Alibaba Cloud heterogeneous computing services provide a comprehensive system in which software is integrated with hardware to help you allocate and scale resources in a flexible and elastic manner, increase computing power, and control costs for AI business.
Heterogeneous computing
Heterogeneous computing is a system that consists of computing units that have different instruction sets and architecture types. The Alibaba Cloud heterogeneous computing service family includes Elastic GPU Service and FPGA as a Service (FaaS). Heterogeneous computing services provide dedicated hardware based on the scenario. In specific scenarios, heterogeneous computing services can provide more than 10 times higher efficiency and cost-effectiveness than regular Elastic Compute Service (ECS) instances. Heterogeneous computing balances performance, costs, and power consumption. Dedicated hardware of heterogeneous computing services can reduce power consumption to achieve optimal performance and costs.
The rapid development of AI technologies, such as deep learning, promotes the complexity and accuracy of AI computing models and increases the demand for computing power and performance. Against this backdrop, an increasing number of AI computing services use heterogeneous computing to improve performance. Alibaba Cloud develops cloud-based AI accelerators for heterogeneous computing services. The AI accelerators use a centralized framework to improve the performance of mainstream AI computing frameworks, such as TensorFlow, PyTorch, MxNet, and Caffe, and optimize the performance of Ethernet and heterogeneous accelerators.
Heterogeneous computing service family
The following section describes the services that belong to the Alibaba Cloud heterogeneous computing service family, such as Elastic GPU Service. Alibaba Cloud provides DeepGPU to enhance the GPU computing capabilities of Elastic GPU Service. DeepGPU provides multiple components, such as FastGPU and cGPU.
Elastic GPU Service
Elastic GPU Service provides GPU-accelerated instances, which are computing servers based on GPUs. GPUs have unique advantages over CPUs in mathematical and geometric computations. In specific scenarios such as floating-point and parallel computing, GPUs can provide more than 100 times higher computing power than CPUs. GPU-accelerated instances combine the computing power of GPUs and CPUs and provide ready-to-use and scalable GPU computing resources for various scenarios, such as AI, high-performance computing, and professional graphics processing. For more information, see What is Elastic GPU Service?
DeepGPU
DeepGPU is a toolkit provided by Alibaba Cloud to enhance the GPU computing capabilities of Elastic GPU Service. You can use DeepGPU to quickly set up enterprise-level services based on Infrastructure as a service (IaaS) products. All components of DeepGPU are free of charge. You can use the components together with Elastic GPU Service to run GPU resources in Alibaba Cloud in a convenient and efficient manner. The following DeepGPU components are included:
FastGPU: a set of fast deployment tools provided by Alibaba Cloud for AI computing. For more information, see What is FastGPU?
cGPU: a container sharing technology provided by Alibaba Cloud to isolate virtual GPUs (vGPUs) based on kernels. You can use cGPU to efficiently and conveniently deploy multiple containers on a single GPU. This allows the containers to share the same GPU. This way, you can securely isolate your business, utilize GPU hardware resources, and reduce costs. For more information, see What is cGPU?