Community-Benchmark: Testet euer System in MisterHs OpenCL Benchmark
Community-Benchmark: Testet euer System in MisterHs OpenCL Benchmark
All Nerds on Deck: Mit dem MrH OpenCL Benchmark von Community-Mitglied MisterH lassen sich CPUs und GPUs einmal ganz anders vergleichen.
www.computerbase.de
November 7, 2026 at 10:00 AM
Community-Benchmark: Testet euer System in MisterHs OpenCL Benchmark
AMD次期APU「Medusa」、10コア試作機がシングル3100点。内蔵GPUのOpenCLは2万2249点
AMD次期APU「Medusa」、10コア試作機がシングル3100点。内蔵GPUのOpenCLは2万2249点
AMDの次世代APU「Medusa Point」の10コア試作機が、Geekbench 7でシングル3100点、マルチ1万8313点を記録。8・6コア機の結果も公開され、4CUの内蔵GPUはOpenCLで2万2249点を記録した。
joho-todai.com
November 7, 2026 at 4:29 PM
AMD次期APU「Medusa」、10コア試作機がシングル3100点。内蔵GPUのOpenCLは2万2249点
Personally I preferred OpenCL, but no one ever seemed to want to support it well enough to get performance where it needed to be
I gotta try Vulcan Compute sooner or later, though, seems more promising
I gotta try Vulcan Compute sooner or later, though, seems more promising
November 7, 2026 at 1:39 PM
Personally I preferred OpenCL, but no one ever seemed to want to support it well enough to get performance where it needed to be
I gotta try Vulcan Compute sooner or later, though, seems more promising
I gotta try Vulcan Compute sooner or later, though, seems more promising
I have just collected "Mandelbox 4D" by @trips.objkt.social on @objkt.com https://objkt.com/tokens/KT1LjmAdYQCLBjwv4S2oFkEzyHVkomAf5MrW/24245?ref=tz1cXWJ9FxAaEKF7j4SaaAqrNXBnRRamffKV
Mandelbox 4D
The power of math and the high calculation systems summarized in one image, just possible by the evolution of the technology and complex software algorithms. High Quality image. Multiple GPU - OpenCL - Monte-Carlo Rendering. Slightly post-processing.
objkt.com
November 8, 2026 at 9:46 PM
I have just collected "Mandelbox 4D" by @trips.objkt.social on @objkt.com https://objkt.com/tokens/KT1LjmAdYQCLBjwv4S2oFkEzyHVkomAf5MrW/24245?ref=tz1cXWJ9FxAaEKF7j4SaaAqrNXBnRRamffKV
I was very excited to start on my OpenCL path tracer in December, I might still do it, but the tiny render in arm64 is going to take a bit! We'll see! Three more weeks to go. Maybe I can make some good progress yet.
November 8, 2026 at 3:59 AM
I was very excited to start on my OpenCL path tracer in December, I might still do it, but the tiny render in arm64 is going to take a bit! We'll see! Three more weeks to go. Maybe I can make some good progress yet.
Absolutely fascinating! A perfect fusion of mathematics, technology, and art. Brilliant work
👏✨
I have just collected "Mandelbox 4D"
by my dear friend
@trips.objkt.social 💞on @objktcom
objkt.com/tokens/KT1Lj...
👏✨
I have just collected "Mandelbox 4D"
by my dear friend
@trips.objkt.social 💞on @objktcom
objkt.com/tokens/KT1Lj...
Mandelbox 4D
The power of math and the high calculation systems summarized in one image, just possible by the evolution of the technology and complex software algorithms.
High Quality image. Multiple GPU - OpenCL...
objkt.com
November 9, 2026 at 4:50 PM
Absolutely fascinating! A perfect fusion of mathematics, technology, and art. Brilliant work
👏✨
I have just collected "Mandelbox 4D"
by my dear friend
@trips.objkt.social 💞on @objktcom
objkt.com/tokens/KT1Lj...
👏✨
I have just collected "Mandelbox 4D"
by my dear friend
@trips.objkt.social 💞on @objktcom
objkt.com/tokens/KT1Lj...
Dammit, *Vulkan, thanks autocorrect
But either way, I’ve always found trying to force everything into tensor frameworks to be even more limiting than the abstractions needed for OpenCL/Vulkan, even if they do perform well
But either way, I’ve always found trying to force everything into tensor frameworks to be even more limiting than the abstractions needed for OpenCL/Vulkan, even if they do perform well
November 7, 2026 at 1:41 PM
Dammit, *Vulkan, thanks autocorrect
But either way, I’ve always found trying to force everything into tensor frameworks to be even more limiting than the abstractions needed for OpenCL/Vulkan, even if they do perform well
But either way, I’ve always found trying to force everything into tensor frameworks to be even more limiting than the abstractions needed for OpenCL/Vulkan, even if they do perform well
TornadoVM has a #YouTube channel now!
