; mixed composition (!); 15x15
22010.479042 Ops/s; mixed composition (!); 75x75
2555.555556 Ops/s; mixed composition (!); 250x250
161325.738916 Ops/s; lines composition (!); 15x15
26366.602687 Ops/s; lines composition (!); 75x75
2595.238095 Ops/s; lines composition (!); 250x250
350250.000000 Ops/s; Simple Blit (!); 15x15
120788.860104 Ops/s; Simple Blit (!); 75x75
16873.849693 Ops/s; Simple Blit (!); 250x250
146709.677419 Ops/s; Transformed Blit Linear (!); 15x15
12956.275033 Ops/s; Transformed Blit Linear (!); 75x75
1343.198992 Ops/s; Transformed Blit Linear (!); 250x250
1547.288777 Ops/s; Transformed Blit Billinear sharp edges (!); 15x15
58.912387 Ops/s; Transformed Blit Billinear sharp edges (!); 75x75
31.346749 Ops/s; Transformed Blit Billinear sharp edges (!); 250x250
30307.453416 Ops/s; transformed texture paint (!); 15x15
1750.628141 Ops/s; transformed texture paint (!); 75x75
173.586572 Ops/s; transformed texture paint (!); 250x250
64899.025070 Ops/s; gradient+temp texture(!); 15x15
6824.362606 Ops/s; gradient+temp texture(!); 75x75
769.067797 Ops/s; gradient+temp texture(!); 250x250
69111.657303 Ops/s; 12pt text grayscale; 5x5
20342.024540 Ops/s; 12pt text grayscale; 15x15
63595.394737 Ops/s; 12pt text lcd; 5x5
20033.093525 Ops/s; 12pt text lcd; 15x15

composite off:
434832.496863 Ops/s; rects (!); 15x15
122139.105058 Ops/s; rects (!); 75x75
31339.772727 Ops/s; rects (!); 250x250
121023.248408 Ops/s; rects composition (!); 15x15
35994.389027 Ops/s; rects composition (!); 75x75
5318.181818 Ops/s; rects composition (!); 250x250
175402.743142 Ops/s; put composition (!); 15x15
26491.889313 Ops/s; put composition (!); 75x75
3078.163772 Ops/s; put composition (!); 250x250
84864.676339 Ops/s; mixed composition (!); 15x15
17957.760989 Ops/s; mixed composition (!); 75x75
2303.079948 Ops/s; mixed composition (!); 250x250
162147.355164 Ops/s; lines composition (!); 15x15
22093.150685 Ops/s; lines composition (!); 75x75
2475.877193 Ops/s; lines composition (!); 250x250
349003.125000 Ops/s; Simple Blit (!); 15x15
118175.824176 Ops/s; Simple Blit (!); 75x75
20220.588235 Ops/s; Simple Blit (!); 250x250
149662.500000 Ops/s; Transformed Blit Linear (!); 15x15
13285.561878 Ops/s; Transformed Blit Linear (!); 75x75
1219.642857 Ops/s; Transformed Blit Linear (!); 250x250
1509.113001 Ops/s; Transformed Blit Billinear sharp edges (!); 15x15
62.765957 Ops/s; Transformed Blit Billinear sharp edges (!); 75x75
31.496063 Ops/s; Transformed Blit Billinear sharp edges (!); 250x250
31043.565684 Ops/s; transformed texture paint (!); 15x15
1734.375000 Ops/s; transformed texture paint (!); 75x75
156.926407 Ops/s; transformed texture paint (!); 250x250
60715.648855 Ops/s; gradient+temp texture(!); 15x15
7489.285714 Ops/s; gradient+temp texture(!); 75x75
804.263566 Ops/s; gradient+temp texture(!); 250x250
65253.510140 Ops/s; 12pt text grayscale; 5x5
23862.053571 Ops/s; 12pt text grayscale; 15x15
64184.615385 Ops/s; 12pt text lcd; 5x5
22251.893939 Ops/s; 12pt text lcd; 15x15

[Linuxhippy]

Thanks a lot, energyman.
Sad, still no serious RENDER accaleration ... and that almost 2 years after the R600 launch.

- Clemens

[energyman]

*shrugg* all I know that this 'benchmarks' are not telling the whole story. Just try 9.2 and see for yourself. The only thing slow is manually resizing. Everything else is fast and smooth. Scrolling, window movement, inimizing, maximizing. No 'start' lag. Everything is instantious.

And even 3d is better. Not so much the fps, but the feeling...

[Linuxhippy]

I don't care about 3D, but 2D is as far as I can see almost only software-only rendering, with the exception of some simple blit and copy accaleration.

