There is also a performance cost with MMU
since you need to translate the virtual address to a physical address.
On the instruction side, one might get
away with a virtual mapped cache but on the data cache this is much harder.
This will lead to that all data accesses
will have one extra clock cycle of latency including data cache hits.
How much this will impact system
performance is very application dependent.
As usual there is no such thing as a free
lunch. Everything costs.
Göran
From:
owner-microblaze-uclinux@xxxxxxxxxxxxxx
[mailto:owner-microblaze-uclinux@xxxxxxxxxxxxxx]
On Behalf Of Tanuj R Oruganti
Sent: Sunday, April 30, 2006 05:52
To: microblaze-uclinux@xxxxxxxxxxxxxx
Subject: RE: [microblaze-uclinux]
Question on MMU
Yes,it does. In fact,I'm trying to get more perspective into
trade-offs between an mmu-less system with stable OS and lower
complexity vs processors with mmu.I guess my question would'nt have been
that vague had I done that earlier.Thanks,for your reply though!
TRO.
Henrik.Westlund@xxxxxxxxxxxxxxx
wrote:
The microblaze was "originally"
intended to replace state machines implemented as logic in a hardware
description language.
Now, the microblaze is capapble of running
a complete operating system! That is a major step.
That was not an answer to you question but
perhaps gave a bit of perspective.
BR Henrik
From:
owner-microblaze-uclinux@xxxxxxxxxxxxxx
[mailto:owner-microblaze-uclinux@xxxxxxxxxxxxxx]
On Behalf Of Tanuj R Oruganti
Sent: den 22 april 2006 15:29
To: microblaze-uclinux@xxxxxxxxxxxxxx
Subject: [microblaze-uclinux]
Question on MMU
I have a question. It is kind of off-topic, so please bear with me. I
don't know how smart a question this is either, so again,bear with me on
that. Going through this list I was just wondering why an MMU for
Microblaze does'nt exist. How much work would doing something like that
involve? Is this a feasible / useful kind of thing for undertaking?
Thanks.
Blab-away for as little as 1¢/min. Make PC-to-Phone
Calls using Yahoo! Messenger with Voice.