Corpus ID: 221703014

The Cost of Software-Based Memory Management Without Virtual Memory

  title={The Cost of Software-Based Memory Management Without Virtual Memory},
  author={Drew Zagieboylo and G. Suh and A. Myers},
Virtual memory has been a standard hardware feature for more than three decades. At the price of increased hardware complexity, it has simplified software and promised strong isolation among colocated processes. In modern computing systems, however, the costs of virtual memory have increased significantly. With large memory workloads, virtualized environments, data center computing, and chips with multiple DMA devices, virtual memory can degrade performance and increase power usage. We… Expand

Figures and Tables from this paper


Efficient virtual memory for big memory servers
This work proposes mapping part of a process's linear virtual address space with a direct segment, while page mapping the rest of thevirtual address space to remove the TLB miss overhead for big-memory workloads. Expand
CARAT: a case for virtual memory through compiler- and runtime-based address translation
This work makes a case for a new model for virtual memory, compiler- and runtime-based address translation (CARAT), which is a co-design between the compiler and the OS kernel, and argues for the feasibility of CARAT through an empirical study of application characteristics and kernel behavior. Expand
Efficient Tagged Memory
The cache behavior of an in-memory tag table is characterized and it is demonstrated that an optimized implementation can typically achieve a near-zero memory traffic overhead. Expand
Do-It-Yourself Virtual Memory Translation
The Do-It-Yourself virtual memory translation architecture is introduced as a flexible complement for current hardware-fixed translation flows and it is shown that different DVMT configurations preserve the native performance, while achieving speedups of 1.2x to 2.0x in virtualized environments. Expand
Supporting Superpages and Lightweight Page Migration in Hybrid Memory Systems
A novel memory management mechanism called Rainbow that supports lightweight page migration without splintering superpages by introducing an efficient hot page identification mechanism and a novel NVM-to-DRAM address remapping mechanism. Expand
Mondrian memory protection
This work extends MMP to support segment translation which allows a memory segment to appear at another location in the address space, and uses this translation to implement zero-copy networking underneath the standard read system call interface. Expand
HawkEye: Efficient Fine-grained OS Support for Huge Pages
The primary ideas behind HawkEye management algorithms are async page pre-zeroing, de-duplication of zero-filled pages, fine-grained page access tracking and measurement of address translation overheads through hardware performance counters. Expand
Thinking about A New Mechanism for Huge Page Management
SysMon-H, a sampling module in OS kernel, which is able to obtain the huge page utilization in a low overhead for both cloud and desktop applications is proposed and H-Policy, a huge page management policy, which splits the underutilized huge pages to mitigate the memory bloating is proposed. Expand
TIMBER-V: Tag-Isolated Memory Bringing Fine-grained Enclaves to RISC-V
TIMBER-V is a new tagged memory architecture featuring flexible and efficient isolation of code and data on small embedded systems, and significantly reduces the problem of memory fragmentation, and improves dynamic reuse of untrusted memory across security boundaries. Expand
Preserving Virtual Memory by Mitigating the Address Translation Wall
This article discusses challenges that pose a threat to the continued progress of computing and presents some promising ways to counter them. Expand