#Optimizations

N-gated Hacker Newsngate
2025-10-03

πŸ“šπŸ”¬ Behold! Another riveting tale of matrix multiplication that promises to make your brain cells do backflips. πŸ€Έβ€β™‚οΈπŸŽ‰ Multithreaded that require you to sacrifice your first-born to hyperparameters just to squeeze out a few extra bytes of performance. βš™οΈπŸ› οΈ And if you want the actual code, here's a hint: .c. Happy debugging! πŸ–₯️πŸ’₯
salykova.github.io/gemm-cpu

2025-06-16

Features deployed for apply.coop!

- Main job listing paginates 25 jobs at time for faster page loads
- Atom/RSS/JSON item titles have working location, pay range, and workplace info
- Jobs/Workplaces sitemaps to help find jobs easier via search engines

#applycoop #Jobs #RSS #Optimizations #SEO

Johnny's Software Labjohnnysswlab@mastodon.online
2025-05-31

Does it matter if we are compiling with optimizations off (O0) or optimizations on (O3) if the problem is memory bound? Let’s find out…

#optimizations #performance #instructionlevelparallelism #ilp #compiler #gcc #memorybound

johnnysswlab.com/an-optimizing

2025-05-09

Π§Π°ΡΡ‚ΡŒ 3. ВСкторизация Π½Π° Go: CGo, Ρ‚Ρ€Π°Π½Π·Π°ΠΊΡ†ΠΈΠΈ, компиляторы, ΠΏΠΎΠ΄Π΄Π΅Ρ€ΠΆΠΊΠ°, Π±Π°ΠΉΡ‚ΠΎΠ²Ρ‹Π΅ инструкции