Software-Only 2D rendering is not bad, and of course far better than weak accaleration causing ping-pong of pixmap contents - but its not good either. Its at best acceptable, but still at a low level.
From the benchmark results I conclude you have quite a fast CPU, which also compensates that.

After all, that makes the ATI HD series the only modern GPU generation without descent 2D accaleration.
- Intel's drivers are EXA based, with full composition and solid support
- OpenChrome via drivers accalerate composition
- NVidia has introduced an excellent RENDER accaleration in 177.80 drivers, even the reverse engineered nouveau driver accalerated composition.
- radeonhd has experimental/unstable/broken accaleration for some composite operations. No idea how long this will take to mature, and I guess it will take even longer to enter fglrx if that should happen at all.

Especially now with many applications switching to RENDER, proper accaleration in the proprietary drivers is quite a must.
Please don't get me wrong, I am not a fanboy of anybody. I am just a software-developer trying to cope with the fact that his software won't be accalerated *only* on AMD hardware these days.
And when seeing that my HD3850 can only render 16 250x250 images/s this impacts real-world animation performance an awful lot. (GF6600 does 850/s, intel 954GM 328/s).

- Clemens

PS: And gtkperf is quite a real-world benchmark, simulating the responsivness of gtk applications. If you compare nvidia with fglrx there's a large difference through the whole benchmark, which tests quite a lot of unrelated / drifferent functionality.

[Kano]

Every single gtkperf test is usually below 1s and combined below 3s with a 8800 GTS 512... So from benchmarking fglrx must be real slow *g*

[energyman]

*shrug* when I switched ~6month ago, 2d with nvidia was unbearable slow. With ati it was bearable. And with the latest liked composite is pleasant. And I tried every hack and workaround with nvidia back then... but I have done nothing so far to 'tune' the ati driver.
And gtkperf is not such a great benchmark anyway - the difference of performance or ondemand as cpu governor is pretty big - with all cards.

[TASADAR-F]

Catalyst 9.2 Beta doesn't work in my Radeon HD 4670.

[energyman]

are you sure you installed it correctly?

[DarkFoss]

doesn't build for me on Mandriva 2009.0 x86..oh well the 9.1 beta's are working fine anyways :)

[Stormking]

Some progress for R300-Owners who were affected by the Cannot-allocate-Memory-Bug (mostly Mobility Radeon 9600 & 9700):

Kernel module finally loads, again. But after that, the screen stays black. No errors in the Xorg.0.log.

[energyman]

where did you download it - and where does it fail?

[Stormking]

where did you download it - and where does it fail?
I used the third (the direct) link in the opening post. It installs without problems and the kernel module loads. But when starting X via GDM (or startx), the screen simply remains black. There are no errors in the Xserver's logfile.

It seems to me that X started up as it should and the driver itself just doesn't display anything. Is that possible?

[RealNC]

It seems to me that X started up as it should and the driver itself just doesn't display anything. Is that possible?

Does the machine actually respond? Can you switch VTs? If not, then the machine is hanging before anything can be written to the log.

[energyman]

hm, strange. I have this problem for a couple of weeks:
at first try to log in, it fails, and I am back at the prompt
second try succeeds, but no glx and no dri
third try, everything is perfect.

I seem to remembewr that the first problem was some kind of kdm screwup. But I forgot where to look at.

[Stormking]

Does the machine actually respond? Can you switch VTs?
I can't remember when switching VTs was possible for the last time. I think two or three years ago. But I will try to login from another machine via ssh to see if the hole system hangs or just X.

If not, then the machine is hanging before anything can be written to the log.
Maybe. But the log looks like X started up. It doesn't just stop somewhere in the process.

[Stormking]

hm, strange. I have this problem for a couple of weeks:
at first try to log in, it fails, and I am back at the prompt
second try succeeds, but no glx and no dri
third try, everything is perfect.

I seem to remembewr that the first problem was some kind of kdm screwup. But I forgot where to look at.
I will experiment a bit, tommorrow. Yesterday i had to switch back to the open source driver because I had some work to do.

[energyman]

btw,

http://www.ati-forum.de/files/Downloads/Driver/Linux_catalyst_9.2_CES_09_preview_driver.zip

is not even extractable for me ....

[Stormking]

btw,

http://www.ati-forum.de/files/Downloads/Driver/Linux_catalyst_9.2_CES_09_preview_driver.zip

is not even extractable for me ....
I had no problem unpacking it using unzip.

[energyman]

unzip is fine. but sh .... .run --extract fails.

[Stormking]

unzip is fine. but sh .... .run --extract fails.
Not for me. I can send you a tarball, if you'd like.