Subscribe here: youtube.com/@TornadoVM-Dev
First up: a 4-minute explainer on how #TornadoVM turns ordinary #Java methods into #GPU tasks, from finding the hot loops to compiling for OpenCL, CUDA or Metal.
👉 youtu.be/BjiEJvNusIk
More videos coming soon.
Subscribe here: youtube.com/@TornadoVM-Dev
First up: a 4-minute explainer on how #TornadoVM turns ordinary #Java methods into #GPU tasks, from finding the hot loops to compiling for OpenCL, CUDA or Metal.
👉 youtu.be/BjiEJvNusIk
More videos coming soon.
From Java to GPU with TornadoVM: Find, Transform, Compose, Execute
YouTube video by TornadoVM
youtu.be
November 4, 2026 at 10:45 AM
TornadoVM has a #YouTube channel now!
Subscribe here: youtube.com/@TornadoVM-Dev
First up: a 4-minute explainer on how #TornadoVM turns ordinary #Java methods into #GPU tasks, from finding the hot loops to compiling for OpenCL, CUDA or Metal.
👉 youtu.be/BjiEJvNusIk
More videos coming soon.
Subscribe here: youtube.com/@TornadoVM-Dev
First up: a 4-minute explainer on how #TornadoVM turns ordinary #Java methods into #GPU tasks, from finding the hot loops to compiling for OpenCL, CUDA or Metal.
👉 youtu.be/BjiEJvNusIk
More videos coming soon.
My thoughts about Epic are pretty well known, I still like reading the part about Epic in this announcement. https://universal-blue.discourse.group/t/bazzite-45-is-here/12702 #gamingonlinux
Bazzite 45 is here!
What’s New & Improved Bazzite is now based on Fedora 45! GNOME 51.1 and KDE 6.8.2 Updated OGC kernels to 7.2.15 and 6.18.60 for NVIDIA LTS images. Mesa 26.2.6. Latest and greatest InputPlumber. Optional gaming networking optimization via bpftune and bpfrag services. ROCM OpenCL is now used over Mesa OpenCL on all non-nvidia images. XWaylandVideoBridge has been removed, fixing an entire class of bugs on our KDE images. Acer Predator Atlas 8 initial support: InputPlumber support, Quick Acce...
universal-blue.discourse.group
November 5, 2026 at 8:23 AM
My thoughts about Epic are pretty well known, I still like reading the part about Epic in this announcement. https://universal-blue.discourse.group/t/bazzite-45-is-here/12702 #gamingonlinux
Intel Compute Runtime Continues Prepping For GPU D Series, OpenCL LEO for Nova Lake P - https://www.phoronix.com/news/Intel-Compute-26.40.40196.8
Intel Compute Runtime Continues Prepping For GPU D Series, OpenCL LEO for Nova Lake P
Released today was the Intel Compute Runtime 26.40.40196.8 as the newest version of their open-source GPU compute stack for both Windows and Linux systems...
www.phoronix.com
November 3, 2026 at 10:55 AM
Intel Compute Runtime Continues Prepping For GPU D Series, OpenCL LEO for Nova Lake P - https://www.phoronix.com/news/Intel-Compute-26.40.40196.8
Snapdragon X2のGPUドライバー32.0.181.0、OpenCL推論で最大87％減速との実測報告
Snapdragon X2のGPUドライバー32.0.181.0、OpenCL推論で最大87％減速との実測報告
Snapdragon X2搭載Surface LaptopでAdreno 32.0.181.0更新後、OpenCLの文章生成が最大87％減速した実測報告。旧版で回復、再更新で再発。4BのCPU・NPUと0.6BのWebGPUはほぼ維持した。
joho-todai.com
November 2, 2026 at 3:53 AM
Snapdragon X2のGPUドライバー32.0.181.0、OpenCL推論で最大87％減速との実測報告
📰 **Intel Compute Runtime Continues Prepping For GPU D Series, OpenCL LEO for Nova Lake P**
Released today was the Intel Compute Runtime 26.40.40196.8 as the newest version of their open-source GPU compute stack for both Windows and Linux systems...
📰 Source: Phoronix
🔗 Link […]
Released today was the Intel Compute Runtime 26.40.40196.8 as the newest version of their open-source GPU compute stack for both Windows and Linux systems...
