Making Huge Pages Actually Useful

Panwar, Ashish; Prasad, Aravinda; Gopinath, K.

doi:10.1145/3173162.3173203

Cited by 58 publications

(21 citation statements)

References 33 publications

(23 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In summary, page-table entries are likely to be present in the sockets processor cache. Memory Fragmentation: Physical memory fragmentation limits the availability of large pages as the system ages, leading to higher page-walk overheads [51,56]. Figure 11 shows the performance of Mitosis under heavy fragmentation while using THP in Linux with 2MB page size.…”

Section: Workload Migration Scenariomentioning

confidence: 99%

“…Existing NUMA policies in commodity OSes migrates data pages to the target socket where the workload has been migrated. Unfortunately, page-table migration is not supported [56], making future TLB misses expensive. Such misplacement of page-tables leads to performance degradation for the workload since 100% of TLB misses require remote memory access as shown in top right table of Figure 1 for one workload (GUPS) from workload migration scenario.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Mitosis: Transparently Self-Replicating Page-Tables for Large-Memory Machines

Achermann

Panwar²,

Bhattacharjee

et al. 2020

Proceedings of the Twenty-Fifth International Conference on Architectural Support for Programming Languages and Operating Syste

Self Cite

View full text Add to dashboard Cite

Multi-socket machines with 1-100 TBs of physical memory are becoming prevalent. Applications running on multi-socket machines suffer non-uniform bandwidth and latency when accessing physical memory. Decades of research have focused on data allocation and placement policies in NUMA settings, but there have been no studies on the question of how to place page-tables amongst sockets. We make the case for explicit page-table allocation policies and show that pagetable placement is becoming crucial to overall performance.We propose Mitosis to mitigate NUMA effects on page-table walks by transparently replicating and migrating page-tables across sockets without application changes. This reduces the frequency of accesses to remote NUMA nodes when performing page-table walks. Mitosis uses two components: (i) a mechanism to enable efficient page-table replication and migration; and (ii) policies for processes to efficiently manage and control page-table replication and migration.We implement Mitosis in Linux and evaluate its benefits on real hardware. Mitosis improves performance for large-scale multi-socket workloads by up to 1.34x by replicating pagetables across sockets. Moreover, it improves performance by up to 3.24x in cases when the OS migrates a process across sockets by enabling cross-socket page-table migration.

show abstract

Section: Workload Migration Scenariomentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Mitosis: Transparently Self-Replicating Page-Tables for Large-Memory Machines

Achermann

Panwar²,

Bhattacharjee

et al. 2020

Proceedings of the Twenty-Fifth International Conference on Architectural Support for Programming Languages and Operating Syste

Self Cite

View full text Add to dashboard Cite

show abstract

“…8.2.1 Memory Fragmentation . Memory fragmentation at the application level and at the system level has been extensively studied in prior literature [7,15,18,19,34,40,42,45,49,55,61,65,66,[66][67][68][69]73]. However, we are aware of little work addressing the effects of fragmentation in hugememory workloads.…”

Section: Explorationmentioning

confidence: 99%

∅sim: Preparing System Software for a World with Terabyte-scale Memories

Mansi

Swift

2020

Proceedings of the Twenty-Fifth International Conference on Architectural Support for Programming Languages and Operating Syste

View full text Add to dashboard Cite

Recent advances in memory technologies mean that commodity machines may soon have terabytes of memory; however, such machines remain expensive and uncommon today. Hence, few programmers and researchers can debug and prototype fixes for scalability problems or explore new system behavior caused by terabyte-scale memories.To enable rapid, early prototyping and exploration of system software for such machines, we built and open-sourced the 0sim simulator. 0sim uses virtualization to simulate the execution of huge workloads on modest machines. Our key observation is that many workloads follow the same control flow regardless of their input. We call such workloads dataoblivious. 0sim harnesses data-obliviousness to make huge simulations feasible and fast via memory compression.0sim is accurate enough for many tasks and can simulate a guest system 20-30x larger than the host with 8x-100x slowdown for the workloads we observed, with more compressible workloads running faster. For example, we simulate a 1TB machine on a 31GB machine, and a 4TB machine on a 160GB machine. We give case studies to demonstrate the utility of 0sim. For example, we find that for mixed workloads, the Linux kernel can create irreparable fragmentation despite dozens of GBs of free memory, and we use 0sim to debug unexpected failures of memcached with huge memories. CCS Concepts. • Computing methodologies → Simulation types and techniques; • Software and its engineering → Memory management; Extra-functional properties.

show abstract

“…To reduce TLB misses, recent studies have proposed to optimize TLB organizations by clustering, coalescing, contiguity [14, 18, 21, 42, 54-56, 64, 72], prefetching [15,41,63], speculative TLBs [9], and large part-of-memory TLBs [47,62]. To increase TLB reach, support for huge pages has been extensively studied [21, 26, 27, 29, 43, 49, 51-53, 57, 60, 67, 69], with OS-level improvements [26,43,51,52]. Other works propose direct segments [10,25] and devirtualized memory [31], and suggest that applications manage virtual memory [2].…”

Section: Other Related Workmentioning

confidence: 99%

Elastic Cuckoo Page Tables

Skarlatos

Kokolis

et al. 2020

Proceedings of the Twenty-Fifth International Conference on Architectural Support for Programming Languages and Operating Syste

View full text Add to dashboard Cite

The unprecedented growth in the memory needs of emerging memory-intensive workloads has made virtual memory translation a major performance bottleneck. To address this problem, this paper introduces Elastic Cuckoo Page Tables, a novel page table design that transforms the sequential pointer-chasing operation used by conventional multi-level radix page tables into fully-parallel look-ups. The resulting design harvests, for the first time, the benefits of memorylevel parallelism for address translation. Elastic cuckoo page tables use Elastic Cuckoo Hashing, a novel extension of cuckoo hashing that supports efficient page table resizing. Elastic cuckoo page tables efficiently resolve hash collisions, provide process-private page tables, support multiple page sizes and page sharing among processes, and dynamically adapt page table sizes to meet application requirements.We evaluate elastic cuckoo page tables with full-system simulations of an 8-core processor using a set of graph analytics, bioinformatics, HPC, and system workloads. Elastic cuckoo page tables reduce the address translation overhead by an average of 41% over conventional radix page tables. The result is a 3-18% speed-up in application execution. CCS Concepts. • Software and its engineering → Operating systems; Virtual memory; • Computer systems organization → Architectures.

show abstract

Making Huge Pages Actually Useful

Cited by 58 publications

References 33 publications

Mitosis: Transparently Self-Replicating Page-Tables for Large-Memory Machines

Mitosis: Transparently Self-Replicating Page-Tables for Large-Memory Machines

∅sim: Preparing System Software for a World with Terabyte-scale Memories

Elastic Cuckoo Page Tables

Contact Info

Product

Resources

About