Open Source LLM

Tesseract 5.0 OCR Engine Bringing Faster Performance With "Fast Floats"

Tesseract as the leading open-source optical character recognition (OCR) engine that employs neural networks for converting images/scans of text into actual recognized text is nearing its 5.0 release.

The Tesseract 5.0 Alpha has been available since the end of last year while marked this weekend was the first beta of Tesseract 5.0. Earlier Tesseract 5.0 Alpha releases have brought improved performance, support for Apple Silicon, build system improvements, an overhaul to its public API, and a lot of code improvements

Open Source LLM

Arm Working On Clang C++ For OpenCL 2021 (OpenCL 3.0 Compatible)

With LLVM's Clang C/C++ compiler front-end squaring away its OpenCL C 3.0 language support, Arm engineers are now working on the C++ equivalent support.

They are pursuing "C++ for OpenCL 2021" as the C++ equivalent to OpenCL 3.0 .

Pending upstream with The Khronos Group is also the C++ for OpenCL 2021 draft -- also being organized by Arm.

The C++ for OpenCL 202

Open Source LLM

An Early Look At LLVM Clang 13 Performance On AMD Zen 3

With LLVM/Clang 13 feature development having ended last week and the 13.0 release candidate being tagged, in starting off the benchmarking cycle first up I was looking at how well this new compiler is performing compared to LLVM Clang 12 stable on an AMD EPYC 7543 (Zen 3) Linux server.

LLVM 13 is shaping up nicely as the H2'2021 update to this leading and widely-used open-source compiler stack. Early benchmarks I've conducted


Open Source LLM

SDL 2.0.16 Released With Better Wayland Support, PipeWire Integration

SDL 2.0.16 is officially out today as the latest update to this widely-used cross-platform library for serving as an abstraction layer around graphics, controllers / input, and more that makes it widely used by cross-platform games as well as the likes of the Steam Runtime.

SDL 2.0.16 is a big update with much better native Wayland support, including handling for client side decorations and other functionality in place. SDL 2.0.16 also has support for

Open Source LLM

AMD Releases AOMP 13.0-5 Radeon OpenMP Compiler

In addition to the AOCC compiler for Zen CPUs , another LLVM/Clang downstream maintained by AMD is the AOMP compiler as where they host their various patches not yet merged around Radeon OpenMP offloading support. This week marked the release of AOMP 13.0-5 as their latest work on that front for the newest OpenMP GPU offloading capabilities.

AOMP 13.0-5 is the newest release of this compiler currently based on the upstream LLVM state as of April with various AMD patches on top for what hasn'

Open Source LLM

O3DE Game Engine Quickly Settling Its Linux Support

Less than one month since Open 3D Engine was announced based on Amazon's Lumberyard engine , the Linux support is nearly in a pleasant state.

As written about in mid-July, the O3DE Linux editor was getting squared away after initially the Linux support was in rough shape, which was rather unfortunate considering O3DE and the new Open 3D Foundation is backed by the Linux Foundation.

Fortunately thanks to the work of Fabio Anderegg and other developers, the Linux editor is now working and most of

Open Source LLM

GraalVM 21.2 Released With New Optimizations, Better Linux AArch64 Support

Oracle has published a new version of GraalVM , its open-source Java JVM/JDK implemented in Java that also supports other programming languages and execution modes. GraalVM continues to be quite an interesting effort given its various languages supported and interesting technical experiments/features in the name of greater Java performance and other innovative features.

GraalVM 21.2 is another interesting release both for the pure community open-source project and also some new extras only within GraalVM Enterprise too. Among the changes catching my eye with GraalVM 21

Open Source LLM

NVIDIA Releases TensorRT 8.0 With Big Performance Improvements

NVIDIA today is making available a much faster version of TensorRT, its SDK for optimized deep learning inference on their GPUs.

With TensorRT 8 that is being made public today, NVIDIA is reporting "2x performance" relative to the existing TensorRT 7 release. That 2x performance is around transformer optimizations while they are also claiming 2x accuracy against TensorRT 7 when using INT8 with quantization aware training.

Continue reading...

Open Source LLM

LibreOffice 7.2 RC1 Released Ahead Of Official Debut Next Month

LibreOffice 7.2 is expected for release before the end of August while today marks the availability of the first release candidate.

Since last month's LibreOffice 7.2 beta process began, more bug fixes have been accumulating into this major half-year update to this cross-platform open-source office suite. LibreOffice 7.2 introduces a command pop-up / heads-up display, initial GTK4 toolkit support, the ability to compile to WebAssembly, font caching for faster rendering, Calc spreadsheet performance improvements

Open Source LLM

OpenCL 3.0.8 Released With New Extension To Help AI Inferencing

The Khronos Group recently released a new minor point release to the OpenCL 3.0 specification.

The new version is OpenCL 3.0.8 and a prominent extension added this time is cl_khr_integer_dot_product. The new cl_khr_integer_dot_product extension adds support for SPIR-V instructions and OpenCL C built-in functions to compute the dote product of vectors of integers. The cl_khr_integer_dot_product extension was worked on by Intel, Arm