ARM Give a man a fire and he's warm for a day, but set fire to him and he's warm for the rest of his life (Terry Pratchett, Jingo)

AMD ROCm Looks Like It Will Finally Be Supporting OpenCL 3.0 Soon

([Radeon] 18 October 06:38 AM EDT AMD ROCm + OpenCL 3.0)

The OpenCL 3.0 compute specification has been out in finalized form since September 2020. Since then NVIDIA's official Windows/Linux drivers have been exposing OpenCL 3.0 going back to 2021, the Intel Compute Runtime stack has also been exposing OpenCL 3.0 support for years, and even with Mesa's Rusticl open-source OpenCL implementation it's beginning to see Gallium3D drivers with conformant OpenCL 3.0. Yet if installing the AMD ROCm compute stack right now, you'll see OpenCL 2.1. But it looks like OpenCL 3.0 will soon be here for ROCm.



Valve Contributes OpenVR Video Driver To SDL

([Valve] 18 October 06:14 AM EDT SDL + OpenVR Video Driver)

Merged to upstream SDL today is an OpenVR video driver that was developed at Valve Software.



Ubuntu Snaps Up Intel's NPU User-Space Software So It's Easier To Accelerate AI

([Ubuntu] 18 October 06:27 AM EDT snap install --beta intel-npu-driver)

Ubuntu Linux maker Canonical has announced the availability of an Intel NPU driver Snap package within their Snap Store to make it easier to leverage the Intel neural processing unit (NPU) on Core Ultra processors within Ubuntu Linux.



Germany's Sovereign Tech Fund Has Invested Over $24.9M In Open-Source In Two Years

([Free Software] 18 October 06:07 AM EDT Sovereign Tech Fund)

Germany's Sovereign Tech Fund (STF) is today celebrating its second anniversary for "empowering public digital infrastructure." In the past two years it has invested more than €23 million (about $24.94M USD) into sixty open technologies.



Microsoft Open-Sources Rust-Written OpenHCL For Running Confidential Intel/AMD VMs

([Microsoft] 17 October 08:28 PM EDT Microsoft OpenHCL)

Microsoft announced today the new and now open-source OpenHCL paravisor for the virtualization stack for enabling Intel TDX and AMD SEV-SNP confidential computing virtual machines (VMs) with this Rust-written software stack. This effort by Microsoft has been five years in the making and is now open-source and will continue to be developed in the open.



Intel NPU Driver Being Updated To Handle Larger AI Workloads

([Intel] 17 October 04:54 PM EDT Intel IVPU Linux Driver)

Following the recent patch work for enabling the Intel 5th Gen NPU premiering with Panther Lake, a new patch series posted today brings a number of improvements for this Intel neural processing unit driver -- including the ability to handle larger workloads.



Exploring The Zen 5 SMT Performance With The AMD EPYC 9755 "Turin" CPU

([Processors] 17 October 01:38 PM EDT 25 Comments)

Continuing on with the testing around the AMD EPYC 9005 series "Turin" processors, today is a look at the Simultaneous Multi-Threading (SMT) performance impact for Turin while using the AMD EPYC 9755 as the highest-end "Turin Classic" processor with 128 cores / 256 threads. Similar SMT on/off tests for "Turin Dense" with the EPYC 9965 192-core / 384-thread will also be coming in a future benchmarking comparison on Phoronix. These tests are mainly intended for reference purposes for those curious about the SMT benefits at such high core counts and what workloads may or may not still benefit from SMT especially when having so many threads while using 12-channel DDR5-6000 memory.



PyTorch 2.5 Released With Improved Intel GPU Support

([Programming] 17 October 12:58 PM EDT PyTorch 2.5)

PyTorch 2.5 is out today as the latest major update to this widely-used machine learning library.



Red Hat Engineer Nikita Popov Now The Lead Maintainer For LLVM

([LLVM] 17 October 10:27 AM EDT LLVM Lead Maintainer)

Following a proposal that began last month, Red Hat engineer Nikita Popov was nominated to become the new lead maintainer for LLVM. Following unaminous approval, as of last week in LLVM Git he's been appointed the official lead maintainer for this critical open-source compiler stack.



Linux 6.13 To Introduce Intel 5th Gen NPU Support For Panther Lake

([Intel] 17 October 08:50 AM EDT Intel 5th Gen NPU In iVPU)

Earlier this month I wrote about Intel's Linux software engineers posting patches adding 5th Gen NPU support to the IVPU accelerator driver for that updated neural processing unit to be found with next-gen Panther Lake processors. Those 5th Gen NPU driver patches for Panther Lake are now queued for introduction with the upcoming Linux 6.13 kernel cycle.



