The Mill CPU Architecture – The Compiler (10 of 13)

Channel:Mill Computing, Inc. · Length:2:02:17 · Published:2015-07-09 · Views:11577

Watch on YouTube

Summary

A long Mill Computing lecture on the Mill CPU's compiler toolchain, covering the specializer that targets family members, scheduling and latency passes, and the belt. It explains how the compiler handles a wide-issue, exposed-pipeline design.

Why HN recommends it

Cited as the authoritative source on how Mill's VLIW-style compiler specializes code and differs from Itanium.

have a design talk that covers these concepts wonderfully in the general form — HN, 2023

Compilers, interpreters & program analysis · Computer architecture & CPUs

Subtitles (excerpt)

the talk tonight is on compilation and how it applies to the mill architecture um the mill is a new general purpose architecture that we've had a fair amount of public uh presentations about um this is essentially talking about how do we compile code for a conventional language for this architecture because it's different enough that compilation techniques are different to F see how what level I should be presenting at how many people have already seen some of our videos who previously attended some of these talks hands up oh probably a little more than half um well you'll get a bit of a repetition for you folks who are already familiar with it but I hope to um touch on the points which are relevant to compilation uh in the course of the talk there's an awful lot more to it than these points and for those you can go to our website and see how it really works it's one of a series of talks all of those prior ones are ones which have been already done at various venues um uh Stamford um Google um many others including this venue and uh we're getting down to compilation technology IES a caution those of you who have ever written the guts at the back end of a compiler will recognize that I am lying

On Hacker News

The (disconcertingly silent) Mill CPU company have a design talk that covers these concepts wonderfully in the general form, while explaining their various innovations. e.g. this video describes how their VLIW compiler does partial compilation/'specialisation' to allow the same input to be consumed on machines with different variations of hardware features. [VIDEO] — anotherhue, 2023-01
It's been too long since I watched the videos, so unfortunately I don't remember the specifics. For reference, here's[1] the relevant one on the compiler aspect. I do seem to recall that they seemingly had studied the failures of Itanium, and supposedly designed their architecture to not fall into the same pitfalls as with the EPIC/Itanium. One aspect I recall is that while they have VLIW, differ… — magicalhippo, 2019-02
The answer at 1:26:12 [1] about the belt being in fact a 'proxy' of the bypass network is very interesting. [1] [VIDEO] — fla, 2015-07

Related videos