Martin Maas

Martin Maas

Auto-generated profile, based on publicly available information

Martin Maas

Google LLC, Mountain View, California, United States

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
30 July 2020 in IEEE Micro
Verified
Journal Article
Open access
MMMartin Maas
Martin Maas
Machine learning has the potential to significantly improve systems, but only under certain conditions. We describe a taxonomy to help identify whether or not machine learning should be applied to particular systems problems, and which approaches are most promising. We believe that this taxonomy can help practitioners and researchers decide how to most effectively use machine learning in their systems, and provide the community with a framework and vocabulary to discuss different approaches for applying machine learning in systems.

22 June 2021
Verified
Conference Paper
Open access
MMMartin Maas
Martin Maas
CKChris Kennelly
Chris Kennelly
KNKhanh Nguyen
Khanh Nguyen
DGDarryl Gove
Darryl Gove
KMKathryn S. McKinley
Kathryn S. McKinley
PTPaul Turner
Paul Turner
Modern C++ server workloads rely on 2 MB huge pages to improve memory system performance via higher TLB hit rates. Huge pages have traditionally been supported at the kernel level, but recent work has shown that user-level, huge page-aware memory allocators can achieve higher huge page coverage and thus performance. These memory allocators deal with a trade-off: 1) allocate memory from the operating system (OS) at the granularity of a huge page, achieve high performance, but potentially waste memory due to fragmentation, or 2) limit fragmentation by breaking up huge pages into smaller 4 KB pages and returning them to the OS, but reduce performance due to lower huge page coverage. For example, the state-of-the-art TCMalloc allocator handles this trade-off by releasing memory to the OS at a configurable release rate, breaking up huge pages as necessary. This approach balances performance and fragmentation well for machines running one workload. For multiple applications on the same machine however, the reduction in memory usage is only useful to overall performance if another workload uses this memory. In warehouse-scale computers, when an application releases and then reacquires the same amount or more memory quickly, but no other application uses the memory in the meantime, the release causes poorer huge page coverage without any system-wide benefit. We introduce a metric, realized fragmentation, to capture this effect. We then present an adaptive release policy that dynamically determines when to break up huge pages and return them to the OS to optimize system-wide performance. We built this policy into TCMalloc and deployed it fleet-wide in our data centers, leading to an estimated 1% fleet-wide throughput improvement at negligible memory overhead.

27 April 2024
Verified
Conference Paper
Open access
PPPhitchaya Mangpo Phothilimthana
Phitchaya Mangpo Phothilimthana
SKSaurabh Kadekodi
Saurabh Kadekodi
SGSoroush Ghodrati
Soroush Ghodrati
SMSelene Moon
Selene Moon
MMMartin Maas
Martin Maas
Representative modeling of I/O activity is crucial when designing large-scale distributed storage systems. Particularly important use cases are counterfactual "what-if" analyses that assess the impact of anticipated or hypothetical new storage policies or hardware prior to deployment. We propose Thesios, a methodology to accurately synthesize such hypothetical full-resolution I/O traces by carefully combining down-sampled I/O traces collected from multiple disks attached to multiple storage servers. Applying this approach to real-world traces that are already routinely sampled at Google, we show that our synthesized traces achieve 95--99.5% accuracy in read/write request numbers, 90--97% accuracy in utilization, and 80--99.8% accuracy in read latency compared to metrics collected from actual disks. We demonstrate how Thesios enables diverse counterfactual I/O trace synthesis and analyses of hypothetical policy, hardware, and server changes through four case studies: (1) studying the effects of changing disk's utilization, fullness, and capacity, (2) evaluating new data placement policy, (3) analyzing the impact on power and performance of deploying disks with reduced rotations-per-minute (RPM), and (4) understanding the impact of increased buffer cache size on a storage server. Without Thesios, such counterfactual analyses would require costly and potentially risky A/B experiments in production.

14 March 2022
Verified
Conference Paper
Closed access
PJParas Jain
Paras Jain
SHSafeen Huda
Safeen Huda
MMMartin Maas
Martin Maas
JGJoseph E. Gonzalez
Joseph E. Gonzalez
ISIon Stoical
Ion Stoical
AMAzalia Mirhoseini
Azalia Mirhoseini
Approximate computing is a promising way to improve the power efficiency of deep learning. While recent work proposes new arithmetic circuits (adders and multipliers) that consume substantially less power at the cost of computation errors, these approximate circuits decrease the end-to-end accuracy of common models. We present AutoApprox, a framework to automatically generate approximate low-power deep learning accelerators without any accuracy loss. AutoApprox generates a wide range of approximate ASIC accelerators with a TPUv3 systolic-array template. AutoApprox uses a learned router to assign each DNN layer to an approximate systolic array from a bank of arrays with varying approximation levels. By tailoring this routing for a specific neural network architecture, we discover circuit designs without the accuracy penalty from prior methods. Moreover, AutoApprox optimizes for the end-to-end performance, power and area of the the whole chip and PE mapping rather than simply measuring the performance of the arithmetic units in iso-lation. To our knowledge, our work is the first to demonstrate the effectiveness of custom-tailored approximate circuits in delivering significant chip-level energy savings with zero accuracy loss on a large-scale dataset such as ImageNet. AutoApprox synthesizes a novel approximate accelerator based on the TPU that reduces end-to-end power consumption by 3.2% and area by 5.2% at a sub-10nm process with no degradation in ImageNet validation top-1 and top-5 accuracy.


25 September 2026 in Communications of the ACM
Verified
Journal Article
Open access
ECEric Christopher
Eric Christopher
KCKevin Crossan
Kevin Crossan
WDWolff Dobson
Wolff Dobson
CKChris Kennelly
Chris Kennelly
DLDrew Lewis
Drew Lewis
BYBrian Yang
Brian Yang
Migrating codebases from one instruction set architecture (ISA) to another is a major engineering challenge. A recent example is the adoption of Arm (in addition to x86) across major cloud hyperscalers. Yet, this problem has seen limited attention from the academic community. Most work assumes migration is straightforward in the presence of source code, or that the main research challenges lie in binary translation used in its absence. We find that neither assumption holds today. Modern ISA migrations can build on a robust open-source ecosystem, making it possible to recompile all relevant software from scratch. However, while this avoids the need for binary translation, modern software ecosystems introduce a new, multifaceted set of challenges. By analyzing a large-scale migration from x86 to Arm at Google spanning nearly 40,000 code commits, we derive a taxonomy of tasks involved in modern ISA migration. We show how Google automated many of these steps and demonstrate how AI can play a major role in performing them. Finally, we identify tasks that remain difficult and highlight research challenges warranting further attention.