Π’ ΠΏΠ΅Ρ€Π²ΠΎΠΉ части ΡΡ‚Π°Ρ‚ΡŒΠΈ ΠΌΡ‹ рассмотрСли , ΠΊΠ°ΠΊ ΠΌΠΎΠΆΠ½ΠΎ Π²Ρ€ΡƒΡ‡Π½ΡƒΡŽ ΡƒΡΠΊΠΎΡ€ΠΈΡ‚ΡŒ Go-ΠΊΠΎΠ΄ с ΠΏΠΎΠΌΠΎΡ‰ΡŒΡŽ Π²Π΅ΠΊΡ‚ΠΎΡ€ΠΈΠ·Π°Ρ†ΠΈΠΈ ΠΈ SIMD-инструкций, Ρ€Π΅Π°Π»ΠΈΠ·ΠΎΠ²Π°Π½Π½Ρ‹Ρ… Ρ‡Π΅Ρ€Π΅Π· Go-ассСмблСр. Написали ΠΏΡ€ΠΎΡΡ‚ΡƒΡŽ, Π½ΠΎ ΠΏΠΎΠΊΠ°Π·Π°Ρ‚Π΅Π»ΡŒΠ½ΠΎ Π±Ρ‹ΡΡ‚Ρ€ΡƒΡŽ Ρ€Π΅Π°Π»ΠΈΠ·Π°Ρ†ΠΈΡŽ sliceContains ΠΈ ΡƒΠ²ΠΈΠ΄Π΅Π»ΠΈ, Ρ‡Ρ‚ΠΎ Π΄Π°ΠΆΠ΅ базовая вСкторизация ΠΌΠΎΠΆΠ΅Ρ‚ Π΄Π°Ρ‚ΡŒ ускорСниС Π² 10–14 Ρ€Π°Π· ΠΏΠΎ ΡΡ€Π°Π²Π½Π΅Π½ΠΈΡŽ со стандартной Ρ€Π΅Π°Π»ΠΈΠ·Π°Ρ†ΠΈΠ΅ΠΉ. Π’ΠΎ Π²Ρ‚ΠΎΡ€ΠΎΠΉ части ΡΡ‚Π°Ρ‚ΡŒΠΈ ΠΏΠΎΠ³Ρ€ΡƒΠ·ΠΈΠ»ΠΈΡΡŒ Π² практичСскоС ΠΏΡ€ΠΈΠΌΠ΅Π½Π΅Π½ΠΈΠ΅ SIMD Π² Go-ассСмблСрС, Ρ€Π΅Π°Π»ΠΈΠ·ΠΎΠ²Π°Π»ΠΈ Ρ„ΡƒΠ½ΠΊΡ†ΠΈΡŽ SliceContainsV1 ΠΈ ΠΈΠ·ΡƒΡ‡ΠΈΠ»ΠΈ, ΠΊΠ°ΠΊ с ΠΏΠΎΠΌΠΎΡ‰ΡŒΡŽ VADD, VDUP ΠΈ Π΄Ρ€ΡƒΠ³ΠΈΡ… инструкций ΠΌΠΎΠΆΠ½ΠΎ Π΄ΠΎΠ±ΠΈΡ‚ΡŒΡΡ 10–14-ΠΊΡ€Π°Ρ‚Π½ΠΎΠ³ΠΎ ускорСния простых Π·Π°Π΄Π°Ρ‡. Но возмоТности ΠΎΠΏΡ‚ΠΈΠΌΠΈΠ·Π°Ρ†ΠΈΠΈ Go-ΠΏΡ€ΠΎΠ³Ρ€Π°ΠΌΠΌ Π½Π° этом Π½Π΅ Π·Π°ΠΊΠ°Π½Ρ‡ΠΈΠ²Π°ΡŽΡ‚ΡΡ. Π’ этой части ΠΌΡ‹ ΠΏΠΎΠΉΠ΄Ρ‘ΠΌ дальшС: рассмотрим Π΄Ρ€ΡƒΠ³ΠΈΠ΅ Ρ‚Π΅Ρ…Π½ΠΈΠΊΠΈ Π½ΠΈΠ·ΠΊΠΎΡƒΡ€ΠΎΠ²Π½Π΅Π²ΠΎΠΉ ΠΎΠΏΡ‚ΠΈΠΌΠΈΠ·Π°Ρ†ΠΈΠΈ β€” ΠΎΡ‚ использования C-ΠΊΠΎΠ΄Π° ΠΈ Π°Π»ΡŒΡ‚Π΅Ρ€Π½Π°Ρ‚ΠΈΠ²Π½Ρ‹Ρ… компиляторов с ΠΏΠΎΠ΄Π΄Π΅Ρ€ΠΆΠΊΠΎΠΉ Π²Π΅ΠΊΡ‚ΠΎΡ€ΠΈΠ·Π°Ρ†ΠΈΠΈ Π΄ΠΎ Ρ€Π°Π±ΠΎΡ‚Ρ‹ с Π°ΠΏΠΏΠ°Ρ€Π°Ρ‚Π½Ρ‹ΠΌΠΈ транзакциями памяти Π½Π° Intel. ΠŸΠΎΠ³ΠΎΠ²ΠΎΡ€ΠΈΠΌ ΠΎ Ρ‚ΠΎΠΌ, ΠΊΠ°ΠΊ Π²Π½Π΅Π΄Ρ€ΡΡ‚ΡŒ ассСмблСр Π² ΠΏΡ€ΠΎΠ΄Π°ΠΊΡˆΠ΅Π½-ΠΊΠΎΠ΄, Π½Π΅ боясь Π·Π° Π΅Π³ΠΎ ΠΏΠΎΠ΄Π΄Π΅Ρ€ΠΆΠΊΡƒ, ΠΈ ΠΊΠ°ΠΊ ΠΎΠ±ΠΎΠΉΡ‚ΠΈ ограничСния стандартного Go-компилятора. ΠŸΡ€ΠΈΠ²Π΅Ρ‚, Π₯Π°Π±Ρ€! МСня Π·ΠΎΠ²ΡƒΡ‚ Π˜Π³ΠΎΡ€ΡŒ Панасюк, я Ρ€Π°Π±ΠΎΡ‚Π°ΡŽ Π² ЯндСкс, ΠΏΡ€Π΅ΠΏΠΎΠ΄Π°ΡŽ Π² ИВМО, Π° Ρ‚Π°ΠΊΠΆΠ΅ Π² свободноС врСмя Π²Ρ‹ΡΡ‚ΡƒΠΏΠ°ΡŽ Π½Π° конфСрСнциях, дСлюсь ΠΎΠΏΡ‹Ρ‚ΠΎΠΌ Π² соцсСтях ΠΈ помогаю Ρ€Π°Π·Π²ΠΈΡ‚ΠΈΡŽ Go-сообщСства, Π²Π΅Π΄Ρƒ Ρ‚Π΅Π»Π΅Π³Ρ€Π°ΠΌ-ΠΊΠ°Π½Π°Π» ΠΈ youtube-ΠΊΠ°Π½Π°Π». Если Π²Ρ‹ ΡƒΠΆΠ΅ Π·Π½Π°ΠΊΠΎΠΌΡ‹ с Π±Π°Π·ΠΎΠ²Ρ‹ΠΌΠΈ Ρ‚Π΅Ρ…Π½ΠΈΠΊΠ°ΠΌΠΈ Π²Π΅ΠΊΡ‚ΠΎΡ€ΠΈΠ·Π°Ρ†ΠΈΠΈ, эта Ρ‡Π°ΡΡ‚ΡŒ ΠΏΠΎΠΌΠΎΠΆΠ΅Ρ‚ Π³Π»ΡƒΠ±ΠΆΠ΅ ΠΏΠΎΠ½ΡΡ‚ΡŒ, ΠΊΠ°ΠΊ устроСны ΠΏΡ€ΠΎΠ΄Π²ΠΈΠ½ΡƒΡ‚Ρ‹Π΅ способы ускорСния Go-ΠΊΠΎΠ΄Π° ΠΈ Π½Π° Ρ‡Ρ‚ΠΎ стоит ΠΎΠ±Ρ€Π°Ρ‚ΠΈΡ‚ΡŒ Π²Π½ΠΈΠΌΠ°Π½ΠΈΠ΅ ΠΏΡ€ΠΈ Ρ€Π°Π±ΠΎΡ‚Π΅ с Π°Ρ€Ρ…ΠΈΡ‚Π΅ΠΊΡ‚ΡƒΡ€Π½ΠΎ-зависимыми оптимизациями.

