rr VS RecursiveFactorization.jl

I think this tool must share a lot techniques and use cases with rr. I wonder how it compares in various aspects.

https://rr-project.org/

rr "sells" as a "reversible debugger", but it obviously needs the determinism for its record and replay to work, and AFAIK it employs similar techniques regarding system call interception and serializing on a single CPU. The reversible debugger aspect is built on periodic snapshotting on top of it and replaying from those snapshots, AFAIK. They package it in a gdb compatible interface.

Hermit also lists record/replay as a motivation, although it doesn't list reversible debugging in general.

Interesting. Perhaps you can inspect the disassembly of the function in question when using Graal and HotSpot. It is likely related to that.

Another debugging technique we use for heisenbugs is to see if `rr` [1] can reproduce it. If it can then that's great as it allows you to go back in time to debug what may have caused the bug. But `rr` is often not great for concurrency bugs since it emulates a single-core machine. Though debugging a VM is generally a nightmare. What we desperately need is a debugger that can debug both the VM and the language running on top of it. Usually it's one or the other.

> In general I’d argue you haven’t fixed a bug unless you understand why it happened and why your fix worked, which makes this frustrating, since every indication is that the bug exists within proprietary code that is out of my reach.

Were you using Oracle GraalVM? GraalVM community edition is open source, so maybe it's worth checking if it is reproducible in that.

[1]: https://github.com/rr-debugger/rr

https://rr-project.org/ had the same problem. They use the retired conditional branch counter instead of instruction counter, and then instruction steeping until at the correct address.

That'll work great for your Distributed QSort Incorporated startup, where the only product is a sorting algorithm.

Formal software verification is very useful. But what can be usefully formalized is rather limited, and what can be formalized correctly in practice is even more limited. That means you need to restrict your scope to something sane and useful. As a result, in the real world running thousands of tests is practically useful. (Well, it depends on what those tests are; it's easy to write 1000s of tests that either test the same thing, or only test the things that will pass and not the things that would fail.) They are especially useful if running in a mode where the unexpected happens often, as it sounds like this system can do. (It's reminiscent of rr's chaos mode -- https://rr-project.org/ linking to https://robert.ocallahan.org/2016/02/introducing-rr-chaos-mo... )

The approach of check-pointing computation such that it is resumable and restartable sounds similar to a time-traveling debugger, like rr or WinDbg:

https://rr-project.org/

https://learn.microsoft.com/windows-hardware/drivers/debugge...

https://github.com/rr-debugger/rr?tab=readme-ov-file#system-...

What about the other benchmarks on the same site? https://docs.sciml.ai/SciMLBenchmarksOutput/stable/Bio/BCR/ BCR takes about a hundred seconds and is pretty indicative of systems biological models, coming from 1122 ODEs with 24388 terms that describe a stiff chemical reaction network modeling the BCR signaling network from Barua et al. Or the discrete diffusion models https://docs.sciml.ai/SciMLBenchmarksOutput/stable/Jumps/Dif... which are the justification behind the claims in https://www.biorxiv.org/content/10.1101/2022.07.30.502135v1 that the O(1) scaling methods scale better than O(log n) scaling for large enough models? I mean.

> If you use special routines (BLAS/LAPACK, ...), use them everywhere as the respective community does.

It tests with and with BLAS/LAPACK (which isn't always helpful, which of course you'd see from the benchmarks if you read them). One of the key differences of course though is that there are some pure Julia tools like https://github.com/JuliaLinearAlgebra/RecursiveFactorization... which outperform the respective OpenBLAS/MKL equivalent in many scenarios, and that's one noted factor for the performance boost (and is not trivial to wrap into the interface of the other solvers, so it's not done). There are other benchmarks showing that it's not apples to apples and is instead conservative in many cases, for example https://github.com/SciML/SciPyDiffEq.jl#measuring-overhead showing the SciPyDiffEq handling with the Julia JIT optimizations gives a lower overhead than direct SciPy+Numba, so we use the lower overhead numbers in https://docs.sciml.ai/SciMLBenchmarksOutput/stable/MultiLang....

