Fair CPU time accounting in CMP+SMT processors

Abstract

Processor architectures combining several paradigms of Thread-Level Parallelism (TLP), such as CMP processors in which each core is SMT, are becoming more and more popular as a way to improve performance at a moderate cost. However, the complex interaction between running tasks in hardware shared resources in multi-TLP architectures introduces complexities when accounting CPU time (or CPU utilization) to tasks. The CPU utilization accounted to a task depends on both the time it runs in the processor and the amount of processor hardware resources it receives. Deploying systems with accurate CPU accounting mechanisms is necessary to increase fairness. Moreover, it will allow users to be fairly charged on a shared data center, facilitating server consolidation in future systems.In this article we analyze the accuracy and hardware cost of previous CPU accounting mechanisms for pure-CMP and pure-SMT processors and we show that they are not adequate for CMP+SMT processors. Consequently, we propose a new accounting mechanism for CMP+SMT processors which: (1) increases the accuracy of accounted CPU utilization; (2) provides much more stable results over a wide range of processor setups; and (3) does not require tracking all hardware shared resources, significantly reducing its implementation cost. In particular, previous proposals lead to inaccuracies between 21% and 79% when measuring CPU utilization in an 8-core 2-way SMT processor, while our proposal reduces this inaccuracy to less than 5.0%.

GDP: Using Dataflow Properties to Accurately Estimate Interference-Free Performance at Runtime

“…Accounting schemes for SMT processors have also been proposed (e.g., [37,38]). Here, the focus is on handling interference in the processor core resources and the per-core memory system.…”

Sensible Energy Accounting with Abstract Metering for Multicore Systems

“…The concept of SEA is inspired by CPU accounting [Luque et al 2009] developed for multicores [Luque et al 2012] and for SMT cores [Eyerman and Eeckhout 2009;Eyerman et al 2006;Luque et al 2013]. CPU accounting measures the CPU utilization of a given task during a period of time when it runs on a multithreaded processor.…”

“…Thus, SEA core is estimated as follows: MIBTA (T i ) stands for the execution time predicted during the MIBTA phases when T i is running with M N of the core resources. Before entering the MIBTA phases (every 2.6 million cycles [Luque et al 2013]), the execution of all tasks is stalled. Then, a controller restores the execution of a particular task to allow it to run alone in the core for 50,000 cycles to warm up.…”

Per-task Energy Accounting in Computing Systems

“…The concept of Energy Accounting is inherited from CPU accounting that was first introduced in [11] and then develop for multicores in [10], [12] and for SMTs in [7]. CPU accounting measures the CPU utilization of a given task during a period of time when it runs in a multicore processor.…”

“…In our solution for computing the LLC energy a task would consume with a fair share of the resources, we combine the proposed per-task energy metering mechanism with hardware support for fair CPU accounting [7], [10]- [12]. In particular, we use ITCA, the technique proposed in [12] that focuses on the progress a task would have with a fair share of the resources.…”