Accelerating your code with CUDA, Vulkan, OpenCL, WebGL and WebGPU.
Custom Development
Building bespoke GPU-powered applications and libraries.
Training & Courses
Upskilling your team with targeted GPU programming training.
Gaussian Splatting
Creating optimized 3D splat models and custom Web3D viewers.
Gaussian Splatting & Custom Viewer
Core11 also provides Gaussian Splatting model creation and custom viewer development. We plan the image capture, handle camera alignment and reconstruction, clean up the trained splat, and prepare compact PLY or SOG assets for reliable web delivery.
Our viewer, geosplat, can be adapted for customer-requested needs, with custom controls, annotations, labels, camera constraints, visual styling, WebXR support, and more.
Camera alignment3D reconstructionPLY/SOG optimizationCustom WebGPU/WebGL viewerWebXR integration
A graphics processing unit, also known as GPU, can process data up to ten times faster than the main
processor of a common PC.
It performs best on large amounts of data, like images, 3D volumes, measurement
data, or physics simulations.
But programming a GPU can be challenging. Its architecture differs a lot from CPUs, and optimal
programming works differently.
Experts at Hand
Core11 GmbH, based in Vienna/Austria,
consults worldwide in application development for GPUs, large and small.
After decades of field experience, we know our way around the perils of GPU programming and can make Computer Vision and High Performance Computing workloads run efficiently on GPUs of any size. Speedups over CPU code are typically ten times or more.
Our Web3D work follows the same engineering model: profiling, GPU-side processing where it pays off, careful asset formats, and viewers that remain responsive on the target hardware.
Our knowledge is based on twenty years of experience with GPGPU (general purpose computation on GPUs). This
includes
pure simulation, image processing, scientific visualization, and interactive prototypes in VR and AR.
We cover embedded systems with GPUs as well (e.g. NVIDIA Jetson).
We constantly evaluate new hardware programming techniques to ensure the highest performance for our customers. Our
preferred hardware platform is NVIDIA hardware, but we are open to working on other platforms like Vulkan and
WebGPU.
