4 ms·

I wonder how this compares to jemalloc, mimalloc, snmalloc?

pizlonator 4y ago
I never got a chance to compare it to those, since I was most interested in beating bmalloc. And I mainly wanted to beat it on Safari workloads. I believe bmalloc was previously compared against jemalloc and tcmalloc, also using Safari workloads, and bmalloc was significantly faster at the time.
- IshKebab 4y ago
  For other people who have never heard of bmalloc - it's a custom allocator used only by WebKit. I guess it's not surprising they added one since the Mac system allocator is extremely slow. A custom allocator is pretty much a free 20% speed up on Mac (depending on your workload) but I found they made no difference on Linux. Haven't tried on Windows.
  - pizlonator 4y ago
    Just some credit where credit is due. I measured the system malloc crushing all other mallocs on some workloads and they were the kind of workloads that some folks run every day. I measured the system malloc crushing most other mallocs on memory efficiency on most workloads. System malloc is really good at reusing memory and has very mature decommit policies that easily rival what I came up with. So there’s that.
    - meisel 4y ago
      What sort of workloads are those? I have yet to see a workload where macOS's system allocator is not substantially slower than alternatives like jemalloc.
    - pcwalton 4y ago
      macOS malloc was incredibly slow for Pathfinder, far behind every other OS. Everything became bottlenecked on it. It was a free 2x speedup if not more to switch to jemalloc. I suspect this is because Pathfinder's CPU portion is a multicore workload and macOS allocator performs poorly when under heavy multithreaded contention. It probably just isn't the kind of workload that macOS allocator was tuned for.
      - pizlonator 4y ago
        My understanding is that the system malloc is excellent under contention, but has a high baseline cost for every malloc/free call. I don't remember exactly what workloads it performed really great at (and if I did I dunno if I could say), but I do remember they were parallel, and the speed-ups got bigger the more cores you added. Everything else about your experience matches mine. Libpas is much faster than system malloc in WebKit and JSC. The difference isn't 2x on my preferred benchmarks (which are large and do lots of things that don't rely on malloc), but it is easily more than 2x on smaller benchmarks. So your 2x result sounds about right.
  - astrange 4y ago
    Alas, many people think performance engineering is only about the wall clock time of their program and not what impact it has on anything else. malloc performance highly depends on how large the allocations are though.
  - KerrAvon 4y ago
    What's the workload where you found that to be the case? The system allocator is heavily tuned for general use, and should be very difficult to beat -- generally.
    - IshKebab 4y ago
      It was a non-traditional compiler written in C++. > The system allocator is heavily tuned Perhaps but it probably has more constraints than a custom allocator too, e.g. backwards compatibility with software that unwittingly depends on its exact behaviour. I found a discussion of this where people reported various speedups (10%, 100%, 300%) on Mac by switching to a custom allocator: https://news.ycombinator.com/item?id=29068828 https://news.ycombinator.com/item?id=29068828 I'm sure there are better benchmarks if you Google it.
- kookamamie 4y ago
  tcmalloc and mimalloc are the ones I'd be interested seeing comparisons to. The former is the fastest in my case, which is essentially many-core HPC.
mjp41 4y ago
libpas is now in mimalloc-bench (https://github.com/daanx/mimalloc-bench https://github.com/daanx/mimalloc-bench). Thanks to Julien Voisin (https://dustri.org https://dustri.org), which makes about 20 allocators that can be compared on a collection of workloads.

I never got a chance to compare it to those, since I was most interested in beating bmalloc. And I mainly wanted to beat it on Safari workloads. I believe bmalloc was previously compared against jemalloc and tcmalloc, also using Safari workloads, and bmalloc was significantly faster at the time.

IshKebab 4y ago
For other people who have never heard of bmalloc - it's a custom allocator used only by WebKit. I guess it's not surprising they added one since the Mac system allocator is extremely slow. A custom allocator is pretty much a free 20% speed up on Mac (depending on your workload) but I found they made no difference on Linux. Haven't tried on Windows.
- pizlonator 4y ago
  Just some credit where credit is due. I measured the system malloc crushing all other mallocs on some workloads and they were the kind of workloads that some folks run every day. I measured the system malloc crushing most other mallocs on memory efficiency on most workloads. System malloc is really good at reusing memory and has very mature decommit policies that easily rival what I came up with. So there’s that.
  - meisel 4y ago
    What sort of workloads are those? I have yet to see a workload where macOS's system allocator is not substantially slower than alternatives like jemalloc.
  - pcwalton 4y ago
    macOS malloc was incredibly slow for Pathfinder, far behind every other OS. Everything became bottlenecked on it. It was a free 2x speedup if not more to switch to jemalloc. I suspect this is because Pathfinder's CPU portion is a multicore workload and macOS allocator performs poorly when under heavy multithreaded contention. It probably just isn't the kind of workload that macOS allocator was tuned for.
    - pizlonator 4y ago
      My understanding is that the system malloc is excellent under contention, but has a high baseline cost for every malloc/free call. I don't remember exactly what workloads it performed really great at (and if I did I dunno if I could say), but I do remember they were parallel, and the speed-ups got bigger the more cores you added. Everything else about your experience matches mine. Libpas is much faster than system malloc in WebKit and JSC. The difference isn't 2x on my preferred benchmarks (which are large and do lots of things that don't rely on malloc), but it is easily more than 2x on smaller benchmarks. So your 2x result sounds about right.
- astrange 4y ago
  Alas, many people think performance engineering is only about the wall clock time of their program and not what impact it has on anything else. malloc performance highly depends on how large the allocations are though.
- KerrAvon 4y ago
  What's the workload where you found that to be the case? The system allocator is heavily tuned for general use, and should be very difficult to beat -- generally.
  - IshKebab 4y ago
    It was a non-traditional compiler written in C++. > The system allocator is heavily tuned Perhaps but it probably has more constraints than a custom allocator too, e.g. backwards compatibility with software that unwittingly depends on its exact behaviour. I found a discussion of this where people reported various speedups (10%, 100%, 300%) on Mac by switching to a custom allocator: https://news.ycombinator.com/item?id=29068828 https://news.ycombinator.com/item?id=29068828 I'm sure there are better benchmarks if you Google it.
kookamamie 4y ago
tcmalloc and mimalloc are the ones I'd be interested seeing comparisons to. The former is the fastest in my case, which is essentially many-core HPC.

mjp41 4y ago

libpas is now in mimalloc-bench (https://github.com/daanx/mimalloc-bench https://github.com/daanx/mimalloc-bench). Thanks to Julien Voisin (https://dustri.org https://dustri.org), which makes about 20 allocators that can be compared on a collection of workloads.