GigaSpaces XAP version 7.0 has introduced a wealth of new features and enhancements which many of our customers have found very useful. But as always, we’ve also worked very hard on improving our performance and scalability, and specifically to this version, multi-core performance and scalability.

Multithreading and parallel computation has recently become necessary to take full advantage of the gains allowed by Moore's law. For years, processor makers such as Intel consistently delivered ever-increasing processor clock speeds, so that single-threaded code executed faster on newer processors with no modification.

Nowadays, processor makers focus on multi-core chip designs, and software has to be written in a multithreaded or manner to take full advantage of the hardware. In line with this trend, the new Intel Nehalem processor architecture offers new levels of parallelism, performance and energy efficiency.

GigaSpaces XAP 7.0 has been specifically optimized to take advantage of such environments and to allow highly multithreaded applications to run as efficiently as possible with the least possible resource contention. Specifically, our in-memory transaction and locking mechanisms have been refactored to use more lightweight locking and synchronization constructs supported by modern processors and later versions of the Java virtual machine. This in turn helped us achieve significant performance gains, as seen below.

GigaSpaces XAP version 7.0 was benchmarked on Intel’s latest Xeon 5500 Nehalem processor and reached 1 million data updates per second and 2.6 million data retrievals per second with four client threads on the Nehalem, while reaching 276,000 updates per second and 453,000 retrievals per second with XAP 6 on the best previous Intel processor. The combination of the latest product versions also demonstrated 30 percent better scalability than previously. This test was part of the fasterAPPS programme, a joint initiative among Intel, MPI Europe and Globant, designed to promote the migration of financial applications from legacy computer architectures to Intel's latest Xeon 5500 processor.

Results Summary

The combination of the Nehalem architecture with GigaSpaces XAP 7.0 provides very impressive raw performance numbers as well as near-linear multi-core scalability.

This is true for both remote and local space configurations.

As can be expected, the performance of embedded in-memory operations is an order of magnitude better than that of remote operations.

This emphasizes the advantages of the GigaSpaces XAP platform and space based architecture, which focus on partitioning and collocation of data and business logic to achieve optimal performance and scalability.

Benchmark Configuration

The benchmark was conducted with 3 servers (see configuration below), The single embedded space scenario was conducted on a single serevr, the single remote space scenario utilized two servers (one for the remote client and one for the space) and the partitioned remote space scenario used all three servers.

Server specifications:

CPU: Two Nehalem 2.80GHz C0 step processors

Memory: 24GB DDR3 1333MHz memory

Speed Step and Hyper Threading: off

Network interface: Mellanox ConnectX QDR Infiniband, connected to Mellanox switch

Operating System: RHEL 5

Java Version: Sun JVM 1.6.0_12

GigaSpaces Version: XAP Premium 7.0.1 GA (build 3818)