habr.com/ru/companies/oleg-bun

#golang #optimizations #компилятор #gccgo #gollvm #ассСмблСр #word #byte #ΠΏΡ€ΠΎΠΈΠ·Π²ΠΎΠ΄ΠΈΡ‚Π΅Π»ΡŒΠ½ΠΎΡΡ‚ΡŒ

2025-04-30

#Database indexes in #Rails - from basic to advanced. πŸ“Š

Post with real migration & model code, you can adapt today!

#rubyonrails #performance #optimizations

jetthoughts.com/blog/turbochar

2025-03-15

Can we please make a new game category just called #UE5 Slop. Games which have no #optimizations they require you to use #upscaling for basic gameplay they force ugly Ray tracing on you. Inferior anti-aliasing choice. Big publisher games have honestly all been garbage. #game #rtx thanks, Nvidia

2025-01-09
πŸ‘€ Do you know? The Raspberry Pi 5 can run #Unvanquished in 1920Γ—1080 at more than 60fps! πŸš€

Here at 70~80fps with #Mesa 25.
We even have some model #optimizations to come!

#indiedev #gamedev #indiegame #floss #idtech #raspberrypi5 #arm #opengl #gameengines #daemon
Platform 23 spectator scene.Platform 23 alien base scene.Platform 23 human base scene.
AndrΓ©s Álvarez Iglesias πŸ“šβœοΈAndresalvareziglesias
2024-10-23

Since Python 3.13, the famous Python GIL is optional. This is a long requested functionality but, is the performance gain of threaded code real?

