Page-based virtual memory improves programmer productivity, security, and memory utilization, but incurs performance overheads due to costly page table walks after TLB misses. This overhead can reach 50% for modern workloads that access increasingly vast memory with stagnating TLB sizes.To reduce the overhead of virtual memory, this paper proposes Redundant Memory Mappings (RMM), which leverage ranges of pages and provides an efficient, alternative representation of many virtual-to-physical mappings. We define a range be a subset of process's pages that are virtually and physically contiguous. RMM translates each range with a single range table entry, enabling a modest number of entries to translate most of the process's address space. RMM operates in parallel with standard paging and uses a software range table and hardware range TLB with arbitrarily large reach. We modify the operating system to automatically detect ranges and to increase their likelihood with eager page allocation. RMM is thus transparent to applications.We prototype RMM software in Linux and emulate the hardware. RMM performs substantially better than paging alone and huge pages, and improves a wider variety of workloads than direct segments (one range per program), reducing the overhead of virtual memory to less than 1% on average.

No abstract

Much attention has been given to the efficient execution of the scale-out applications that dominate in datacenter computing. However, the effects of the hardware support in the Memory Management Unit (MMU) in combination with the distinct characteristics of the scale-out applications have been largely ignored until recently. In this paper, we comprehensively quantify the MMU overhead on a real machine leveraging the use of performance counters on a collection of emerging scale-out applications. We show that the MMU overhead accounts for up to 16% of the total execution time due to the high TLB miss rates and the interference between page walks and application data in the cache hierarchy. We find that decreasing the MMU overhead -with large pages -may improve the application performance by up to 13.9%. However, the limited MMU support for large pages in combination with the workloads' low memory locality may even harm the performance when large pages are enabled. By comparing the expected and measured application speedup, we observe a performance gap of up to 3.8%, indicating that any improvements in the MMU may result in more efficient utilization of the available execution resources. Finally, we find that the MMU overhead remains high for most scale-out applications even in the presence of large pages, leaving ample space for optimizations. In response, we present upper bounds for perfect MMU optimizations that motivate rethinking its design in the context of the scale-out applications.