Performance Expectations and Guidelines for MPI Derived Datatypes

Gropp, William; Hoefler, Torsten; Thakur, Rajeev; Träff, Jesper Larsson

doi:10.1007/978-3-642-24449-0_18

Cited by 20 publications

(16 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Since some MPI implementations struggle to fulfill elementary performance expectations when handling derived datatypes [6], their adoption is not widespread yet. More and more success stories about improving performance using MPI DDTs are reported [1,8] and tools are available that enable users to quickly change their code from using manual pack loops to leveraging derived datatypes [10].…”

Section: Related Workmentioning

confidence: 99%

MPI datatype processing using runtime compilation

Schneider

Kjølstad

Hoefler

2013

Proceedings of the 20th European MPI Users' Group Meeting

Self Cite

View full text Add to dashboard Cite

Data packing before and after communication can make up as much as 90% of the communication time on modern computers. Despite MPI's well-defined datatype interface for non-contiguous data access, many codes use manual pack loops for performance reasons. Programmers write accesspattern specific pack loops (e.g., do manual unrolling) for which compilers emit optimized code. In contrast, MPI implementations in use today interpret datatypes at pack time, resulting in high overheads. In this work we explore the effectiveness of using runtime compilation techniques to generate efficient and optimized pack code for MPI datatypes at commit time. Thus, none of the overhead of datatype interpretation is incurred at pack time and pack setup is as fast as calling a function pointer. We have implemented a library called libpack that can be used to compile and (un)pack MPI datatypes. The library optimizes the datatype representation and uses the LLVM framework to produce vectorized machine code for each datatype at commit time. We show several examples of how MPI datatype pack functions benefit from runtime compilation and analyze the performance of compiled pack functions for the data access patterns in many applications. We show that the pack/unpack functions generated by our packing library are seven times faster than those of prevalent MPI implementations for 73% of the datatypes used in a scientific application and in many cases outperform manual pack loops.

show abstract

Section: Related Workmentioning

confidence: 99%

MPI datatype processing using runtime compilation

Schneider

Kjølstad

Hoefler

2013

Proceedings of the 20th European MPI Users' Group Meeting

Self Cite

View full text Add to dashboard Cite

show abstract

“…One effort looked at exporting the MPI datatype capabilities to other programming systems [23]. In collaboration with other researchers, we examined the implications of adding performance requirements on MPI implementations for the use of datatypes in [14]; this is important in ensuring that users of MPI can count on reasonable behavior from their MPI implementations. Following up on this work, we have begun developing datatype optimization approaches that extend "just in time" compilation techniques to MPICH.…”

Section: Some Of the Most Interesting Results From This Project Addrementioning

confidence: 99%

Final Report for Enhancing the MPI Programming Model for PetaScale Systems

Gropp

2013

View full text Add to dashboard Cite

show abstract

“…However, performance of current datatype implementations remains suboptimal and has not received as much attention as latency and bandwidth, probably due to the lack of a reasonable and simple benchmark. For example Gropp et al found that several basic performance expectations are violated by MPI implementations in use today [9].…”

Section: Related Workmentioning

confidence: 99%

“…Not many scientific codes leverage MPI DDTs, even though their usage would be appropriate in many cases. One of the reasons might be that current MPI implementations in some cases still fail to deliver the expected performance, as shown by Gropp et al in [9], even though a lot of work is done on improving DDT implementations [6,18,20]. Most of this work is guided by a small number of micro-benchmarks.…”

Section: Introductionmentioning

confidence: 99%

Micro-applications for Communication Data Access Patterns and MPI Datatypes

Schneider

Gerstenberger

Hoefler

2012

Recent Advances in the Message Passing Interface

Self Cite

View full text Add to dashboard Cite

Abstract. Data is often communicated from different locations in application memory and is commonly serialized (copied) to send buffers or from receive buffers. MPI datatypes are a way to avoid such intermediate copies and optimize communications, however, it is often unclear which implementation and optimization choices are most useful in practice. We extracted the send/recv-buffer access pattern of a representative set of scientific applications into micro-applications that isolate their data access patterns. We also observed that the buffer-access patterns in applications can be categorized into three different groups. Our microapplications show that up to 90% of the total communication time can be spent with local serialization and we found significant performance discrepancies between state-of-the-art MPI implementations. Our microapplications aim to provide a standard benchmark for MPI datatype implementations to guide optimizations similarly to SPEC CPU and the Livermore loops do for compiler optimizations.

show abstract

Performance Expectations and Guidelines for MPI Derived Datatypes

Cited by 20 publications

References 10 publications

MPI datatype processing using runtime compilation

MPI datatype processing using runtime compilation

Final Report for Enhancing the MPI Programming Model for PetaScale Systems

Micro-applications for Communication Data Access Patterns and MPI Datatypes

Contact Info

Product

Resources

About