📰 Source: Phoronix
🔗 Link […]
Original post on igeek.gamer-geek-news.com
igeek.gamer-geek-news.com
November 5, 2026 at 11:02 AM
📰 **Intel Compute Runtime Continues Prepping For GPU D Series, OpenCL LEO for Nova Lake P**
Released today was the Intel Compute Runtime 26.40.40196.8 as the newest version of their open-source GPU compute stack for both Windows and Linux systems...
📰 Source: Phoronix
🔗 Link […]
Released today was the Intel Compute Runtime 26.40.40196.8 as the newest version of their open-source GPU compute stack for both Windows and Linux systems...
📰 Source: Phoronix
🔗 Link […]
O Intel Compute Runtime 26.40.40196.8 chegou com USM ativado por padrão no GPU D Series (Crescent Island), OpenCL LEO no Nova Lake P e standby mode no Battlemage. Análise completa. -> tinyurl.com/54y69e26 #Intel
Intel Compute Runtime 26.40.40196.8 chega com USM no GPU D Series e OpenCL LEO no Nova Lake P
Blog com notícias sobre, Linux, Android, Segurança , etc
tinyurl.com
November 4, 2026 at 11:20 AM
O Intel Compute Runtime 26.40.40196.8 chegou com USM ativado por padrão no GPU D Series (Crescent Island), OpenCL LEO no Nova Lake P e standby mode no Battlemage. Análise completa. -> tinyurl.com/54y69e26 #Intel
OpenCL just works™ as well after running the "GPU rendering" option in the installer
November 3, 2026 at 10:51 AM
OpenCL just works™ as well after running the "GPU rendering" option in the installer
Mas a pesquisa não foi em vão, Vulkan Profile é bem útil, por meio de gpuinfo.org consigo acessar um banco de dados contendo informações de compatibilidade de hardware extensões e outros muito detalhes.
GPU hardware info database launchpad
Home of the GPU hardware databases for Vulkan, OpenGL, OpenGL ES and OpenCL
gpuinfo.org
November 6, 2026 at 5:34 AM
Mas a pesquisa não foi em vão, Vulkan Profile é bem útil, por meio de gpuinfo.org consigo acessar um banco de dados contendo informações de compatibilidade de hardware extensões e outros muito detalhes.
Herní distribuce Bazzite 45 staví na Fedoře 45 a rozšiřuje podporu hardwaru
Herní distribuce Bazzite 45 staví na Fedoře 45 a rozšiřuje podporu hardwaru
Vyšla nová verze herní distribuce Bazzite 45, která je postavená na čerstvé Fedoře 45 a přináší GNOME 51.1 a KDE Plasma 6.8.2. Aktualizována jsou také linuxová jádra, grafická knihovna Mesa na verzi 26.2.6 a InputPlumber, který zajišťuje podporu ovladačů a dalších vstupních zařízení. U ne-Nvidia obrazů nově systém používá ROCm OpenCL místo Mesa OpenCL.
Mezi hlavní novinky patří podpora handheldu Acer Predator Atlas 8, včetně mapování zadních tlačítek, a zařazení podpory ROG Ally X20 do stabilní verze. Bazzite nově podporuje také závodní volanty Logitech TrueForce. Na obrazech KDE řeší odstranění XWaylandVideoBridge celou skupinu dřívějších chyb. Uživatelé mohou využít také volitelnou optimalizaci síťového provozu pro hraní.
Pro přípravu na chystanou podporu Easy Anti-Cheat Linux v2 přibyl příkaz ujust check-anticheat. Prověří požadavky, jako jsou Secure Boot, TPM 2.0, podepsané jádro či IOMMU, a vypíše, co případně chybí. Kontrola systém nijak nemění. Vývojáři zároveň upozorňují, že uživatelé Steam Decku nemají zapínat Secure Boot a u handheldů obecně doporučují zatím nezapínat kernel lockdown.
