I've a Peg I with a voodoo5 board on AGP bus. Sound outpout is totally corrupted as soon as I scroll in Voyager window, even when Voyager loads a web page the music is slowing down....
I'm really not qualified to entirely answer your question, but, from my observation, Voyager will eat up to 100% CPU when scrolling through some windows, this according to "TaskManager".
IBrowse does not exhibit this "sound" problem, and it's CPU usage when scrolling through windows is considerably less.
Basically, my *guess* is that the sound corruption problem is related to both Voyager (specifically), as well as general instances of high CPU usage. I do not believe, from my experience, that there is any "sound bug" which afflicts MorphOS.
Both my Mac and Athlon experince similar characteristics, only actually a bit *worse* than my Pegasos.
Slightly unrelated...but I just ordered a new SB for my Peg... I'm VERY curious to see how it handles these situations in comparison to the onboard sound.
Have you test the Warp3D-Demo "Engine" on the Voodoo5?
I see Screencorruptions in 800x600 but 137 FPS.
Same way looks Freespace in Warp3D-Mode.
Did have any access to a quake-, doom- and heretic-clone for morphos that works fine? Especially quake-Warp3D or GLQuake with a working sound-support for ahi?
I have try many port's, but only working thing is adoom-mos, but it does'nt load my saved games. Any tips?
Give it any Warp3D- or Rave3D-Demos to test the Voodoo5(4)-drivers funktionability. I think this actual Voodoo4-driver has errors, i had never seen clean graphical results. Did you see better results?
i don't have a V5, but a Voodoo4 in my Peg! There is no problems at all on several games in Warp3D-mode!!! Here my tested applications/games: HereticII, Shogo, QuakeII (Hyperion), GlQuake-WOS (beta11 from Massimilano Tretene) , WipeOut2097. All games runs properly without any errors! I think, it was better when you test the Voodoo5 or check your configuration...
Now, after a long time of testing, i can say not the mos-quarkkernel is the probleme, for me it looks now, that the OF is the reason for the V5-problems, some pci-listing shows me a badly addressed pci-pipeline, so that only the second GPU is "mountet" on this pipeline, also the colorchannels are inverse'd.
Please test it with LinuxPPC-XServer (tdfxfb), the same on MOL (MacOSX)!
I wish that i can present better news, but now it's no more in our hands, this job goes to Genesi or MDC! it looks like the little- vs. big- endian thing is real source. :(