mimircache

Cache is a major component in the memory hierarchy, which bridges the gap between fast computing devices and slow storage devices. A considerable amount of effort has been devoted to improving cache performance. However, as cache are used in different settings and environments, they usually deviate from the models researchers have long proposed. In order to provide researchers a better tool to study different cache replacement algorithms, and provide system administrators a better tool to understand the characteristics of their workload, we design Mimircache.

What is Mimircache

Mimircache is the name for two platforms, PyMimircache and CMimircache, both platforms provide the same functions, however, one is writtne in Python3 while the other in C. Besides, PyMimircache also embeds CMMimircache into it when possible, so that PyMimircache provides same performance as CMimircache with a much simple interface. The website of Mimircache is at http://mimircache.info.

Philosolphy of Mimircache

• Performance
• Flexibility
• Easy-to-Use

Features of Mimircache

• Static analysis: computation and visualization of different cache replacement algorithms, different cache statistical information.
• Dynamic analysis: computation and visualization of different cache behavior with changing time; variants of heatmap.
• Easy plugins: write and test your own cache replacement algorithms has been made super easy in mimircache, just following the tutorial, you can write your own algorithm in a few miniutes.
• Open source: mimircache is released under public domain with GPLv3 License.

Roadmap

• Feature: Add sampling layer for all profilers (v0.8.0)
• Milestone: Allow mimircache to interact with Memcached and Redis (v1.0.0)
• Feature: Add GPU profiling support (v1.2.0)
• Feature: Add Full Windows Support (v2.0.0)