> you must compile/write whole programs in each of the respective languages to enable full compiler/interpreter optimizations

You do realize that a .so has lower overhead to call from a JIT compiled language than from a static compiled language like C because you can optimize away some of the bindings at the runtime right? https://github.com/dyu/ffi-overhead is a measurement of that, and you see LuaJIT and Julia as faster than C and Fortran here. This shouldn't be surprising because it's pretty clear how that works?

I mean yes, someone can always ask for more benchmarks, but now we have a site that's auto updating tons and tons of ODE benchmarks with ODE systems ranging from size 2 to the thousands, with as many things as we can wrap in as many scenarios as we can wrap. And we don't even "win" all of our benchmarks because unlike for you, these benchmarks aren't for winning but for tracking development (somehow for Hacker News folks they ignore the utility part and go straight to language wars...).

If you have a concrete change you think can improve the benchmarks, then please share it at https://github.com/SciML/SciMLBenchmarks.jl. We'll be happy to make and maintain another.

Ask them to download Julia and try it, and file an issue if it is not fast enough. We try to have the latest available.

See for example: https://github.com/JuliaLinearAlgebra/RecursiveFactorization...

Julia defaults to OpenBLAS but libblastrampoline makes it so that `using MKL` flips it to MKL on the fly. See the JuliaCon video for more details on that (https://www.youtube.com/watch?v=t6hptekOR7s). The recursive comparison is against OpenBLAS/LAPACK and MKL, see this PR for some (older) details: https://github.com/YingboMa/RecursiveFactorization.jl/pull/2... . What it really comes down to in the end is that OpenBLAS is rather bad, and MKL is optimized for Intel CPUs but not for AMD CPUs, so when the best CPUs are now all AMD CPUs, having a new set of BLAS tools and mixing that with recursive LAPACK tools is either as good or better on most modern systems. Then we see this in practice even when we build BLAS into Sundials for 1,000 ODE chemical reaction networks (https://benchmarks.sciml.ai/html/Bio/BCR.html).

>I hope those benchmarks are coming in hot

M1 is extremely good for PDEs because of its large cache lines.

https://github.com/SciML/DiffEqOperators.jl/issues/407#issue...

The JuliaSIMD tools which are internally used for BLAS instead of OpenBLAS and MKL (because they tend to outperform standard BLAS's for the operations we use https://github.com/YingboMa/RecursiveFactorization.jl/pull/2...) also generate good code for M1, so that was giving us some powerful use cases right off the bat even before the heroics allowed C/Fortran compilers to fully work on M1.

Not necessarily true with Julia. Many libraries like DifferentialEquations.jl are Julia all of the way down because the pure Julia BLAS tools outperform OpenBLAS and MKL in certain areas. For example see:

https://github.com/YingboMa/RecursiveFactorization.jl/pull/2...

So a stiff ODE solve is pure Julia, LU-factorizations and all.

Also, the major point is that BLAS has little to no role played here. Algorithms which just hit BLAS are very suboptimal already. There's a tearing step which reduces the problem to many subproblems which is then more optimally handled by pure Julia numerical linear algebra libraries which greatly outperform OpenBLAS in the regime they are in:

https://github.com/YingboMa/RecursiveFactorization.jl#perfor...

And there are hooks in the differential equation solvers to not use OpenBLAS in many cases for this reason:

https://github.com/SciML/DiffEqBase.jl/blob/master/src/linea...

Instead what this comes out to is more of a deconstructed KLU, except instead of parsing to a single sparse linear solve you can do semi-independent nonlinear solves which are then spawning parallel jobs of small semi-dense linear solves which are handled by these pure Julia linear algebra libraries.

And that's only a small fraction of the details. But at the end of the day, if someone is thinking "BLAS", they are already about an order of magnitude behind on speed. The algorithms to do this effectively are much more complex than that.