Comment by cwzwarich
1 year ago
This is me! Didn’t expect to see this on here, but I’m looking forward to working with everyone else at the Lean FRO and the wider Lean community to help make Lean even better.
My background is in mathematics and I’ve had an interest in interactive theorem provers since before I was ever a professional software engineer, so it’s a bit of a dream come true to be able to pursue this full-time.
Thank you! My work laptop is a M4 Macbook Pro so I really appreciate the beauty of Rosetta. Thank you for the effort!
I just checked your LinkedIn and realized you joined Apple since 2009 (with one year of detour to Mozilla). You also graduated from Waterloo as a Pure Math Graduate student (I absolutely love Waterloo, the best Math/CS school IMO in my country - at the age of 40+ I'd go without doubt if they accept me).
May I ask, what is the path that leads you to the Rosetta 2 project? I even checked your graduate paper: ( https://uwspace.uwaterloo.ca/items/4bc518ca-a846-43ce-92f0-8... ), but it doesn't look like it's related to compiler theory.
(I myself studied Mathematics back in the day, but I was not a good student and I studied Statistics, which I joked that was NOT part of Mathematics, so I didn't take any serious Algebra classes and understand nothing of your paper)
> May I ask, what is the path that leads you to the Rosetta 2 project?
The member of senior management who was best poised to suggest who should work on it already knew me and thought I would be the best choice. Getting opportunities in large companies is a combination of nurturing relationships and luck.
Btw, followup question, and don't take this the wrong way at all, but what is impressive is someone with a mathematical background worked on something that seems to be one of the pinnacles of software engineering: a translator working at the binary level that creates executables interacting directly with the OS. Did you also double in CS back in school? Or did you pick up the knowledge afterwards? Yeah, it seems like a long list: operating systems, compilers, computer architecture, UNIX/MacOS systems-calls and internals...
... not to mention all the performance considerations and optimizations, also requiring a strong sense of algorithms and computational complexity. Wow!
Yeah, seems like most mathematicians (and physicists) I know who go into tech don't get past learning a couple of programming languages and don't have an interest in learning the depths of a how a computer works. Very impressive!
5 replies →
Thank you for the information! I'm sure your skills are well trusted.
Waterloo really is the best CS school in the world.
I have never been there, what do you consider to be its speciality comparing to say MIT and Berkeley?
2 replies →
Belgium, London, ABBA, Canada, or San Dimas? Why better than others?
1 reply →
Rosetta 2 is easily one of the most technically impressive things I've seen in my life. I've done some fairly intense work applying binary translation (DynamoRIO) and Rosetta 2 still feels totally magical to me.
Thanks. It means a lot coming from someone with experience in our niche field.
We're you really _the_ creator of Rosetta 2? How big was the team, what was your role in it?
Rosetta 2 is amazing, I'm genuinely surprised this is the work of just one person!
I was the only person working on it for ~2 years, and I wrote the majority of the code in the first version that shipped. That said, I’m definitely glad that I eventually found someone else (and later a whole team) to work on it with me, and it wouldn’t have been as successful without that.
When people think of a binary translator, they usually just think of the ISA aspects, as opposed to the complicated interactions with the OS etc. that can consume just as much (or even more) engineering effort overall.
As someone frustrated in a team of 10+ that is struggling to ship even seemingly trivial things due to processes and overheads and inefficiencies, I would really appreciate some insights on how do you organize the work to allow a single developer to achieve this.
How do you communicate with the rest of the organization? What is the lifecycle and release process like? Do you write requirements and specs for others (like validation or integration) to base their work on? Basically, what does the day to day work look like?
8 replies →
That's super impressive. I remember being astonished that the x86 executable of Python running through Rosetta 2 on my M1 was just a factor of 2 slower than the native version.
QEMU was something like a factor of 5-10x slower than native, IIRC.
5 replies →
That is fascinating that this amazing system was the work of largely one person. You mentioned that interacting with the OS was super difficult. What were the most enjoyable aspects of building Rosetta?
1 reply →
It's a shame that Apple's stated intent is to throw the project away after a while. Personally, I really hope it sticks around forever, though I'm not optimistic.
23 replies →
It is my experience that it is easier to create good quality things as an individual than as a team. Especially for the core of a product. Also look at Asahi.
However, to really finish/polish a product you need a larger group of people. To get the UI just right, to get the documentation right, to advocate the product, to support it.
It is easily possible to have 10 people working on the team and only having a single core person. Then find someone to act as product manager while as the core person you can focus on the core of the product while still setting the direction without having to chase all the other work.
It is possible, but not easy to set up in most organisations. You need a lot of individual credit/authority and/or the business case needs to be very evident.
Do you have book recommendations in regards to disassembly, syscalls, x86/64 assembler etc?
What do I need to know to be able to build something as advanced as rosetta?
