Messages in this thread |  | | Date | Mon, 14 Jun 1999 11:50:56 -0400 (EDT) | | From | hagopiar@vuser ... | | Subject | Re: EPIC |
| |
It should actually be faster... 64-bit CPUs will have 128-bit memory busses so with SDRAM you can supply one instruction bundle every cycle of the FSB. With a lot of code using predicates (I'm really starting to like the predicate idea :-) ) instead of real branches you can keep SDRAM (4-1-1-1 for pc100) going at full speed. That's 3 instructions per clock cycle with no stalls. You could blow through a 12 instruciton branch (4 cycles * 3 instrs/bundle) done with predicates at the same speed as branching around it.
When you add an I-cache (the docs say nothing of specific implemtations yet) things will get even faster (but they also change a bit - it's no longer quicker to blow through a 12 instruction stream). I can't see any reason that a multi-unit processor would come into existance any time soon. It basicly defeats the purpose of the architecture which is to get _away_ from superscalar processors and put the parallelism work into the compiler. -Rob
On Sun, 6 Jun 1999, Matti Aarnio wrote:
> Date: Sun, 6 Jun 1999 19:21:35 +0300 > From: Matti Aarnio <matti.aarnio@sonera.fi> > To: David S. Miller <davem@redhat.com> > Cc: linux-kernel@vger.rutgers.edu > Subject: Re: EPIC > > On Sun, Jun 06, 1999 at 09:08:28AM -0700, David S. Miller wrote: > > You have precluded the possibility of dispatching multiple bundles per > > clock cycle, and I am rather sure this will be a characteristic of > > various EPIC processors. ... > > You are right, but lets see; presumed 1 clock per bundle (or multiple > bundles) means the instruction stream must be at least 16 bytes wide, and > if two bundles at the time are to be executed, 32 byte memory read width > is needed. Gee, that is quite a monster.. > > Of course having wide cache-cpu bus does not mandate as wide main-memory > interface. > > Definitely not cheap systems. > > > Later, > > David S. Miller > > davem@redhat.com > > /Matti Aarnio <matti.aarnio@sonera.fi> > > - > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > the body of a message to majordomo@vger.rutgers.edu > Please read the FAQ at http://www.tux.org/lkml/ >
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
|  |