Re: Memory access

From: Magnus Kollberg <Magnus.Kollberg_at_emw.ericsson.se>
Date: Tue, 13 Jan 1998 15:28:22 +0100

Hi,

> > the roof.... they were at 2400%. Sounds more correct. Maybe the AB040 fools
>
> It does?

Well as you get 1000% on a AB040 and 2400% on ROM on the Hades it's a bit more than
twice the performance of the AB040 which doesn't sound to far fetched.
 
> A normal Falcon has something like 6Mbyte/s bandwidth to RAM. 24 times
> that would be 144 Mbyte/s, which given a 32 bit bus would require all
> accesses to take less than a single clock cycle at 32MHz.

And as the bus is 64-bit but ineterleaved it will get a bit faster.
 
> Let's say that it was really 128Mbyte/s to keep within the physical limits.
> That still requires an instant setup of each burst transfer, which just
> isn't possible. I'd be very surprised if even 2-1-1-1 was possible.

You are ofcourse right but the interleaved memory architecture of the Hades
will give you better performance. When the CPU has read one 32-bit word the
next is already available. Or????
 
> An AB040 can reach slightly more than 30MByte/s and I'd say that the
> Hades can at best manage a little more than double that. I'd be surprised
> if it did, though.

And so 2400% for GEMBENCH sounds correct as it's about twice as fast as on the
AB040. This is if you use _GEMBENCH_ you will get this 24 figure. :-)
 
> > But also, Petr did say that Quake on the Hades were slower than on the AB040 so
> > something fishy is going on in the Hades.
>
> An '060 should be _very_ good at Quake.
> I don't think the difference between a Hades040 and an AB040 should be
> significant, though.

Probably not and I think it was a Hades040 Petr had tested it on. Talking about Quake
has anyone looked at it and see if anything can be improved?

//Magnus Kollberg
Received on ti. jan. 13 1998 - 20:41:00 CET

This archive was generated by hypermail 2.3.0 : ti. nov. 03 2015 - 20:07:53 CET