Liqun Cheng
Liqun Cheng is an Engineering Fellow at Google. His research interests range from architecture, distributed systems, energy proportional computing and machine learning infrastructure. He is particularly interested in interactions across domains with a major focus on software-hardware co-design. He obtained his PhD from the University of Utah and BS degree from Shanghai Jiao Tong University. Prior to Google, Liqun was a performance architect at Intel.
Authored Publications
Sort By
Preventing Network Bottlenecks: Accelerating Datacenter Services with Hotspot-Aware Placement for Compute and Storage
Hamid Bazzaz
Yingjie Bi
Minlan Yu
Ramesh Govindan
Chloe Tsai
Chris DeForeest
Charlie Carver
Jan Kopański
2025
Preview abstract
Datacenter network hotspots, defined as links with persistently high utilization, can lead to performance bottlenecks.In this work, we study hotspots in Google’s datacenter networks. We find that these hotspots occur most frequently at ToR switches and can persist for hours. They are caused mainly by bandwidth demand-supply imbalance, largely due to high demand from network-intensive services, or demand exceeding available bandwidth when compute/storage upgrades outpace ToR bandwidth upgrades. Compounding this issue is bandwidth-independent task/data placement by data-center compute and storage schedulers. We quantify the performance impact of hotspots, and find that they can degrade the end-to-end latency of some distributed applications by over 2× relative to low utilization levels. Finally, we describe simple improvements we deployed. In our cluster scheduler, adding hotspot-aware task placement reduced the number of hot ToRs by 90%; in our distributed file system, adding hotspot-aware data placement reduced p95 network latency by more than 50%. While congestion control, load balancing, and traffic engineering can efficiently utilize paths for a fixed placement, we find hotspot-aware placement – placing tasks and data under ToRs with higher available bandwidth – is crucial for achieving consistently good performance.
View details
Searching for Fast Models on Datacenter Accelerators
Ruoming Pang
Andrew Li
Norm Jouppi
Conference on Computer Vision and Pattern Recognition (2021)
Preview abstract
Neural Architecture Search (NAS), together with model scaling, has shown remarkable progress in designing high accuracy and fast convolutional architecture families. However, as neither NAS nor model scaling considers sufficient hardware architecture details, they do not take full advantage of the emerging datacenter (DC) accelerators. In this paper, we search for fast and accurate CNN model families
for efficient inference on DC accelerators. We first analyze DC accelerators and find that existing CNNs suffer from insufficient operational intensity, parallelism, and execution efficiency and exhibit FLOPs-latency nonproportionality. These insights let us create a DC-accelerator-optimized search space, with space-to-depth, space-to-batch, hybrid fused convolution structures with vanilla and depthwise convolutions, and block-wise activation functions. We further propose a latency-aware compound scaling (LACS), the first multi-objective compound scaling method optimizing both accuracy and latency. Our LACS discovers that network depth should grow much faster than image size and network width, which is quite different from the observations from previous compound scaling. With the new search space
and LACS, our search and scaling on datacenter accelerators results in a new model series named EfficientNet-X. EfficientNet-X is up to more than 2X faster than EfficientNet (a model series with state-of-the-art trade-off on FLOPs and accuracy) on TPUv3 and GPUv100, with comparable accuracy. EfficientNet-X is also up to 7X faster than recent RegNet and ResNeSt on TPUv3 and GPUv100. Source code is at https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet/tpu
View details
Autonomous Warehouse-Scale Computers
Proceedings of the 57th Annual Design Automation Conference 2020, Association for Computing Machinery, New York, NY United States
Preview abstract
Modern Warehouse-Scale Computers (WSCs), composed of many generations of servers and a myriad of domain specific accelerators, are becoming increasingly heterogeneous. Meanwhile, WSC workloads are also becoming incredibly diverse with different communication patterns, latency requirements, and service level objectives (SLOs). Insufficient understanding of the interactions between workload characteristics and the underlying machine architecture leads to resource over-provisioning, thereby significantly impacting the utilization of WSCs.
We present Autonomous Warehouse-Scale Computers, a new WSC design that leverages machine learning techniques and automation to improve job scheduling, resource management, and hardware-software co-optimization to address the increasing heterogeneity in WSC hardware and workloads. Our new design introduces two new layers in the WSC stack, namely: (a) a Software-Defined Server (SDS) Abstraction Layer which redefines the hardware-software boundary and provides greater control of the hardware to higher layers of the software stack through stable abstractions; and (b) a WSC Efficiency Layer which regularly monitors the resource usage of workloads on different hardware types, autonomously quantifies the performance sensitivity of workloads to key system configurations, and continuously improves scheduling decisions and hardware resource QoS policies to maximize cluster level performance. Our new WSC design has been successfully deployed across all WSCs at Google for several years now. The new WSC design improves throughput of workloads (by 7-10%, on
average), increases utilization of hardware resources (up to 2x), and reduces performance variance for critical workloads (up to 25%).
View details
Kelp: QoS for Accelerators in Machine Learning Platforms
Haishan Zhu
Rama Govindaraju
Mattan Erez
International Symposium on High Performance Computer Architecture (2019)
Preview abstract
Development and deployment of machine learning (ML) accelerators in Warehouse Scale Computers (WSCs) demand significant capital investments and engineering efforts. However, even though heavy computation can be offloaded to the accelerators, applications often depend on the host system for various supporting tasks. As a result, contention on host resources, such as memory bandwidth, can significantly discount the performance and efficiency gains of accelerators. The impact of performance interference is further amplified in distributed learning for large models.