[energyman]

you don't have to. I downloaded it from the chinese site and it worked fine.

[Kano]

Maybe your unzip tool is broken or your ram ;)

[Tares]

hm, strange problem:
aticonfig --odgt
ERROR - Could not find library: libatiadlxx.so

but:
ls -lh /usr/lib64/libatiadlxx.so
-rwxr-xr-x 1 root root 147K 9. Jan 15:37 /usr/lib64/libatiadlxx.so

Same here. I've switched from 9.1 to 9.2 and get this. I remember that switching from 8.12 to 9.1 gave me the same thing. In the meantime I've moved to ubuntu amd64, so on fresh install I've installed 9.1 and everything was fine. I think that the installer don't override the old file, but when I've copied the file from extracted 9.2 to /usr/lib64 I still recive the same error (although those files were different ;/).

[Kano]

Why don't you use the unistall feature first? Usually at:

/usr/share/ati/fglrx-uninstall.sh

[grege]

I have an ASUS EAH3450 running on a Gigabyte GA-G33M-S2H and the DPMS functionality has never worked. With this release it still doesn't. With options in xorg.conf I can get the display to go to sleep, but the only way to wake it up is ctrl-alt-bkspce or with a normal config it just never goes to sleep. The same card on an NVidia chipset motherboard (my spare machine) works as expected. The onboard G33 or an NVidia 8400 has no problem on the same motherboard. Go figure.

I think that I may have some conflict between the HD3450 and the G33 chipsets that is unresolvable.

I did report the fault to AMD. Maybe I am the only sufferer?

Question - does anyone with a G33 motherboard have a successful setup?

[energyman]

I usually use ebuilds. Thank you
btw, my ram is fine.

[energyman]

I have an ASUS EAH3450 running on a Gigabyte GA-G33M-S2H and the DPMS functionality has never worked. With this release it still doesn't. With options in xorg.conf I can get the display to go to sleep, but the only way to wake it up is ctrl-alt-bkspce or with a normal config it just never goes to sleep. The same card on an NVidia chipset motherboard (my spare machine) works as expected. The onboard G33 or an NVidia 8400 has no problem on the same motherboard. Go figure.

I think that I may have some conflict between the HD3450 and the G33 chipsets that is unresolvable.

I did report the fault to AMD. Maybe I am the only sufferer?

Question - does anyone with a G33 motherboard have a successful setup?

no, I see the same. With my 8600GT dpms worked, with my HD3870 it does not

[Tares]

Why don't you use the unistall feature first? Usually at:

/usr/share/ati/fglrx-uninstall.sh

Because I'm using distro builds, so I think when I purge debs all unnessesary files are deleted.

[Kano]

Depends, the newer Ubuntu packageing works better, it just leaves the /etc/ati dir, which you should purge manually after install. I gave up using the Debian packageing because it was broken by default - I fixed it at the beginning but then never used it again. I don't know of other packages, only those which I use in my script.

[Linuxhippy]

*shrug* when I switched ~6month ago, 2d with nvidia was unbearable slow. With ati it was bearable.

Yes, because 6 months ago nvidia had weak hw accaleration, AMD had and has almost no 2D accaleration.
Weak accaleration means some operations were accalerated while others were not, leading to a lot of image-moving between VRAM and normal RAM.
If nothing is accalerated of course, you don't have this problem ;)

Starting with the 178.70 drivers NVidia released pretty good RENDER/2D accaleration. I am not saying that their driver is without problems, but when it comes to 2D accaleration is probably the best driver arround. It even accalerates RENDER's filter effects (blur, ...) with shaders on GF8+.


And gtkperf is not such a great benchmark anyway - the difference of performance or ondemand as cpu governor is pretty big - with all cards.

gtkperf is a real-world benchmark. Typical 2D software consumes both, CPU and GPU time, but usually a lot more CPU resources.
Therefor it pretty well simulates real-world GUI behaviour, so what you get when you work (resize,...) with gtk based programs.

[energyman]

a nvidia using friend of mine really loves the nvidia drivers! With them the text in some kde menus (like the one to add plasmoids) is white on white - whatever he does to change that, it does not help .... he is using nouveau now ...

[Kano]

Then his card must be really old. Most fixes are for GF8+. Older drivers got last update over 6 month ago. Ok, compared to fglrx which supports (in theory) all DX8+ cards you get more updates for fglrx, but the driver quality and feature support (look at VDPAU) is better with NV.

[grege]

no, I see the same. With my 8600GT dpms worked, with my HD3870 it does not

Thanks for the answer. I just installed the radeohd driver and the DPMS works! Unfortunately the radeonhd can't do much else, but at least I know it can work.

