II.
LibraryProcess overview
Reference · livelib-process:gpu-programming--cuda-stream-concurrency
specializations/gpu-programming/cuda-stream-concurrency overview
CUDA Stream and Concurrency Management - Process for implementing concurrent kernel execution and overlapping operations using CUDA streams and events.
Attributes
displayName
specializations/gpu-programming/cuda-stream-concurrency
description
CUDA Stream and Concurrency Management - Process for implementing concurrent kernel execution
and overlapping operations using CUDA streams and events.
libraryPath
library/specializations/gpu-programming/cuda-stream-concurrency.js
specialization
gpu-programming
references
- - CUDA Streams: https://docs.nvidia.com/cuda/cuda-c-programming-guide/
- - CUDA Graphs: https://developer.nvidia.com/blog/cuda-graphs/
example
const result = await orchestrate('specializations/gpu-programming/cuda-stream-concurrency', {
projectName: 'data_pipeline',
concurrencyPatterns: ['kernel-overlap', 'transfer-compute'],
streamCount: 4,
useGraphs: true
});
usesAgents
- 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:concurrency-multithreading·SkillAreaConcurrency and Multithreading
uses_agent2
- lib-agent:gpu-programming--cuda-kernel-expert·LibraryAgentcuda-kernel-expert
- lib-agent:gpu-programming--gpu-performance-engineer·LibraryAgentgpu-performance-engineer
Incoming edges
None.