Transactional WaveCache: Towards Speculative and Out-of-Order DataFlow Execution of Memory Operations

Abstract

The WaveScalar is the first DataFlow Architecture that can efficiently provide the sequential memory semantics required by imperative languages. This work presents an alternative memory ordering mechanism for this architecture, the Transaction WaveCache. Our mechanism maintains the execution order of memory operations within blocks of code, called Waves, but adds the ability to speculatively execute, out-of-order, operations from different waves. This ordering mechanism is inspired by progress in supporting Transactional Memories. Waves are considered as atomic regions and executed as nested transactions. If a wave has finished the execution of all its memory operations, as soon as the previous waves are committed, it can be committed. If a hazard is detected in a speculative Wave, all the following Waves (children) are aborted and re-executed. We evaluate the WaveCache on a set artificial benchmarks. If the benchmark does not access memory often, we could achieve speedups of around 90%. Speedups of 33.1% and 24% were observed on more memory intensive applications, and slowdowns up to 16% arise if memory bandwidth is a bottleneck. For an application full of WAW, WAR and RAW hazards, a speedup of 139.7% was verified.

Gamma — General Abstract Model for Multiset mAnipulation and dynamic dataflow model: An equivalence study

“…In this case, the intention is to use a data fusion algorithm written in Gamma 1 to be converted into a dataflow graph. After that, It will be applied studies related to speculative and out‐of‐order dataflow execution 3 and reuse of instruction traces 4 . Another interesting research thread left unexplored is the exploitation of interest‐based communication protocols, like Information Centric Networks (ICN), in the development of an Internet of Things (IoT) environment, via the implementation of Gamma distributed multisets.…”

“…The equivalence study makes it possible to transform a dataflow graph into Gamma code and vice versa. Hence, the Gamma model would take advantage of studies developed for dataflow model, for instance, speculative and out‐of‐order dataflow execution 3 and reuse of instruction traces 4 . On the other hand, a program represented initially by a dataflow graph can be exploited in an execution environment quite suitable, for instance, to an Internet of Things (IoT) environment, after its conversion to Gamma.…”

Exploring the Equivalence between Dynamic Dataflow Model and Gamma - General Abstract Model for Multiset mAnipulation

“…For instance, considering the problem of choosing the smaller element in a some multiset. In the Gamma paradigm, this operation can be performed through a unique reaction, as depicted in Equation (2). We use the syntax of Gamma reactions based on the Gamma implementation provided by Juarez Muylaert [13].…”

“…Through empirical and formal tests we show that these models are equivalent, making it possible to transform a dataflow graph into Gamma code and vice versa. This way, it is possible to provide the scientific community with a series of benefits that comprise exploring and analyzing, in a code written in Gamma, speculative and out-of-order dataflow execution [2], performing instructions trace reuse [3], among others. In addition, a program initially represented by a dataflow graph can be exploited in an execution environment quite suitable, for instance, to an Internet of Things (IoT) environment.…”

TALM: A Hybrid Execution Model with Distributed Speculation Support

“…The Transactional WaveCache [6] is a mechanism that allows speculative out-of-order execution of memory operations in WaveScalar. In this mechanism, memory instructions in a loop iteration are treated as a transaction, which is nested inside transactions of previous iterations.…”