On Tue, Aug 26, 2008 at 19:56, Benjamin Huntsman
<[EMAIL PROTECTED]> wrote:
>>I think trap.c would be easier to work with or extend to other
>>instructions, l.s might give better performance.
>
> That was my concern as well.  The Alpha compiler doesn't optimize much.  Not 
> that optimization is going to matter much on an EV4.

Well, even on EV4 it could matter. At least I don't remember Alpha
having any real cache-coherency (or memory coheency at all), also EV4
and EV5 which don't have BWX can get much higher speed from using
8-byte aligned data, as well as sticking to 64-bit data where
possible. All those masking operations are probably slowing the code
much.

> I've got the EBSDK, which was the PALCode developer's kit from the AlphaPC 
> boards.  I don't think there's any pressing need to patch up the PALCode 
> though.  The OSF/1 PALCode that Plan 9 runs with is already designed for 
> UNIX-like systems.  Besides, since they're all implementation-specific, it'd 
> be a bad idea to try to run the EBSDK PALCode on a GS-series machine, should 
> anyone want to run Plan 9 on one in the future. :)

Well, the code for updating PALcode probably should be in bootloader
anyway, and HWPRB *does* report processor type, doesn't it? Patching
only "illegal instruction trap" should be enough. As for GS-series, at
my current job we still have some GS1280 (those were EV7z?) Pity I
wouldn't be able to play with those machines....

And could you share EBSDK? If the license permits, of course.

> -Ben
>



-- 
Paweł Lasek

Reply via email to