Serverless computing is currently receiving much attention from both academia and industry. It has a straightforward interface that abstracts the complex internal structure of cloud computing resource usage and configuration. The fine grained pay-per-use model of serverless computing can dramatically reduce the cost of using cloud computing resources for users. Thus, today more and more traditional cloud applications are moving to the serverless architecture. In serverless computing, functions executing in containers are the basic unit of scheduling. However, the impact of resource allocation on function performance in serverless platform is still not clear. It is very challenging to improve the function performance while reducing the resource costs in serverless platform. In this paper, we select several typical workloads in serverless and analyze the function performance by controlling the CPU and memory resources. Experimental results reveal the impact of resource allocation on the performance of different types of functions. We also classify the functions in serverless according to their dependence on CPU resources and memory resources.