Let's test the performance of a threaded application with and without the GIL.

open.substack.com/pub/andresal

Kir4ik52 :blobfoxnerd:kir4ik52@mastodon.ml
2024-09-18

Power Options β€” это графичСскоС ΠΏΡ€ΠΈΠ»ΠΎΠΆΠ΅Π½ΠΈΠ΅ для управлСния энСргопотрСблСниСм Π½Π° Linux, ΠΏΡ€Π΅Π΄Π»Π°Π³Π°ΡŽΡ‰Π΅Π΅ ΠΊΠ°ΠΊ простыС, Ρ‚Π°ΠΊ ΠΈ Ρ€Π°ΡΡˆΠΈΡ€Π΅Π½Π½Ρ‹Π΅ Ρ„ΡƒΠ½ΠΊΡ†ΠΈΠΈ.

Оно создаСт ΠΎΠΏΡ‚ΠΈΠΌΠΈΠ·ΠΈΡ€ΠΎΠ²Π°Π½Π½Ρ‹Π΅ ΠΏΡ€ΠΎΡ„ΠΈΠ»ΠΈ ΠΏΡ€ΠΈ установкС, позволяя ΠΏΠΎΠ»ΡŒΠ·ΠΎΠ²Π°Ρ‚Π΅Π»ΡΠΌ ΡƒΠΏΡ€Π°Π²Π»ΡΡ‚ΡŒ настройками аккумулятора ΠΈ сСти с Π²ΠΎΠ·ΠΌΠΎΠΆΠ½ΠΎΡΡ‚ΡŒΡŽ настройки количСства ΠΏΡ€ΠΎΡ„ΠΈΠ»Π΅ΠΉ ΠΈ ΠΏΠ΅Ρ€Π΅ΠΎΠΏΡ€Π΅Π΄Π΅Π»Π΅Π½ΠΈΠΉ.
github.com/TheAlexDev23/power-
#poweroptions #linux #gnu_linux #battery #optimization #optimizations #github #energy

Zeeshan Ali Khan :rust: πŸ‡ΊπŸ‡¦zeenix@toot.cat
2024-09-05

bow to your king!

#optimizations #rustlang

Screenshot of benchmarks showing 99.419% improvement in performance
Shakthi Kannanshakthimaan
2024-08-29

Scheduling Model in LLVM Compiler myhsu.xyz/llvm-sched-model-1/ @llvmorg @risc_v

2024-07-06

'Multi-Objective Neural Architecture Search by Learning Search Space Partitions', by Yiyang Zhao, Linnan Wang, Tian Guo.

jmlr.org/papers/v25/23-1013.ht

#optimizers #optimizer #optimizations

Patrick Baumgartnerpatbaumgartner
2024-05-22

Had a great time giving my rockstar talk at . Thanks to the organizers, sponsors, speakers and attendees for creating such a great experience. Thanks for having me! ☺️ Looking forward to the third edition.

2024-02-09

"[Hash tables] need to be fast, meaning that they can quickly retrieve the necessary information. But they also need to be compact, using as little memory as possible. These twin objectives are fundamentally at odds.
[...]
Computer scientists have now mathematically proved that they have found the optimal trade-off."

quantamagazine.org/scientists-

#Mathematics #HashTables #Optimizations #ComputerScience

2024-01-28

Assessing the Impact of Compiler Optimizations on GPUs Reliability

#CUDA #Compilers #Optimizations

hgpu.org/?p=29032

JiΕ™Γ­ VanΔ›kvanek
2023-11-13

ŘeΕ‘Γ­te zrychlenΓ­ svΓ©ho webu a lepΕ‘Γ­ metriky? A napadl vΓ‘s Lazy load komentΓ‘Ε™Ε―? Co to je a jak funguje se dočtete ve člΓ‘nku.

blog.jirivanek.eu/cs/2023/11/1

Client Info

Server: https://mastodon.social
Version: 2025.07
Repository: https://github.com/cyevgeniy/lmst