[energyman]

Then his card must be really old. Most fixes are for GF8+. Older drivers got last update over 6 month ago. Ok, compared to fglrx which supports (in theory) all DX8+ cards you get more updates for fglrx, but the driver quality and feature support (look at VDPAU) is better with NV.

really old - like a 7900 (afaik)?

[madman2k]

After all, that makes the ATI HD series the only modern GPU generation without descent 2D accaleration.
- Intel's drivers are EXA based, with full composition and solid support
- OpenChrome via drivers accalerate composition
- NVidia has introduced an excellent RENDER accaleration in 177.80 drivers, even the reverse engineered nouveau driver accalerated composition.
- radeonhd has experimental/unstable/broken accaleration for some composite operations. No idea how long this will take to mature, and I guess it will take even longer to enter fglrx if that should happen at all.
the radeon driver has pretty solid EXA/ Render support. It was just the EXA implementation itself holding it back, but with Xserver1.6 everything flies. Just try a Jaunty liveCD.

[Tares]

Depends, the newer Ubuntu packageing works better, it just leaves the /etc/ati dir, which you should purge manually after install. I gave up using the Debian packageing because it was broken by default - I fixed it at the beginning but then never used it again. I don't know of other packages, only those which I use in my script.

Still no luck with fglrx-uninstall.sh and manual remove of /etc/ati. Do you know where are all the fglrx files for manual remove ?

[grege]

Still no luck with fglrx-uninstall.sh and manual remove of /etc/ati. Do you know where are all the fglrx files for manual remove ?

To get rid of fglrx in Ubuntu open Synaptic and install the Ubuntu fglrx, then uninstall. This will clean up a lot and restore symlinks to default.

I also delete xorg.conf then run "sudo dpkg-reconfigure -phigh xserver-xorg" to generate a clean xorg.conf.

I hope this helps, I change my video a lot, switching between my HD3450 and the onbord G33, because fglrx causes me such grief. I live in hope that radeon, radeonhd and ati all merge into a great open source driver that just works and I can forget fglrx forever.

[Linuxhippy]

a nvidia using friend of mine really loves the nvidia drivers! With them the text in some kde menus (like the one to add plasmoids) is white on white - whatever he does to change that, it does not help .... he is using nouveau now ...
I don't see how this relates to the original topic, it seems the old phenomena - people defending their hardware ;)
I don't even use a NVidia GPU, the only one I have is a GF6600 I plug in sometimes to test my software on.

As I said I am not a fanboy of anyone and I am not interested in a flame, I also explicitly said their drivers *do* have problems, but so do the ATI drivers...


the radeon driver has pretty solid EXA/ Render support.

Yes the readeon driver does pretty well for RENDER, but only for R500 and below. There's no accaleration code for R600 and up, released in early 2007.

- Clemens

[BlackStar]

Installed these drivers:

1. no - or very slight - improvement to compiz

2. aticonfig --odgc returns "ERROR - Could not find library: libatiadlxx.so"

Haven't tried video playback yet.

Maybe amd should consider a public beta program?

[RealNC]

Then his card must be really old. Most fixes are for GF8+. Older drivers got last update over 6 month ago. Ok, compared to fglrx which supports (in theory) all DX8+ cards you get more updates for fglrx.

Nah, DX8 cards have been abandoned long ago. The last fglrx supporting them was even before the "Catalyst for Linux" appeared; last release for them was fglrx 8.28.8. Everything after that was DX9+ only (Radeon 9500 and above).

[Kano]

Right, they are DX9, but not the fastest ones ;) So basically only GF5 series has lower status from the DX9 NV cards then.

[TASADAR-F]

are you sure you installed it correctly?

Yes.

The problem is the first screen "login screen" appear with lots lots lots pixels of lots lots colors. You can't read anything.

[energyman]

btw, friend just told me that composite with the latest stable nvidia is still broken. If you try to 'pull' marked text in ff, composite is switched off. The good news - the same happens with ati. So both are broken.

[Kano]

Well I personally don't need composite. I prefer using vdpau without tearing. Desktop 3d effects are nice to show some V user that Linux can handle more 3d then V, but for everyday use nobody really needs it. U enables it by default and uses only minimal effects, you would hardy see a difference when it is out anyway.

[gsacks]

Well I personally don't need composite. I prefer using vdpau without tearing. Desktop 3d effects are nice to show some V user that Linux can handle more 3d then V, but for everyday use nobody really needs it. U enables it by default and uses only minimal effects, you would hardy see a difference when it is out anyway.