I am assuming that you reimplemented the syscalls for each host/guest system as a reliable abstraction layer to test against. But so many things are way beyond my level of understanding.
Did you build your own assembler debugger? What kind of tools did you use along the way? Were reversing tools useful at all (like ghidra, binaryninja etc)?
"Virtual Machines: Versatile Platforms for Systems and Processes" by Jim Smith and Ravi Nair is a great book on the topic.
Thank you. Other than papers, I think this is one of the rare books that talk extensively about dynamic recompilation. I was hoping to learn more about the PPC M68K emulator (early version interpreter style and later version dynamic recompilation style) and definitely will read it.
1 reply →
Out of curiosity, if you’ve been at a FAANG since at least 2009, have you ever retired or taken a “sabbatical” for a year or two, since you would have made enough money to retire and live passively at amounts similar to annual compensation and taxed way better
Just curious how the decisions have formed, its totally fine if FAANG or specifically Apple was fulfilling for you, I also wonder if its financial fear to an irrational extent just because I see that on Blind a lot
I did go to Mozilla Research to work on Servo/Rust for a bit in 2015, which didn’t turn out to be the best decision.
I always assumed that I would stick around at Apple until some singular event that would motivate me to quit, and that would be it. I have been so lucky at Apple to have been in the right place at the right time for several projects: relatively early iPhone team, original iPad team, involved in the GCC -> Clang transition, involved in the 64-bit ARM transition, involved in early Apple Watch development, first engineer working full-time on the Apple silicon transition for the Mac, etc. Obviously I was doing something right if I kept getting these chances, but if I went to another FAANG I wouldn’t have the same history, and I don’t think it would be the same experience.
My projected path to parting ways with Apple didn’t really take place, since I’m now working at a non-profit dedicated to developing an interactive theorem prover and left Apple without any animosity in either direction.
It would be incredible if you could write a book someday about all those experiences. I would very happily buy that book.
Thanks for all that incredible work and your insights here.
Wow that's really a ton of memorable experience. I hope you write a book or some blog posts or do an interview.
sorry to hijack the discussion but do you see any chance of consolidating the theoretical framework of real numbers with practical calculations of floats? That is if I proof the correctness of some theorem for real numbers ideally I would just use that as the algorithm to compute things with floats.
also I was shocked to learn that the simple general comparison of (the equality of) two real numbers is not decidable, which is very logical if you think about it but an enormous hindrance for practical applications. Is there any work around for that?
You can use floats to accelerate interval arithmetic (which is "exact" in the sense of constructive real numbers) but that requires setting the correct rounding modes, and being aware of quirks in existing hardware floating point implementations, some of which may e.g. introduce non-exact outputs in several of the least significant digits, or even flush "small" (for unclear definitions of "small", not always restricted to FP-denormal numbers) results to zero.
Equality is not computable in the general case, but apartness can be stated exactly. For some practical cases, one may also be able to prove that two real numbers are indeed equal.
This is exciting!
Given your experience with Rosetta 2 and your deep understanding of code translation and optimization, what specific areas in Lean’s code generation pipeline do you see as ‘low-hanging fruit’ for improvement?
Additionally, which unique features or capabilities of Lean do you find most promising or exciting to leverage in pushing the boundaries of efficient and high-quality code generation?
What was the tipping point that made you want to work on Lean?
I don't think there was a single tipping point, just a growing accumulation of factors:
- the release of Lean 4 slightly over a year ago, which impressed me both as a proof assistant and a programming language
- the rapid progress in formalization of mathematics from 2017 onward, almost all of which was happening in Lean
- the growing relevance of formal reasoning in the wake of improvements in AI
- seeing Lean's potential (a lot of which is not yet realized) for SW verification (especially of SW itself written in Lean)
- the establishment of the Lean FRO at the right time, intersecting all of the above
How does Lean compares with Coq? (I am not familiar with Lean but am familiar with Coq)
2 replies →
Can Lean can do what TLA+ does - model check thorny concurrency problems ?
Surprised you didnt go into something AI adjacent
I don't know what his reasons are but it makes sense to me. Yes, there are incredible results coming out of the AI world but the methods aren't necessarily that interesting (i.e. intellectually stimulating) and it can be frustrating working in a field with this much noise.
I don't want to come across as too harsh but having studied machine learning since 2015 I find the most recent crop of people excited about working on AI are deep in Dunning-Kruger. I think I conflate this a bit with the fascination of results over process (I suppose that befuddlement is what led me to physics over engineering) but working in ML research for so long it's hard to gin up a perspective that these things are actually teleologically useful, and not just randomly good enough most of the time to keep up the illusion.
12 replies →
Lean is AI adjacent.
Only because the AI people find it interesting. It's not really AI in itself.
8 replies →
This is *VERY* AI-adjacent... the next batch of AI algos will need to integrate reasoning through theorem provers to go next level