cnx-software.com [Unofficial]2d ago

GEEKOM A7 2026 Edition Review – Part 2: AMD Ryzen 5 7545U Mini PC tested with Windows 11 Pro

https://fed.brid.gy/r/https://www.cnx-software.com/2026/06/22/geekom-a7-2026-edition-review-part-2-amd-ryzen-5-7545u-mini-pc-tested-with-windows-11-pro/

GEEKOM A7 2026 Edition Review – Part 2: AMD Ryzen 5 7545U Mini PC tested with Windows 11 Pro

We have already checked out the GEEKOM A7 2026 Edition mini PC's hardware with an unboxing and a teardown in the first part of the review, before quickly trying the pre-installed Windows 11 Pro. We've now had time to test the AMD Ryzen 5 7545U mini PC more thoroughly in Windows 11, and we'll report our experience, including feature testing, benchmark results, 4K and 8K YouTube video playback, 2.5GbE and WiFi 6E network performance evaluation, and cooling performance, as well as fan noise and power consumption measurements. Software Overview and Feature Testing – GEEKOM A7 2026 Edition The System > About window confirms we have an A7 computer equipped with a 3.2 GHz (base frequency) AMD Ryzen 5 7545U processor with Radeon 740M graphics, 16 GB of RAM, and 466 GB of storage preloaded with Windows 11 Pro 25H2. HWiNFO 64 provides additional details about the AMD Ryzen 5 7545U

CNX Software - Embedded Systems News
cnx-software.com [Unofficial]3d ago

SpacemiT K3 Pico-ITX RISC-V Chassis Kit Review – Part 2: What works, what doesn’t in Bianbu OS 4.0

https://fed.brid.gy/r/https://www.cnx-software.com/2026/06/21/spacemit-k3-pico-itx-risc-v-chassis-kit-review-part-2-what-works-what-doesnt-in-bianbu-os-4-0/

SpacemiT K3 Pico-ITX RISC-V Chassis Kit Review – Part 2: What works, what doesn’t in Bianbu OS 4.0

Last month, I received the SpacemiT K3 Pico-ITX Chassis Kit based on the company's K3 16-core RISC-V SoC, and started the review with an unboxing, a teardown, and a first boot to Bianbu OS 4.0. Since the system features a 10Gbps Ethernet SFP+ cage, I also had to order a 10GbE SFP+ to Copper adapter, as my 10GbE networking gear is exclusively based on RJ45 ports. In this review, I'll check system information in Bianbu OS 4.0.1, run a few benchmarks, test 10GbE, GbE, and WiFi 6 networking performance, play YouTube videos at various resolutions, run AI workloads (LLM), check all/most features work as expected, and measure the power consumption of the SpacemiT K3 "Pico-ITX Chassis Kit" mini PC. Bianbu OS 4.0 System Information Before running anything, I updated the system: 594 packages were updated. For reference, the update also involved updating the EC (Embedded Controller) firmware: Let's double-check the

CNX Software - Embedded Systems News
Benjamin Han4d ago

Just how capable are today's frontier AI models? Artificial Analysis condenses the answer into one number, its Intelligence Index, a weighted average of nine evaluations that counts agentic tasks the most. Closed models lead all three of its indices, but the comparison is black-box: the winners disclose neither size nor architecture, while the best open-weight model, GLM-5.2, is a fully documented 753B mixture of experts.

https://benjaminhan.net/posts/20260620-intelligence-benchmarking/?utm_source=mastodon&utm_medium=social

#LLMs #Benchmarks #AI

How Artificial Analysis Builds Its Intelligence Index – synesis

Artificial Analysis’s published methodology for its Intelligence Index, the single score it uses to rank language models across nine evaluations.

synesis
LinuxLinks [Unofficial]Jun 16

BOSGAME VTA-439 Mini PC running Linux – Benchmarks

https://fed.brid.gy/r/https://www.linuxlinks.com/bosgame-vta-439-mini-pc-running-linux-benchmarks/

Politico.eu (Unofficial RSS)Jun 17
Turbulent skies: The stealth erosion of EC261 https://www.politico.eu/sponsored-content/turbulent-skies-the-stealth-erosion-of-ec261/?utm_source=RSS_Feed&utm_medium=RSS&utm_campaign=RSS_Syndication #Councilpresidency #Accountability #Lawenforcement #Enforcement #Benchmarks #Parliament #Regulation #Technology #Companies #Inflation #Airlines #Airports #Aviation #History #Markets #Rights #Trade #Data
Turbulent skies: The stealth erosion of EC261

Reducing compensation to symbolic amounts strips the regulation of its primary purpose: consumer protection and accountability.