Well, I need it for everyday use. I run Ubuntu on an HTPC and a 720P plasma screen. When I am using that computer and browsing the web from 10 feet away, text is unreadable. I use the 'zoom' feature of compiz to make the text of the article I am reading large enough to fill the screen. I do this reading Phoronix forums ALL THE TIME. (Although I am using an XP desktop at the moment). I also use it to zoom in on undersized video clips in web pages and make them as big as I want. Everybody's individual use case is different.

I've read a lot of your posts on Phoronix and I respect your knowledge of Linux. It is far greater than mine, but I am just getting sick of reading people state that composite desktop features have to value, just because they don't use it.

[Kano]

Maybe you don't know that FF has like all common web browsers a built-in function to resize text when you hold ctrl + use mouse wheel?

[yotambien]

Maybe you don't know that FF has like all common web browsers a built-in function to resize text when you hold ctrl + use mouse wheel?

I personally prefer CTRL + "+", I hate to take my hands off the keyboard, but good trick anyway.

\:D/

[Aphax]

I gave xf86-video-ati-6.10.0 a try the other day and noticed 2D was blitzing fast compared to fglrx; doing a 'dmesg' in xterm was now instanteneous (with fglrx I'd see it scroll text for a while), resizing gvim was superfast too (seems to do a lot of redrawing, more so than normal gtk+ apps), even the graphs rendered by gnome-system-monitor which are horribly slow on fglrx were smooth as butter on ati.

People in #radeon on freenode told me this is likely because my (C2D E8400) CPU which was now doing all the work because there was no 2D accel for r600+ chips. which I guess makes sense. So I then tried to see what performance was like with the NoAccel option set with fglrx, but that was really way worse :/

I'm not sure what the downsides are of having my CPU doing 2D as it is with xf86-video-ati, I tried it for a few days and didn't notice any other adverse effects other than playing video (understandably) using a lot more CPU time. I was especially happy about gvim, I use it a lot when I hack on code and faster scrolling and window manipulation (split screens etc.) really makes it much more pleasurable

[Tares]

Anyone has solved the libatiadlxx.so problem ?

[storma]

See this thread:

http://www.phoronix.com/forums/showthread.php?t=14252&highlight=libatiadlxx.so

[bridgman]

I gave xf86-video-ati-6.10.0 a try the other day and noticed 2D was blitzing fast compared to fglrx; doing a 'dmesg' in xterm was now instanteneous (with fglrx I'd see it scroll text for a while), resizing gvim was superfast too (seems to do a lot of redrawing, more so than normal gtk+ apps), even the graphs rendered by gnome-system-monitor which are horribly slow on fglrx were smooth as butter on ati.

People in #radeon on freenode told me this is likely because my (C2D E8400) CPU which was now doing all the work because there was no 2D accel for r600+ chips. which I guess makes sense. So I then tried to see what performance was like with the NoAccel option set with fglrx, but that was really way worse :/

I'm not sure what the downsides are of having my CPU doing 2D as it is with xf86-video-ati, I tried it for a few days and didn't notice any other adverse effects other than playing video (understandably) using a lot more CPU time. I was especially happy about gvim, I use it a lot when I hack on code and faster scrolling and window manipulation (split screens etc.) really makes it much more pleasurable

What you are seeing is normally called "shadowfb" acceleration. The primary copy of the frame buffer is kept in system memory, where the CPU can access it really quickly. As the CPU updates portions of your screen, those portions are copied to the "real" frame buffer in video memory. For 2D acceleration it runs pretty much as fast as any acceleration hardware, and faster than some hardware.

Running "no-accel" is different -- the CPU is still doing all the drawing but it is drawing into video memory, where the accesses are much slower. This is why the performance is so different.

The real drawback of shadowfb is that you can't really use it with ANY hardware acceleration, ie it's not practical to have 2D accel done by shadowfb but have video and 3D acceleration done by hardware. It might be possible to make some kind of hybrid but nobody is doing it yet.

When all the HW vendors started dropping dedicated 2D acceleration blocks and using the 3D engine for 2D acceleration, some things (like EXA Render) got better while other things (XAA acceleration) got worse. I think you will see 2D hardware acceleration continue to improve over the next 6-12 months. Shadowfb will continue to be an annoying thorn in our sides as a reminder of what is possible for 2D-only on the CPU. Where the CPU falls down is where you get into computationally expensive things like video scaling and 3D acceleration -- there the 3D engine shines and can run rings around a typical CPU.

[gsacks]

Maybe you don't know that FF has like all common web browsers a built-in function to resize text when you hold ctrl + use mouse wheel?

