DPMS screen saver.

Acorn machines without VRAM share the normal RAM with the display, resulting in progressively worse performance as the screen mode gets bigger or deeper.

I seem to recall old machines (I’m not sure if it included acorns or not) having an option to blank the screen to reduce processing time.

Does an A7000 (or similar) with the DPMS screen saver act in the same way?

ie does the video stop contending with the processing for RAM access, when the screen is blank? (if not, could that be changed?)

And does the same apply to the Beagleboard?

Does video stop contending with the processing for RAM

No, the DPMS driver uses the, er, DPMS lines to merely tell the monitor to go into various power saving modes.

 - ChangeFSI in 'Fast' mode (switches down to a lower mode)
 - ADFS when the floppy DMA underruns (does OS_UpdateMEMC)

ChangeFSI in ’Fast’ mode (switches down to a lower mode)

That’s what I was thinking of.

So if the screensaver were modified to switch to a very low screen mode after enabling DPMS, then the system would be able to process things faster.

(This is more of a curiosity for me, since I have an Iyonix and RPCs with VRAM, but might be allow an A7000 to play MP3 with better quality.)

And does the same apply to the Beagleboard?

I haven’t implemented the GraphicsV_SetBlank code in the driver yet, so at the moment the DPMS screensaver won’t do anything. But I’m fairly certain that the beagleboard does support DPMS.

On the OMAP all the video is done via overlays, so by turning the overlays off we can stop the video controller from accessing RAM without disabling the output signal entirely. All that you’d see on the monitor is the background colour (which I think the current driver sets to the RISC OS border colour).

Is there a trick needed getting the screen saver to work? Mine is always disabled after a re-boot, well switched to “delay off”.

The screen saver delay is saved to CMOS RAM. So until Dave’s SD/MMC code is ready, it’s probably best to add something to your boot sequence that sets the right blank time using *BlankTime.

Thought you might say something like that. Thanks for the clarification!