POLITICO
Kaa  🧶Jun 16
#IA Beaucoup de collègues pensent que LA source des #stéréotypes des grands modèles de langues (ou IA génératives) à la #gpt c'est les données d'entrainement... Certes, elles jouent un rôle, mais il y a d'autres sources de biais stéréotypés. L'une d'elles est assez inattendue : l'évaluation.
En gros, on utilise de très gros jeux de données, souvent de type QCM, pour évaluer les modèles (voire les entraîner), qu'on appelle des #benchmarks. Mes étudiant·e·s sont allés regarder ces benchmarks de plus près et iels ont trouvé pas mal de pbs : des biais explicites (ça avait déjà été mentionné par d'autres), mais également des biais implicites et des erreurs factuelles. Si cela vous intéresse, nous avons écrit sur le sujet un papier en français pour la conférence TALN, que vous trouverez ici : https://inria.hal.science/hal-05618509/file/Benchmarks_trad_TALN-VF.pdf
Bernd von Mallinckrodt 🦋Jun 15
Unruhige Zeiten … Systeme … Standard … Messen … Stabilität von nichtlinearen komplexen Systemen … WarnSignale #EWS … Braucht die #Wissenschaft nicht dieselben reproduzierbaren #Benchmarks für #Frühwarnsignale, … die andere Disziplinen längst als Standard nutzen? doi.org/10.5281/zeno... 🖖
Hacker NewsJun 14

Rio de Janeiro's city government model Rio3.5 beats Qwen3.7 in recent benchmarks

https://twitter.com/zenmagnets/status/2065796012820848699

#HackerNews #RioDeJaneiro #Rio3.5 #Qwen3.7 #benchmarks #citygovernment #technology

𝗭𝗲𝗻 𝗠𝗮𝗴𝗻𝗲𝘁𝘀 (@ZenMagnets) on X

Alibaba Qwen3.7 slowly fading into irrelevance at the frontier due to proprietary stance. In it's place we have Minimax M3 and... *checks notes* Rio 3.5 397b, made by the municipal IT company of Rio de Janeiro's city government. https://t.co/JgIJYVhoEi

X (formerly Twitter)
Linux ✅Jun 13

Windows 11 vs. Linux 🎮 🔌 📊

GPU here used is AMD Radeon RX 6650 XT. Recorded w/ a capture card (no FPS loss). Benchmarked games performing by the enthusiastic NJ Tech:

LEGO Batman: Legacy of the Dark Knight
Cronos: The New Dawn
Resident Evil: Requiem
The Last of Us Part 2
Gothic (remake)
Crimson Desert
Cyberpunk 2077
Star Wars Jedi: Survivor
Clair Obscure: Expedition 33
Death Stranding 2

https://youtu.be/CglxLdJXrxs

#gaming #benchmarks #Linux #CachyOS #Windows #tests #AMD

Windows 11 vs CachyOS - RX 6650 XT (1440p, FSR 4, VRAM Fix)

YouTube
HabrJun 11

Модели почти год, а она всё ещё №1 по цене/качеству. Прогнал свежий батл-тест — и опять не сдвинул

Я собрал четыре модели в один батч — две Gemma, DeepSeek V4 Flash и Qwen3-235B — потому что не поверил кросс-сессионным данным: выходило, что маленькая Gemma обошла большую. В честной очной ставке сюрприз развеялся, обе Gemma оказались вровень. Зато всплыло другое. DeepSeek V4 Flash, который у меня числился на 83, выдал 89 — был недооценён на 6 баллов. А Qwen3-235B-A22B-2507 — модель, вышедшая 21 июля 2025 года, — снова взяла первое место по цене/качеству. Почти год прошёл: Gemini прыгнул с 57 до 97, DeepSeek я перетестировал трижды, MiniMax все хвалят как ровню Opus. А этот июльский чекпойнт просто держит трон, и я уже почти обижаюсь за прогресс. В статье: разбор очной ставки, хронология ретестов одной модели за год, новый критерий — скорость генерации, ещё восемь моделей июня, непопулярное мнение про раскрутку MiniMax и две новые колонки на лидерборде.

https://habr.com/ru/articles/1046582/

#LLM #Qwen #DeepSeek #Gemma #Benchmarks #OpenRouter #Quality_Assessment #Methodology #Russian_NLP

Модели почти год, а она всё ещё №1 по цене/качеству. Прогнал свежий батл-тест — и опять не сдвинул

Собрал четыре модели в один батч — одна сессия, один судья, те же пять тем. Сравнивать баллы из разных тестов нельзя: разница в пару баллов между сессиями — это шум, а не сигнал....

Хабр