Adaptive Kernel Merge and Fusion for Multi-Tenant Inference in Embedded GPUs

Keyphrases

• Multi-tenant 100%
• Adaptive Kernel 100%
• Kernel Fusion 100%
• Embedded GPU 100%
• Execution Time 50%
• Stall 50%
• Scheduling Techniques 50%
• Embedded Systems 50%
• Short Tails 50%
• Number of Groups 50%
• Software Development Methodology 50%
• Queuing Delay 50%
• Fusion Technique 50%
• Number of Requests 50%
• Deep Learning Model 50%
• Context Switch 50%
• Tail Latency 50%
• Multiple Inference 50%
• Concurrent Kernels 50%
• GPU-based Systems 50%
• Kernel Scheduling 50%

Computer Science

• Graphics Processing Unit 100%
• Embedded Systems 33%
• Software Technique 33%
• Deep Learning Model 33%
• Context Switching 33%
• Multiple Inference 33%
• Kernel Execution Time 33%
• Queueing Delay 33%