Code specialization for memory efficient hash tries (short paper)

Abstract

The hash trie data structure is a common part in standard collection libraries of JVM programming languages such as Clojure and Scala. It enables fast immutable implementations of maps, sets, and vectors, but it requires considerably more memory than an equivalent array-based data structure. This hinders the scalability of functional programs and the further adoption of this otherwise attractive style of programming.In this paper we present a product family of hash tries. We generate Java source code to specialize them using knowledge of JVM object memory layout. The number of possible specializations is exponential. The optimization challenge is thus to find a minimal set of variants which lead to a maximal loss in memory footprint on any given data. Using a set of experiments we measured the distribution of internal tree node sizes in hash tries. We used the results as a guidance to decide which variants of the family to generate and which variants should be left to the generic implementation.A preliminary validating experiment on the implementation of sets and maps shows that this technique leads to a median decrease of 55% in memory footprint for maps (and 78% for sets), while still maintaining comparable performance. Our combination of data analysis and code specialization proved to be effective.

M3: A general model for code analytics in rascal

“…Soft references mitigate memory limitations nicely at the cost of runtime efficiency. On the other hand we are investing in low level optimizations for in-memory hierarchical data-structures [5], [15]. These results are hopeful, making it feasible to scale to even larger analyses without sacrificing immutability.…”

Enabling PHP software engineering research in Rascal

“…Recent work has focused on optimizing the sets (which are also used for relations) and maps in Rascal [31,32], with the latter [32] using control-flow graphs extracted by PHP AiR as part of the evaluation dataset. Work on PHP AiR has also led to improvements in Rascal libraries for statistics, for accessing external data sources, and for serializing large volumes of data.…”

“…This also enables easy experimentation at the algorithmic level. Memory performance is critical for this [31,32] since a common reason to not use such straightforward implementations is to increase performance.…”

“…There are still a number of cases where memory use or execution performance is unsatisfactory, but, more positively, these cases are driving improvements in the algorithms used in PHP AiR and in the implementation of Rascal [31,32].…”

Fast and Lean Immutable Multi-Maps on the JVM based on Heterogeneous Hash-Array Mapped Tries

“…Our point of departure is the Hash-Array Mapped Trie (HAMT) data structure [2], which has proven to be an efficient immutable alternative to array-based implementations. In contrast to arrays, HAMTs enable fine-grained memory layout optimizations [17]. There exists an optimized encoding [18] of HAMTs tailored the JVM, named Compressed Hash-Array Mapped Prefix-tree (CHAMP).…”