chapel VS tigerbeetle

Thanks, @subharmonicon!

While Chapel can run on many different systems, the main goal is making HPC programming much easier. Therefore, we are currently focusing on hardware that you can find in HPC systems (NVIDIA, AMD and Intel). Metal doesn't fall into that category, unfortunately. So far, the name came up infrequently in our discussions IIRC (especially targetting SPIRV), but we haven't heard from any [potential] user who may be interested in it. I would encourage you or anybody else interested in it to create an issue asking for the feature: https://github.com/chapel-lang/chapel/issues/new. Seeing public interest in that direction can change our prioritization.

One thing that I wanted to add that's not in the blogpost is the "cpu-as-device" mode. With that mode, you can use any machine, even one without a GPU, to write applications using Chapel's GPU features. That mode is for those who want to do initial development/debugging on their personal laptops before putting their application on an HPC system. In other words, while you can't use Metal directly, you can still write GPU-enabled applications in your Mac using Chapel, if the end goal is to run it on an HPC system. More details on cpu-as-device: https://chapel-lang.org/docs/main/technotes/gpu.html#cpu-as-...

Agreed. Here is a serious contender[0] minus all the hype and the $100M in VC money. You would expect a minimum of interest given how Mojo is received by the community, but not really in practice.

[0]: https://chapel-lang.org/

Cray is pushing their own language as well, Chapel.

https://chapel-lang.org/

As for Julia on Cray,

"Julia — The Newest Petaflop Family Language We Have Started to Love"

https://www.avenga.com/magazine/julia-programming-language

> Julia is one of the few languages that are in the so-called PetaFlop family; the other languages are C, C++ and Fortrant. It achieved 1.54 petaflops with 1.3 million threads on the Cray XC40 supercomputer.

https://chapel-lang.org if possible, Nix was also recently mentioned in Chapel Workshop https://chapel-lang.org/CHIUW2023.html https://github.com/twesterhout/nix-chapel

See Val for a possible step into that direction.

https://www.val-lang.dev/

Or how the Chapel language for HPC is going at it,

https://chapel-lang.org/

I suggest posting to a PLT focused resource, such as http://lambda-the-ultimate.org/

That said, a bit confused about the languages you reference in this context (Python, C#, JS) - didn't see any mention here or at your github repo of languages (some relatively ancient) in this space designed.

Sandia: Programming Languages for HPC [high performance computing] - is there life after MPI?

https://www.sandia.gov/app/uploads/sites/179/2022/04/SOS10-T...

Chapel:

https://chapel-lang.org/

https://en.wikipedia.org/wiki/Category:Array_programming_lan...

We needed the implicit conversion to `uint` in order for the overload resolution rules to make reasonable choices when faced with binary overloads for all of the numeric types. The document I linked talks through the examples. The case we were facing is something that we shared with `C#` -- in `C#` terms, if I make overloads for `f` for all numeric types (see https://github.com/chapel-lang/chapel/blob/main/test/types/coerce/allNumericsBinary.cs if you want to know exactly what I am talking about), then `f( myInt, myUlong )` runs `f(float, float)` which makes no sense. Especially if you care about numerical accuracy or program performance.

Code | Blog Walkthrough

Indeed!

I was so glad to see you cite not only the Rebello paper but also Protocol-Aware Recovery for Consensus-Based Storage. When I read your first comment, I was about to reply to mention PAR, and then saw you had saved me the trouble!

UW-Madison are truly the vanguard where consensus hits the disk.

We implemented Protocol-Aware Recovery for TigerBeetle [1], and I did a talk recently at the Recurse Center diving into PAR, talking about the intersection of global consensus protocol and local storage engine. It's called Let's Remix Distributed Database Design! [2] and owes the big ideas to UW-Madison.

[1] https://github.com/coilhq/tigerbeetle

[2] https://www.youtube.com/watch?v=rNmZZLant9o

> It sounds like payments might be part of the larger concept of declarative programming (DP)

Yes, exactly! The idea with TigerBeetle's state machine [1] is to expose double-entry accounting as higher level financial primitives, so that developers can think in terms of declaring transfers from one account to another. The business logic behind the scenes is detailed, but the interfaces and data structures are simple.

[1] https://github.com/coilhq/tigerbeetle/blob/main/src/state_ma...

> Maybe TigerBeetle could be generalized to support any multi-step distributed process?

That's part of the plan, that the distributed database framework of TigerBeetle can be used as a ”distributed Iron Man suit” to support any kind of state machine.

It's a pleasure. Let me know if you have any more questions about TigerBeetle. Our design doc is also here: https://github.com/coilhq/tigerbeetle/blob/main/docs/DESIGN....

Surprisingly, some of the most powerful distributed systems algorithms or tools are actually deterministic. They're powerful because they can "load the dice" and so make the distributed system more intuitive for humans to reason about, more resilient to real world network faults, and do all this with more performance.

For example, Barbara Liskov and James Cowling's deterministic view change [1], which isn't plagued by the latency issues of RAFT's randomized dueling leader problem. Viewstamped Replication Revisited's deterministic view change can react to a failed primary much quicker than RAFT (heartbeat timeouts don't require randomized "padding" as they do in RAFT), commence the leader election, and also ensure that the leader election succeeds without a split vote.

Determinism makes all that possible.

Deterministic testing [2][3] is also your best friend when it comes to testing distributed systems.

[1] I did a talk on VSR, including the benefits of the view change — https://www.youtube.com/watch?v=Wii1LX_ltIs

[2] FoundationDB are pioneers of deterministic testing — https://www.youtube.com/watch?v=OJb8A6h9jQQ

[3] TigerBeetle's deterministic simulation tests — https://github.com/coilhq/tigerbeetle#simulation-tests

This is the chasm problem, where people don't use a technology because people aren't using that technology, thus the technology has difficulty gaining adoption. I did see that Zig does have it's own killer app and startup that's using Zig: TigerBeattle.

Control flow statements should always be on their own lines, then it's easy to find all of them by visually scanning top-down, without needing to look all the way down each line.

[1]: https://github.com/coilhq/tigerbeetle/blob/main/src/vsr/repl...

> In hindsight, data logic should be in the database itself.

This is the reason we are creating TigerBeetle [1] at Coil, as an open source distributed financial accounting database, with the double entry logic and financial invariants enforced through financial primitives within the database itself.

This is all the more critical for financial data, because raw data consistency is not enough for financial transactions, you also need financial consistency, not to mention immutability.

The performance of doing it this way is also easier. For example, around a million financial transactions per second on commodity hardware, with p100 latency around 10-20ms.

[1] https://github.com/coilhq/tigerbeetle

You probably already know this — because we’ve mentioned it a few times — but Coil champions and supports open-source projects and is privacy-first, by default. Over the years, Developer Relations at Coil has championed and sponsored teams that write Open Web Documentations and projects that empower open-source developers to get paid. Coil has also incubated many open-source projects like Tigerbeetle and Rafiki.

[6] Partial logical sector reads/writes even when using O_DIRECT — https://github.com/coilhq/tigerbeetle/blob/main/src/storage....