Home
Fujitsu AI Computing Broker
Your GPUs are not lazy —Your Scheduler is:
AI computing broker (ACB) makes GPUs work harder for AI
— not cost more
Facing GPU challenges for AI?
- Your workloads involve alternating phases of CPU and GPU tasks
- You have multiple GPU-capable jobs that can be queued or batched
- Your workloads can utilize full GPU memory intermittently
- You need to host multiple domain-specific LLMs with uneven demand on shared GPU infrastructure
How can we drive peak GPU utilization for AI?
- Increase throughput on shared memory
- Dynamically allocate resources to application need
- Use fewer GPUs across heterogenous workloads
AI computing broker achieves more with your current GPU infrastructure
Runtime-aware GPU allocation
Full memory access
Advanced scheduling algorithms
Docker support
No code changes
Cloud or on-premise
Customer Stories
TRADOM streamlined GPU resources for multi-instance AI training
Xtreme-D Improved efficiency and throughput for AI workloads utilizing GPUs, contributing to reduced computational costs
Learn more about this technology and its applications.
Learn more about ACB
Video 1 : Overview and use cases
Video 2 : Feature and demo
Documents
Related infromation
Learn more about this technology and its applications.