OS4 on the pegasos

[Karlos]

Mr_Capehill wrote:
@Kenny:

Out of interest, how have they implemented memory protection and altivec support in MorphOS?

That's quite an intresting question actually. I don't really know much about MOS kernel. Has anybody got any information about this?

[Karlos]

Acill wrote:

When the hardware was being shown to them they also had a Teron (A1 is the same as this) and a G5 dual system. The Pegasos beat them all out.

So, your'e saying a single CPU G3 based machine beat a dual CPU G5 based one. At what, exactly?

[Karlos]

Acill wrote:

Karlos wrote:

Acill wrote:

When the hardware was being shown to them they also had a Teron (A1 is the same as this) and a G5 dual system. The Pegasos beat them all out.

So, your'e saying a single CPU G3 based machine beat a dual CPU G5 based one. At what, exactly?

Where did I say G3? It was a 1GHZ G4 system running linux that was used. It out preformed the G5 system in several systems. In the end it was picked for qualiety and prince point with matching performance. Read the Genesi and Morphzone news and take a look at hese things from time to time. Freescale is the branch of Motorola that seperated BTW.

My understanding was that the common or garden variety Peg was a G3 600MHz based system (with the CPU on a replacable card) and that the Peg II was a 1GHz G4 machine.

So basically, nowhere did you say it wasn't a basic Peg 1, hence my assumption you meant the G3 600MHz.

[Karlos]

Acill wrote:
Just get a Pegasos and run MorphOS. I dont even miss my A3000 since I started using the Pegasos II I have now. Its so much better then the current state of the A1 and I wouldnt put OS4 on it even if it did work as it is now! MOS is light years ahead of it!

All this MOS v OS4 is frankly a bit wasted on me - I only have a 1200T. Still, OS4 is pretty damn nippy on my BlizzPPC, so nerr :-P :-)

[Karlos]

Slightly OT,

Cymric wrote:

Apparently games are no longer using table lookups in order to avoid some costly mathematical function evaluations...  

Well, that's because the CPU is now so much faster than the main memory that unless your lookup table is very small (ie will fit in the cache), reading a value from memory can often prove more expensive than calculating it there and then.