Actually, no, I didn't know that, since I never needed that function before I built my HTPC. Thanks for the tip. That will come in handy. It isn't quite as smooth as using compiz zoom, and still does not allow me to quickly make the browser window fill the screen and then scroll it off to the side so I can work on another part of the desktop. (Yes, I do have multiple desktop screens). But I can definitely see myself using that feature sometimes. I will still maintain that composite can add to the user experience, but there are always many different ways to solve a task, and I guess composite desktops don't fit the bill for you. They do for me. So I look forward to a driver that will allow it to work better alongside GL apps. I agree that most of it is eye-candy, but there are a few key features that work well for me in my set-up. Can I live without them? Sure. Do I want to? No. Besides, what is wrong with eye candy, anyway. We could all still be using amber screen 24x80 text console displays. Wouldn't that be fun...

[russell_h]

Installed flawlessly for me. The only issue I've encountered so far is that sometimes, especially if I am skipping around or resizing the window, video playback (mplayer, xv) will just go black. Only way to fix it seems to be to restart mplayer.

Tearing is definitely still a problem though. Anyone know anything about that?

[TASADAR-F]

Tearing is definitely still a problem though. Anyone know anything about that?

I have the same problem with fglrx,ati and radeonHD. In my 4670.

Last version of open source ati solve problem in r300,r400 and r500.

R600 (HD2000 and HD3000 series) and R700 (HD4000 series) will have wait to ati or radeonHD support XV output.

I hate the tears.

[Aphax]

What you are seeing is normally called "shadowfb" acceleration. The primary copy of the frame buffer is kept in system memory, where the CPU can access it really quickly. As the CPU updates portions of your screen, those portions are copied to the "real" frame buffer in video memory. For 2D acceleration it runs pretty much as fast as any acceleration hardware, and faster than some hardware.

Running "no-accel" is different -- the CPU is still doing all the drawing but it is drawing into video memory, where the accesses are much slower. This is why the performance is so different.

The real drawback of shadowfb is that you can't really use it with ANY hardware acceleration, ie it's not practical to have 2D accel done by shadowfb but have video and 3D acceleration done by hardware. It might be possible to make some kind of hybrid but nobody is doing it yet.

When all the HW vendors started dropping dedicated 2D acceleration blocks and using the 3D engine for 2D acceleration, some things (like EXA Render) got better while other things (XAA acceleration) got worse. I think you will see 2D hardware acceleration continue to improve over the next 6-12 months. Shadowfb will continue to be an annoying thorn in our sides as a reminder of what is possible for 2D-only on the CPU. Where the CPU falls down is where you get into computationally expensive things like video scaling and 3D acceleration -- there the 3D engine shines and can run rings around a typical CPU.


Thanks for the explanation :) That answers a bunch of questions I had about why 2D acceleration seems to be in such a state of flux at the moment. I'm looking forward to the day it's all properly done on the 3D hardware

[BlackStar]

Has anyone managed to install these on Archlinux? The PKGBUILD scripts complete, the packages are installed, but I get a black screen with some corruption on restart (VT switching works fine, however).

I'll try to find out what is going wrong, but I wonder if anyone has managed to install these correctly.

(System is Archlinux x86_64 with a 4850 video card.)

Edit: It seems that the kernel module fails to load due to "flush_tlb_page" missing.
WARNING: "flush_tlb_page" [/tmp/catalyst/src/archive_files/common/lib/modules/fglrx/build_mod/fglrx.ko] undefined!

A quick search showed this to be an old problem (around kernel 2.6.25) - I'll try the solutions listed there and see if they work.

Edit 2: Aha, the 2.6.27.patch seems to already contain a fix - maybe it didn't get applied?

Edit 3: Indeed, it didn't. Even worse, it doesn't seem to apply anymore. I give up - waiting 2 months isn't so bad after all :)

[Tares]

See this thread:

http://www.phoronix.com/forums/showthread.php?t=14252&highlight=libatiadlxx.so

That doesn't help me, even if it would, that just a workaround not a solution :/

[nikopol]

Last version of open source ati solve problem in r300,r400 and r500.

R600 (HD2000 and HD3000 series) and R700 (HD4000 series) will have wait to ati or radeonHD support XV output.

The tearing has really been the bane of my life too (rv630 - hd 2600 xt). I'm not even that bothered about the 3d, all I want it to be able to watch DVDs and video files that don't look like someone is slicing every image in half with a blade.

Is there a version of catalyst that doesn't have tearing on it for my hardware? I'd be tempted to go old school to resolve the issue.

[grege]

The tearing has really been the bane of my life too (rv630 - hd 2600 xt). I'm not even that bothered about the 3d, all I want it to be able to watch DVDs and video files that don't look like someone is slicing every image in half with a blade.

