September 26, 2011
Background – Promises And Potential
Last year I wrote about Oracle’s new plans for SPARC, anchored by a new line of SPARC CPUs engineered in conjunction with Fujitsu (Does SPARC have a Future?), and commented that the first deliveries of this new technology would probably be in early 2012, and until we saw this tangible evidence of Oracle’s actual execution of this road map we could not predict with any confidence the future viability of SPARC.
The T4 CPU
Fast forward a year and Oracle has delivered the first of the new CPUs, ahead of schedule and with impressive gains in performance that make it look like SPARC will remain a viable platform for years. Specifically, Oracle has introduced the T4 CPU and systems based on them. The T4, an evolution of Oracle’s highly threaded T-Series architecture, is implemented with an entirely new core that will form the basis, with variations in number of threads versus cores and cache designs, of the future M and T series systems. The M series will have fewer threads and more performance per thread, while the T CPUs will, like their predecessors, emphasize throughput for highly threaded workloads. The new T4 will have 8 cores, and each core will have 8 threads. While the T4 emphasizes highly threaded workload performance, it is important to note that Oracles has radically improved single-thread performance over its predecessors, with Oracle claiming performance per thread improvements of 5X over its predecessors, greatly improving its utility as a CPU to power less thread-intensive workloads as well.
The SPARC SuperCluster
Drawing on the legacy of its Exadata and Exalogic offerings, Oracle has expanded its roster of “Engineered Systems” with the introduction of the T4-based SPARC SuperCluster. The SuperCluster consists of a combination of T4 compute nodes, Exalogic Storage Cells (storage and processors integrated to perform low-level Oracle DB storage operations), a general-purpose ZFS storage array and controllers, and a 40 Gb Infiniband network to connect all the pieces. The SPARC SuperCluster is intended to support multiple Exalogic and Exadata virtual images as well as general-purpose workloads such as ERP and other legacy applications.
How well does the SuperCluster work? Oracle’s announcement includes the obligatory vendor charts showing the system is head and shoulders above all competition, but when has a vendor announced a new system without such similar fanfare? However, underneath all the highly integrated superlatives, this looks like a general purpose platform that will attract users of Oracle’s database and other major software products as well as serve as a platform for continued enterprise use of Solaris. I expect that further announcements for T4 systems and the general availability of T4 servers will follow shortly.
What Does This Mean For SPARC/Solaris Users?
This is a major milestone for Oracle and its server community. The virtues of the SuperCluster aside, it is the first tangible product of their commitments to a renewed investment in SPARC processor technology, and as such, it looks impressive. It retains the highly threaded throughput-oriented architecture of the T-series, and makes major improvements in single-threaded performance, which was a weakness in previous generations of T-series technology. But most importantly, it is early, laying to rest the ghosts of previous disasters at Sun and Oracle, validating not only Oracle’s intentions but their ability to execute with this new stream of CPU architectures.
My take is that this announcement goes a long way toward supporting a claim that SPARC is a viable platform for future investment. Committed SPARC/Solaris users can breathe a sigh of relief and shelve their migration plans with a multi-year promise of increases in both CPU and systems performance.
I'd like to hear from SPARC users — does this change your plans?