By: RichardC (tich.delete@this.pobox.com), April 21, 2015 12:57 pm
Room: Moderated Discussions
Ivan Godard (ivan.delete@this.millcomputing.com) on April 20, 2015 5:27 pm wrote:
> You seem to have mistaken limited resources for lack of interest :-) We are putting
> all we can into the tool chain. The software model of the Mill has been running in sim
> for six years now, although the Mill of six years ago is not the Mill of today.
Running *what* in sim ? What kind of benchmarks, and hand-coded or compiled by what compiler ?
And with what kind of simulated memory system ?
This doesn't seem like a general-purpose architecture, at all. It doesn't really seem to expect
to do procedure call/return very often - and it sure as heck doesn't seem to expect to switch
between threads in a hurry. So what's it for ?
> As for my goals: PDing might get me glory or tenure, but I'm not an academic and already
> have all the glory I'll ever want. The Mill is a commercial product. Yes, hundreds (not
> thousands) of people, and deep pockets ($120M estimated) are needed. The early days of
> design are best done with a small team - but why assume we intend to stay that way?
If you've been at this for 6+ years, as you claim, the "early days" are over. ARM went
from a blank slate to working silicon in about 17 months ...
> The classic entrypoint for a disruptor (which we are) is the low end of the market.
I'm pretty skeptical. The low end these days is Cortex-M0 chips for < $1, with a mature
toolchain and excellent code density. Above that are embedded DSP/motor controller chips.
And then there are mobile phone/tablet chips, but the Mill looks like a terrible architecture to
run a branchy JVM, and there's already a profusion of good ARM implementations, plus Intel's
heavily-subisidized x86's, so that's an awfully unattractive market for a new architecture
with immature tools and (initially) low volume.
> You seem to have mistaken limited resources for lack of interest :-) We are putting
> all we can into the tool chain. The software model of the Mill has been running in sim
> for six years now, although the Mill of six years ago is not the Mill of today.
Running *what* in sim ? What kind of benchmarks, and hand-coded or compiled by what compiler ?
And with what kind of simulated memory system ?
This doesn't seem like a general-purpose architecture, at all. It doesn't really seem to expect
to do procedure call/return very often - and it sure as heck doesn't seem to expect to switch
between threads in a hurry. So what's it for ?
> As for my goals: PDing might get me glory or tenure, but I'm not an academic and already
> have all the glory I'll ever want. The Mill is a commercial product. Yes, hundreds (not
> thousands) of people, and deep pockets ($120M estimated) are needed. The early days of
> design are best done with a small team - but why assume we intend to stay that way?
If you've been at this for 6+ years, as you claim, the "early days" are over. ARM went
from a blank slate to working silicon in about 17 months ...
> The classic entrypoint for a disruptor (which we are) is the low end of the market.
I'm pretty skeptical. The low end these days is Cortex-M0 chips for < $1, with a mature
toolchain and excellent code density. Above that are embedded DSP/motor controller chips.
And then there are mobile phone/tablet chips, but the Mill looks like a terrible architecture to
run a branchy JVM, and there's already a profusion of good ARM implementations, plus Intel's
heavily-subisidized x86's, so that's an awfully unattractive market for a new architecture
with immature tools and (initially) low volume.
