From Jacket Wiki
Jacket is a run-time platform to execute your code faster on GPUs. It includes:
|
|
- Automatic detection of memory transfers
- Optimized kernel launch configurations
|
- Just-in-time (JIT) compilation allowing code to efficiently run on the GPU
- Simultaneous access to multiple GPUs
|
|
|
Getting Started
Licensing
|
Base Libraries
Premium Libraries
- DLA: Dense linear algebra library
- SLA: Sparse linear algebra library
|
Premium Features
- MGL: 2 to 8 GPUs in a single computer
- HPC: 8+ GPUs in multiple computers
- SDK: Integrate custom CUDA code
- JMC: Deploy with MATLAB CompilerTM
Extra Resources
|
|
If you are interested in contributing to this wiki documentation, you may request access by sending an email to support@accelereyes.com. Also, we invite you to share your feedback, questions, and experiences on the Jacket Forums.