Announcement

Collapse
No announcement yet.

Ubuntu 21.10 Performance Continues In The Right Direction For AArch64

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Ubuntu 21.10 Performance Continues In The Right Direction For AArch64

    Phoronix: Ubuntu 21.10 Performance Continues In The Right Direction For AArch64

    As a good sign ahead of the important Ubuntu 22.04 LTS release in the spring, Ubuntu 21.10 further ups the 64-bit ARM (AArch64) performance. Here is a look at some of the gains in going from Ubuntu 21.04 to the recently released Ubuntu 21.10.

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Love when the benchmarks are all consistent.

    Side note: Has anyone tried the new KFENCE?

    With kernel 5.13, Ubuntu 21.10 adds support for Kernel Electric Fence (KFENCE), a new run-time memory error detector designed for production environments. KFENCE keeps the overhead low whilst detecting the most common memory errors. Enabled by default, Ubuntu 21.10 will randomise the memory location of the kernel stack at each system-call entry on both the amd64 and arm64 architectures with a minimum impact on performance.
    Source: Ubuntu Blog

    Kernel Electric-Fence (KFENCE) is a low-overhead sampling-based memory safety error detector. KFENCE detects heap out-of-bounds access, use-after-free, and invalid-free errors.

    KFENCE is designed to be enabled in production kernels, and has near zero performance overhead. Compared to KASAN, KFENCE trades performance for precision. The main motivation behind KFENCE’s design, is that with enough total uptime KFENCE will detect bugs in code paths not typically exercised by non-production test workloads. One way to quickly achieve a large enough total uptime is when the tool is deployed across a large fleet of machines.
    Source: docs.kernel.org

    Comment


    • #3
      Originally posted by perpetually high View Post
      Love when the benchmarks are all consistent.
      Same - I'm always wary of overall improvements when I see something has regressed, but when all tests across the board have improved, that's a great sign. Of course, this could potentially mean higher power consumption, but so long as it's proportionate, that isn't a bad thing. For ARM, a 5% rise in performance is fine if it means <=5% increase in wattage.

      Comment


      • #4
        Originally posted by schmidtbag View Post
        Same - I'm always wary of overall improvements when I see something has regressed, but when all tests across the board have improved, that's a great sign. Of course, this could potentially mean higher power consumption, but so long as it's proportionate, that isn't a bad thing. For ARM, a 5% rise in performance is fine if it means <=5% increase in wattage.
        5% overall is quite a large gain indeed, I guess from compiler optimization and better vectorization. Generally performance improvements from optimization improve perf/Watt since you execute fewer instructions and finish a task faster.

        Comment

        Working...
        X