Mastodawn

Considering how efficient Apple A*/M* CPUs are, do you think x86 is dead? Or it's just Apple's optimization that makes them work so efficient?
I'm using an Intel 12th gen and I hope ARM CPUs (or #RiscV 🙃 ) will be available for Linux in the next few years.

#Apple #x86 #x86_64 #Mprocessors #m5 #macbook #macbookNeo

N-gated Hacker News Mar 9

🚀🎉 #JSLinux now supports #x86_64, because running outdated operating systems in your #browser is exactly what everyone was asking for! 🤔👌 Experience the thrill of Alpine Linux and the #nostalgia of Windows 2000 in glorious console mode, because who needs modernity when you can have retro-tech headaches instead? 🤓🔧
https://bellard.org/jslinux/ #retro #computing #tech #innovation #HackerNews #ngated

JSLinux

Hacker News Mar 9

JSLinux Now Supports x86_64

https://bellard.org/jslinux/

#HackerNews #JSLinux #x86_64 #LinuxDevelopment #TechNews #OpenSource #Emulator

JSLinux

openSUSE Linux Feb 18

#openSUSE Leap 16.0 public cloud images are now available for both #x86_64 and #Arm64. Launch it today via Compute Engine and deploy your workloads on a stable #Linux distribution. https://sysadmin-journal.com/opensuse-leap-16-0-is-now-available-on-google-cloud-platform/

openSUSE Leap 16.0 is now available on Google Cloud Platform

openSUSE Leap 16.0 is now available on Google Cloud Platform with x86_64 and Arm64 images. You can launch a Compute Engine instance by selecting openSUSE Leap 16.0 under Public images. The Cloud Observability Ops Agent is not yet supported.

SysAdmin Journal

Ammar Faizi Feb 15

https://godbolt.org/z/nn967Wdb9

Nova Lake is live on Godbolt. We are still early. I took the latest Clang/GCC trunk for a spin on Godbolt, and -march=novalake is officially here.

If the rumors hold true, we are looking at the single biggest architectural jump in Intel's history.

With the arrival of Intel APX, the x86-64 ISA is finally getting the modernization it deserves. Several notable things are:

- 16 new GPRs (%r16 - %r31).
- NDD (New Data Destination): 3-operand instructions help stop wasting cycles on mov to preserve src/dst registers.
- push2/pop2: Push and pop two registers at once.
- jmpabs: 64-bit jumps.
- NF Suppression: Math ops without touching %rflags.
- ccmp: Conditional comparisons, branchless code.

Additionally, with the AVX10 support, Intel will officially restore AVX-512 capabilities to its consumer lineup for the first time since Tiger Lake in 2020.

Both P-core and E-core will support AVX-512.

#x86 #x86_64 #apx #llvm #clang #gcc

Ammar Nofan Faizi Feb 13

LLVM Bug: x86-64 inline assembly fails with "symbol is already defined" for local labels under optimization (-O1+)

When compiling x86-64 inline assembly that contains explicitly named local labels (e.g., `.Lcoro_resume_task_ret_addr`), Clang throws a "symbol is already defined" error if optimizations (-O1, -O2, -O3) are enabled. The exact same code compiles successfully at -O0 in Clang, and works perfectly across all optimization levels in GCC.

https://github.com/llvm/llvm-project/issues/181310

#llvm #x86_64 #x86

Bug: x86-64 inline assembly fails with "symbol is already defined" for local labels under optimization (-O1+) · Issue #181310 · llvm/llvm-project

Ammar Faizi - 2026-02-13 Summary When compiling x86-64 inline assembly that contains explicitly named local labels (e.g., .Lcoro_resume_task_ret_addr), Clang throws a "symbol is already defined" er...

GitHub

T4$Jan 23

#computers #laptops #linux #x86_64 Is it wrong that I am using the bottom half of a ultra thin laptop/notebook (I hate the term notebook), without the original monitor because it still works, and I've named it "Slabby"? 🤣

Habr Jan 16

Библиотека JIT-компиляции Loops для оптимизации нейросетей, СУБД и не только

Привет, Хабр! Меня зовут Пётр Чекмарёв, я старший инженер компании YADRO, занимаюсь компьютерным зрением на мобильных устройствах и низкоуровневой оптимизацией плотных вычислительных функций. Оптимизация кода — вечная тема, особенно актуальная в дни триумфального шествия искусственного интеллекта. Оптимально написанные, но изолированные ядра сетей составляются в разные последовательности в зависимости от архитектуры модели. Однако, если дать им информацию друг о друге во время компиляции, сеть удастся заметно ускорить. Выгружать программу для перекомпиляции, будь она движком инференса или СУБД — бессмысленно, поэтому компилировать надо во время работы, Just-In-Time. В предыдущей статье AI-дивизиона YADRO Илья Знаменский рассказывал про JIT на базе Xbyak. В продолжении темы, я расскажу про пет-проект векторной JIT-кодогенерации, который я веду, и покажу, как она может помогать в оптимизации.

https://habr.com/ru/companies/yadro/articles/985422/

#jitкомпилятор #riscv #arm #arm64 #x86_64 #компиляторы

Библиотека JIT-компиляции Loops для оптимизации нейросетей, СУБД и не только

Loop fusion is a compiler transformation in which two adjacent loops are merged into a single loop over the same index range. This transformation is typically applied to reduce loop overhead and...

Хабр