In this work, we study the performance of four production machine learning workloads on three accelerator platforms. Our experiments show that these workloads are highly sensitive to host memory bandwidth contention, which can cause 40% average performance degradation when left unmanaged. To tackle this problem, we design and implement Kelp, a software runtime that isolates high priority accelerated ML tasks from memory resource interference. We evaluate Kelp with both production and artificial aggressor workloads, and compare its effectiveness with previously proposed solutions. Our evaluation shows that Kelp is effective in mitigating performance degradation of the accelerated tasks, and improves performance by 24% on average. Compared to previous work, Kelp reduces performance degradation of ML tasks by 7% and improves system efficiency by 17%. Our results further expose opportunities in future architecture designs.
View details
WSMeter: A Fast, Accurate, and Low-Cost Performance Evaluation for Warehouse-Scale Computers
Jaewon Lee
Changkyu Kim
Rama Govindaraju
Jangwoo Kim
Proceedings of the Twenty-Third International Conference on Architectural Support for Programming Languages and Operating Systems (2018) (to appear)
Preview abstract
A warehouse-scale computer (WSC) is a vast collection
of tightly networked computers providing modern internet
services, that is becoming increasingly popular as the most
cost-effective approach to serve users at global scale. It is
however extremely difficult to accurately measure the holistic
performance of WSC. The existing load-testing benchmarks
are tailored towards a dedicated machine model and do not
address shared infrastructure environments. Evaluating the
performance of a live shared production WSC environment
presents many challenges due to the lack of holistic performance
metrics, high evaluation costs, and potential service
disruptions they may cause. WSC providers and customers
are in need of a cost effective methodology to accurately evaluate
the holistic performance of their platforms and hosted
services.
To address these challenges, we propose WSMeter, a cost
effective framework and methodology to accurately evaluate
the holistic performance of WSC in a live production environment.
We define a new performance metric to accurately reflect
the holistic performance of a WSC running a wide variety of
unevenly distributed jobs. We propose a model to statistically
embrace the performance variances amplified by co-located
jobs, to evaluate holistic performance with minimum costs.
For validation of our approach, we analyze two real-world
use cases and show that WSMeter accurately discerns 7% and
1% performance improvements, using only 0.9% and 6.6% of
the machines in the WSC, respectively. We show through a
Cloud customer case study, where WSMeter helped quantify
the performance benefits of service software optimization with
minimal costs.
View details
Improving Resource Efficiency at Scale with Heracles
Rama Govindaraju
Christos Kozyrakis
ACM Transactions on Computer Systems (TOCS), 34 (2016), 6:1-6:33
Preview abstract
User-facing, latency-sensitive services, such as websearch, underutilize their computing resources during daily periods of low traffic. Reusing those resources for other tasks is rarely done in production services since the contention for shared resources can cause latency spikes that violate the service-level objectives of latency-sensitive tasks. The resulting under-utilization hurts both the affordability and energy efficiency of large-scale datacenters. With the slowdown in technology scaling caused by the sunsetting of Moore’s law, it becomes important to address this opportunity.
We present Heracles, a feedback-based controller that enables the safe colocation of best-effort tasks alongside a latency-critical service. Heracles dynamically manages multiple hardware and software isolation mechanisms, such as CPU, memory, and network isolation, to ensure that the latency-sensitive job meets latency targets while maximizing the resources given to best-effort tasks. We evaluate Heracles using production latency-critical and batch workloads from Google and demonstrate average server utilizations of 90% without latency violations across all the load and colocation scenarios that we evaluated.
View details
Heracles: Improving Resource Efficiency at Scale
Rama Govindaraju
Christos Kozyrakis
Proceedings of the 42th Annual International Symposium on Computer Architecture (2015)
Preview abstract
User-facing, latency-sensitive services, such as websearch, underutilize their computing resources during daily periods of low traffic. Reusing those resources for other tasks is rarely done in production services since the contention for shared resources can cause latency spikes that violate the service-level objectives of latency-sensitive tasks. The resulting under-utilization hurts both the affordability and energy-efficiency of large-scale datacenters. With technology scaling slowing down, it becomes important to address this opportunity.
We present Heracles, a feedback-based controller that enables the safe colocation of best-effort tasks alongside a latency-critical service. Heracles dynamically manages multiple hardware and software isolation mechanisms, such as CPU, memory, and network isolation, to ensure that the latency-sensitive job meets latency targets while maximizing the resources given to best-effort tasks. We evaluate Heracles using production latency-critical and batch workloads from Google and demonstrate average server utilizations of 90% without latency violations across all the load and colocation scenarios that we evaluated.
View details
Towards Energy Proportionality for Large-Scale Latency-Critical Workloads
Rama Govindaraju
Luiz André Barroso
Christos Kozyrakis
Proceedings of the 41th Annual International Symposium on Computer Architecture, ACM (2014)
Preview abstract
Reducing the energy footprint of warehouse-scale computer (WSC) systems is key to their affordability, yet difficult to achieve in practice. The lack of energy proportionality of typical WSC hardware and the fact that important workloads (such as search) require all servers to remain up regardless of traffic intensity renders existing power management techniques ineffective at reducing WSC energy use.
We present PEGASUS, a feedback-based controller that significantly improves the energy proportionality of WSC systems, as demonstrated by a real implementation in a Google search cluster. PEGASUS uses request latency statistics to dynamically adjust server power management limits in a fine-grain manner, running each server just fast enough to meet global service-level latency objectives. In large cluster experiments, PEGASUS reduces power consumption by up to 20%. We also estimate that a distributed version of PEGASUS can nearly double these savings.
View details