Is there a version of catalyst that doesn't have tearing on it for my hardware? I'd be tempted to go old school to resolve the issue.

An NVidia 8400 GS does not cost much ($30 ish) and will solve all of your problems. I have a HD3450 sitting in the drawer waiting for real drivers. I also have a 780G (= HD3100 ) motherboard and fglrx 8.12 works reasonably well, dpms works and xvid avis play well enough. HD TV does not play nicely, but standard def DVB-T is OK. Eventually the ati - radeon - radeonhd driver will work better in a HTPC setup than fglrx. I wait in hope, meanwhile the cheap nvidia option is not silly.

[legume]

The tearing has really been the bane of my life too (rv630 - hd 2600 xt). I'm not even that bothered about the 3d, all I want it to be able to watch DVDs and video files that don't look like someone is slicing every image in half with a blade.

Is there a version of catalyst that doesn't have tearing on it for my hardware? I'd be tempted to go old school to resolve the issue.

If you use current svn mplayer, turn on vsync with amdcccle and use gl instead of xv -

mplayer -vo gl:yuv=4 ... it should work.

It may not matter for dvd, but 8-10 used less CPU that 8-12 when I tried it.

[nikopol]

An NVidia 8400 GS does not cost much ($30 ish) and will solve all of your problems. I have a HD3450 sitting in the drawer waiting for real drivers. I also have a 780G (= HD3100 ) motherboard and fglrx 8.12 works reasonably well, dpms works and xvid avis play well enough. HD TV does not play nicely, but standard def DVB-T is OK. Eventually the ati - radeon - radeonhd driver will work better in a HTPC setup than fglrx. I wait in hope, meanwhile the cheap nvidia option is not silly.

Will look into that - good tip, thanks.

[Boerkel]

If you use current svn mplayer, turn on vsync with amdcccle and use gl instead of xv -

mplayer -vo gl:yuv=4 ... it should work.

It may not matter for dvd, but 8-10 used less CPU that 8-12 when I tried it.

It doesn't work for me. It helps to minimize the tearing but, it is still there. There is one horizontal "tearing-line" that wanders up my monitor. Unless this annoying "thing" there is absolutely no tearing.

I'm using Ubuntu 64bit Hardy Heron with a Radeon HD 2600. My mplayer version is mplayer - 2:1.0~rc2svn28238-0hardy2, but i usual use smplayer.

[RealNC]

You need to enable VSync in Catalyst Control Center.

[Melcar]

Speaking of Vsync. Do any of the options work for you guys (on any driver version)? For me, the controls either force vsync on globally or force it off globally. The in between options (off/on unless specified) don't work as intended.

[Boerkel]

You need to enable VSync in Catalyst Control Center.

It is enabled...

[nikopol]

If you use current svn mplayer, turn on vsync with amdcccle and use gl instead of xv -

mplayer -vo gl:yuv=4 ... it should work.

It may not matter for dvd, but 8-10 used less CPU that 8-12 when I tried it.

Well maximum kudos - you resolved the long standing playing of avi files for me. It's pretty much undetectable now.

Using Radeon HD 2600XT on ubuntu intrepid

[nikopol]

Actually I seem to get a similarly good result with xine witn opengl - or maybe my eyes are starting to fail me from staring at the screen too close for too long.

[legume]

It is enabled...

Strange - does the framerate if you run glxgears = your refresh rate?

I guess running out of CPU or double rate deinterlacing so framerate = refresh rate could also mess it up.

[legume]

Speaking of Vsync. Do any of the options work for you guys (on any driver version)? For me, the controls either force vsync on globally or force it off globally. The in between options (off/on unless specified) don't work as intended.

That's what I see as well

[legume]

Actually I seem to get a similarly good result with xine witn opengl - or maybe my eyes are starting to fail me from staring at the screen too close for too long.

Cool - I haven't tried xine, so I don't know what it's gl is like WRT CPU usage (eg it may do color space conversion in software, whereas the yuv=4 option uses GPU).

It doesn't really matter for SD mpeg2, but if you want to play HD h264 and have an older CPU like mine then it can make a difference. If you are scaling small res file to a large res display with mplayer you can also try bicubic rather than bilinear with lscale=1.

[nstamoul0]

One kinda off-topic question.

I have a debian/sid setup running the latest 2.6.26 kernel.The only way I can get new drivers to run (outside my repo) is to compile a new kernel (because of the flush tlb page bug (http://www.phoronix.com/forums/archive/index.php/t-7631.html)).

Anybody know a way around this (besides compiling a new kernel ?).

[Vighy]

One kinda off-topic question.

