II.
LibraryProcess overview
Reference · livelib-process:gpu-programming--custom-cuda-operator-development
specializations/gpu-programming/custom-cuda-operator-development overview
Custom CUDA Operator Development for Deep Learning - Process for developing custom CUDA kernels integrated with deep learning frameworks (PyTorch, TensorFlow) for specialized operations.
Attributes
displayName
specializations/gpu-programming/custom-cuda-operator-development
description
Custom CUDA Operator Development for Deep Learning - Process for developing custom CUDA kernels
integrated with deep learning frameworks (PyTorch, TensorFlow) for specialized operations.
libraryPath
library/specializations/gpu-programming/custom-cuda-operator-development.js
specialization
gpu-programming
references
- - PyTorch CUDA Extensions: https://pytorch.org/tutorials/advanced/cpp_extension.html
- - TensorFlow Custom Ops: https://www.tensorflow.org/guide/create_op
- - Custom Operator Best Practices: https://pytorch.org/docs/stable/notes/extending.html
example
const result = await orchestrate('specializations/gpu-programming/custom-cuda-operator-development', {
operatorName: 'fused_attention',
framework: 'pytorch',
operation: 'scaled_dot_product_attention',
supportBackward: true
});
usesAgents
- ml-inference-optimizer
- cuda-kernel-expert
- gpu-performance-engineer
Outgoing edges
lib_applies_to_domain1
- domain:scientific-computing·DomainScientific Computing
lib_belongs_to_specialization1
- specialization:gpu-programming·Specialization
lib_implements_workflow1
- workflow:performance-profiling-cycle·WorkflowPerformance Profiling Cycle
lib_involves_role2
- role:computational-scientist·RoleComputational Scientist
- role:ml-engineer·RoleMachine Learning Engineer
lib_requires_skill_area3
- skill-area:cuda-kernels·SkillAreaCUDA Kernel Programming
- skill-area:compute-shaders·SkillAreaCompute Shaders
- skill-area:profiling-cuda·SkillAreaCUDA Profiling
uses_agent3
- lib-agent:gpu-programming--ml-inference-optimizer·LibraryAgentml-inference-optimizer
- lib-agent:gpu-programming--cuda-kernel-expert·LibraryAgentcuda-kernel-expert
- lib-agent:gpu-programming--gpu-performance-engineer·LibraryAgentgpu-performance-engineer
Incoming edges
None.