Have there been any reports of the agpgart speed of the radeon
drivers being downgraded for certain cards? Today, because of breakage
in the current fglrx livna package, we reverted our machine with a
ATI FireGL 8700/8800 QH (AGP) back to the stock radeon driver. This
provided excellent 3D performance when we first installed FC5. Now
even though DRI is initialized the glxgears speed is down near 400 fps.
I believe I have identified the cause of this. Where the fglrx drivers
result in...
Jun 12 10:12:32 carbo kernel: agpgart: Detected an Intel E7505 Chipset.
Jun 12 10:12:32 carbo kernel: agpgart: AGP aperture is 128M @ 0xe8000000
Jun 12 10:18:15 carbo kernel: agpgart: Found an AGP 3.0 compliant device at 0000:00:00.0.
Jun 12 10:18:15 carbo kernel: agpgart: Device is in legacy mode, falling back to 2.x
Jun 12 10:18:15 carbo kernel: agpgart: Putting AGP V2 device at 0000:00:00.0 into 4x mode
Jun 12 10:18:15 carbo kernel: agpgart: Putting AGP V2 device at 0000:00:01.0 into 4x mode
Jun 12 10:18:15 carbo kernel: agpgart: Putting AGP V2 device at 0000:01:00.0 into 4x mode
the radeon drivers cause...
Jun 12 10:33:42 carbo kernel: agpgart: Detected an Intel E7505 Chipset.
Jun 12 10:33:42 carbo kernel: agpgart: AGP aperture is 128M @ 0xe8000000
Jun 12 10:33:53 carbo kernel: agpgart: Found an AGP 3.0 compliant device at 0000:00:00.0.
Jun 12 10:33:53 carbo kernel: agpgart: Device is in legacy mode, falling back to 2.x
Jun 12 10:33:53 carbo kernel: agpgart: Putting AGP V2 device at 0000:00:00.0 into 1x mode
Jun 12 10:33:53 carbo kernel: agpgart: Putting AGP V2 device at 0000:00:01.0 into 1x mode
Jun 12 10:33:53 carbo kernel: agpgart: Putting AGP V2 device at 0000:01:00.0 into 1x mode
to appear in the messages file. Is this a known regression and is there
a why to force the radeon driver back up to 4x AGP?
Jack