I have a debian/sid setup running the latest 2.6.26 kernel.The only way I can get new drivers to run (outside my repo) is to compile a new kernel (because of the flush tlb page bug (http://www.phoronix.com/forums/archive/index.php/t-7631.html)).

Anybody know a way around this (besides compiling a new kernel ?).

change distro.

[the-me]

One kinda off-topic question.

I have a debian/sid setup running the latest 2.6.26 kernel.The only way I can get new drivers to run (outside my repo) is to compile a new kernel (because of the flush tlb page bug (http://www.phoronix.com/forums/archive/index.php/t-7631.html)).

Anybody know a way around this (besides compiling a new kernel ?).

8-12-4 is in Debian Sid now, which does not have this issue.

[nstamoul0]

change distro.

When hell frezes over :p

I'm already back to 8.12

[Vighy]

When hell frezes over :p

I'm already back to 8.12

so unstable to be not worth-trying? :) I will peacefully wait for 9-01 then...

...when it will ever arrive :D

[Boerkel]

Strange - does the framerate if you run glxgears = your refresh rate?

I guess running out of CPU or double rate deinterlacing so framerate = refresh rate could also mess it up.

Yes glxgears = refresh rate :
301 frames in 5.0 seconds = 60.195 FPS
301 frames in 5.0 seconds = 60.066 FPS

I don't think CPU is a problem. I've got a 2x2.0 GHz Turion and CPU-usage is far below 100%. So there must be another reason for this strange "tearing-line". I really don't know why this happens... :(

[biasquez]

(EE) Failed to load module "fglrx" (module requirement mismatch, 0)

error with jaunty (doesn't detect xorg 7.4)

driver support jaunty (infact Ubuntu/Jaunty is present in --listpkg and --buildpkg Ubuntu/Jaunty compile without error)

[suokko]

Yes glxgears = refresh rate :
301 frames in 5.0 seconds = 60.195 FPS
301 frames in 5.0 seconds = 60.066 FPS

I don't think CPU is a problem. I've got a 2x2.0 GHz Turion and CPU-usage is far below 100%. So there must be another reason for this strange "tearing-line". I really don't know why this happens... :(

That is usually because your driver is waiting for display sync. You should check the settings. driconf has option "synchronization with vertical refresh" for that

[Aphax]

Yes glxgears = refresh rate :
301 frames in 5.0 seconds = 60.195 FPS
301 frames in 5.0 seconds = 60.066 FPS

I don't think CPU is a problem. I've got a 2x2.0 GHz Turion and CPU-usage is far below 100%. So there must be another reason for this strange "tearing-line". I really don't know why this happens... :(

With proper vsync you should never see tearing even if you can't draw at at your display's refreshrate, because as far as I know it will just delay swapping the buffers on the next refresh. So basically if you see tearing then either vsync isn't turned on or the driver is failing in some way.

[energyman]

one question: which wm are you using again?

[Melcar]

fglrx tears. I though it was common knowledge by now. Vsync. "works" in so far that it keeps frames in sync with your monitor's refresh rate, but images still suffer from tears (whether it be videos or games).

[energyman]

it doesn't tear here. But I am not using broken compiz but kde 4.2 prereleases.

[Boerkel]

I'm using Gnome and switching Compiz off doesn't solve anything. It's really strange. Perhaps fglrx works fine and my xserver or anything else is broken?

[frep]

Hello! i installed beta driver on ubuntu hardy its working perfectly. My card is a x800xl.
Lightsmark work and its fast:)

I followed this guide for the installation and replaced 8.561 with 8.580:
http://wiki.cchtml.com/index.php/Ubuntu_Hardy_Installation_Guide

tested also google earth and compiz enabled, there is anyway the hassle of flickering. i really dislike it.
Im going to make further tests with an agp card.

[SolidSteel144]

Can you guys help me correctly install this driver? I'm new to installing them manually.
On Jaunty...

[biasquez]

Can you guys help me correctly install this driver? I'm new to installing them manually.
On Jaunty...

you can't install this driver on jaunty: there is support for jaunty in --listpkg and directory x740 but fglrx module is compiled for xorg 7.1 and not 7.4...

[legume]

Yes glxgears = refresh rate :
301 frames in 5.0 seconds = 60.195 FPS
301 frames in 5.0 seconds = 60.066 FPS

I don't think CPU is a problem. I've got a 2x2.0 GHz Turion and CPU-usage is far below 100%. So there must be another reason for this strange "tearing-line". I really don't know why this happens... :(

Strange - I don't know why it's not working.

The only thing I can think of is to check the output of mplayer to make sure it really is using gl and not falling back to x11 for some reason.

If you are using composite/compiz that will mess things up.