AMD Working On GPU Compute Virtualization Support With ROCm/HIP For VMs

([Radeon] 17 October 06:46 AM EDT ROCm VMs)

Last week at XDC 2024 in Montreal was a status update on AMD's GPU compute virtualization support around their open-source Linux GPU driver and ROCm compute stack.



The Maturing State Of Rusticl For Rust-Based OpenCL Within Mesa

([Mesa] 17 October 06:58 AM EDT Rusticl 2024 Status Update)

Karol Herbst of Red Hat presented in Montreal last week at the X.Org Developers' Conference (XDC 2024) on the current state of Rusticl as the Rust-written OpenCL implementation for Gallium3D drivers within Mesa.



OGRE-Next 3.0 Released For This Open-Source 3D Engine

([Linux Gaming] 17 October 06:30 AM EDT OGRE-Next 3.0)

OGRE-Next 3.0 has debuted this week as the newest version of the Object-Oriented Graphics Rendering Engine Next Generation for serving as an open-source 3D graphics rendering engine.



AMD Releases AOMP 20.0-0 For Radeon/Instinct Compiler Offloading

([AMD] 16 October 04:30 PM EDT AOMP 20.0-0)

Following last week's release of the LLVM/Clang-downstream AOCC 5.0 for optimized compiler support extended to Zen 5 CPUs, the GPU side of the house at AMD this week released AOMP 20.0-0 as their LLVM/Clang downstream focused on GPU device offloading.



Qualcomm Announces Mesa VCL Driver For OpenCL Acceleration Within VMs

([Virtualization] 16 October 02:01 PM EDT OpenCL + VirtIO-GPU)

Qualcomm engineers have developed VCL as a new open-source OpenCL driver for use with VirtIO-GPU for providing OpenCL hardware acceleration within virtual machines.



Open-Source Radeon Vulkan Driver "RADV" Demonstrated On Windows

([Radeon] 16 October 01:11 PM EDT RADV On Windows)

Last week at the X.Org Developer's Conference (XDC2024) in Montreal there was a talk showcasing Mesa's open-source Radeon "RADV" Vulkan driver running atop Windows 11.



Intel Lunar Lake vs. AMD Strix Point Platform Profile Performance Comparison

([Computers] 16 October 11:30 AM EDT 8 Comments)

For those that have been eager to see more Intel Core Ultra Series 200V Lunar Lake Linux testing, here is the latest installment of testing as well as an update from Intel following my Lunar Lake Linux testing recent reports. Today's article is looking at Intel Lunar Lake versus AMD Strix Point across different ACPI Platform Profile configurations for whether you are after peak performance or the most power savings.



GCC Preparing To Set C23 "GNU23" As Default C Language Version

([GNU] 16 October 10:15 AM EDT -std=gnu23)

The GNU Compiler Collection (GCC) support for the C23 programming language standard is now considered "essentially feature-complete" with GCC 15. As such they are preparing to enable the C23 language version (using the GNU23 dialect) by default for the C language version of GCC when not otherwise specified.



Red Hat Enterprise Linux AI 1.2 Brings AMD ROCm + Instinct Tech Preview

([Red Hat] 16 October 10:00 AM EDT RHEL AI 1.2)

Red Hat has announced the GA release of Red Hat Enterprise Linux AI 1.2. RHEL AI was announced earlier this year as Red Hat's AI solution for a foundation model platform to develop / test / run Granite GenAI models. Not to be confused with the RHEL operating system itself, RHEL AI is all about building large language models for enterprise software with Granite LLMs and InstructLab tooling.



AMD Linux Graphics Driver To Switch To More Aggressive Power Heuristics By Default

([Radeon] 16 October 06:51 AM EDT Better Performance)

It looks like for the upcoming Linux 6.13 kernel cycle there could be a nice performance boost for AMD Radeon discrete graphics cards with the AMDGPU kernel driver poised to set more aggressive power heuristics by default.



More

It is a very humbling experience to make a multimillion-dollar mistake, but
it is also very memorable. I vividly recall the night we decided how to
organize the actual writing of external specifications for OS/360. The
manager of architecture, the manager of control program implementation, and
I were threshing out the plan, schedule, and division of responsibilities.
The architecture manager had 10 good men. He asserted that they
could write the specifications and do it right. It would take ten months,
three more than the schedule allowed.
The control program manager had 150 men. He asserted that they
could prepare the specifications, with the architecture team coordinating;
it would be well-done and practical, and he could do it on schedule.
Furthermore, if the architecture team did it, his 150 men would sit twiddling
their thumbs for ten months.
To this the architecture manager responded that if I gave the control
program team the responsibility, the result would not in fact be on time,
but would also be three months late, and of much lower quality. I did, and
it was. He was right on both counts. Moreover, the lack of conceptual
integrity made the system far more costly to build and change, and I would
estimate that it added a year to debugging time.
-- Frederick Brooks Jr., "The Mythical Man Month"