Výrazné změny se odehrály i pod kapotou: sestavování obrazů bylo přepracováno a zrychleno. Všechny varianty se nyní obvykle sestaví přibližně za 30 minut místo dřívějších téměř 50 minut, zatímco aktualizace mohou díky změně komprese a dělení obrazů zabrat jen 100 MB. Bazzite 45 navíc zpřístupňuje údržbu Btrfs v grafickém nástroji Btrfs Assistant.
www.root.cz
November 5, 2026 at 7:40 AM
Herní distribuce Bazzite 45 staví na Fedoře 45 a rozšiřuje podporu hardwaru
[Phoronix] Intel Compute Runtime Continues Prepping For GPU D Series, OpenCL LEO for Nova Lake P
#Linux #OpenSource
#Linux #OpenSource
Intel Compute Runtime Continues Prepping For GPU D Series, OpenCL LEO for Nova Lake P
Released today was the Intel Compute Runtime 26.40.40196.8 as the newest version of their open-source GPU compute stack for both Windows and Linux systems...
www.linuxnews.net
November 3, 2026 at 12:00 PM
[Phoronix] Intel Compute Runtime Continues Prepping For GPU D Series, OpenCL LEO for Nova Lake P
#Linux #OpenSource
#Linux #OpenSource
📰 **Intel Compute Runtime Continues Prepping For GPU D Series, OpenCL LEO for Nova Lake P**
Released today was the Intel Compute Runtime 26.40.40196.8 as the newest version of their open-source GPU compute stack for both Windows and Linux systems...
📰 Source: Phoronix
🔗 Link […]
Released today was the Intel Compute Runtime 26.40.40196.8 as the newest version of their open-source GPU compute stack for both Windows and Linux systems...
📰 Source: Phoronix
🔗 Link […]
Original post on igeek.gamer-geek-news.com
igeek.gamer-geek-news.com
November 3, 2026 at 10:55 AM
📰 **Intel Compute Runtime Continues Prepping For GPU D Series, OpenCL LEO for Nova Lake P**
Released today was the Intel Compute Runtime 26.40.40196.8 as the newest version of their open-source GPU compute stack for both Windows and Linux systems...
📰 Source: Phoronix
🔗 Link […]
Released today was the Intel Compute Runtime 26.40.40196.8 as the newest version of their open-source GPU compute stack for both Windows and Linux systems...
📰 Source: Phoronix
🔗 Link […]
Bazzite 45 is here! - Bazzite - Universal Blue
https://universal-blue.discourse.group/t/bazzite-45-is-here/12702
https://universal-blue.discourse.group/t/bazzite-45-is-here/12702
Bazzite 45 is here!
What’s New & Improved Bazzite is now based on Fedora 45! GNOME 51.1 and KDE 6.8.2 Updated OGC kernels to 7.2.15 and 6.18.60 for NVIDIA LTS images. Mesa 26.2.6. Latest and greatest InputPlumber. Optional gaming networking optimization via bpftune and bpfrag services. ROCM OpenCL is now used over Mesa OpenCL on all non-nvidia images. XWaylandVideoBridge has been removed, fixing an entire class of bugs on our KDE images. Acer Predator Atlas 8 initial support: InputPlumber support, Quick Acce...
universal-blue.discourse.group
November 4, 2026 at 2:33 PM
Bazzite 45 is here! - Bazzite - Universal Blue
https://universal-blue.discourse.group/t/bazzite-45-is-here/12702
https://universal-blue.discourse.group/t/bazzite-45-is-here/12702
TornadoVM 8.0 introduces a pure-Java compiler that targets CUDA C++, OpenCL, and Metal, replacing Graal-based compilation while keeping existing APIs unchanged.
Save What Matters
Curate Feeds | Make Collections | Customize Email Briefs
briefly.co
November 2, 2026 at 5:18 PM
TornadoVM 8.0 introduces a pure-Java compiler that targets CUDA C++, OpenCL, and Metal, replacing Graal-based compilation while keeping existing APIs unchanged.
C'est l'occasion de rappeler que:
2 trillion dollars (USA) =
2 billions de dollars (France).
Pour nVidia, le pire risque (outre la bulle IA elle-même), c'est OpenCL, qui permet de programmer des CPU massivement parallèles autres que ceux de nVidia (qui promeut son langage propriétaire CUDA).
2 trillion dollars (USA) =
2 billions de dollars (France).
Pour nVidia, le pire risque (outre la bulle IA elle-même), c'est OpenCL, qui permet de programmer des CPU massivement parallèles autres que ceux de nVidia (qui promeut son langage propriétaire CUDA).
October 31, 2026 at 2:08 PM
C'est l'occasion de rappeler que:
2 trillion dollars (USA) =
2 billions de dollars (France).
Pour nVidia, le pire risque (outre la bulle IA elle-même), c'est OpenCL, qui permet de programmer des CPU massivement parallèles autres que ceux de nVidia (qui promeut son langage propriétaire CUDA).
2 trillion dollars (USA) =
2 billions de dollars (France).
Pour nVidia, le pire risque (outre la bulle IA elle-même), c'est OpenCL, qui permet de programmer des CPU massivement parallèles autres que ceux de nVidia (qui promeut son langage propriétaire CUDA).