14 DAY TRIAL // Oracle did say it intends to release the new SPARC Supercluster at some point in the fourth quarter of 2011 and, for now, things appear to be going smoothly, with no delays in sight.
The supercomputing market is one that, like all other segments of the industry, has been advancing and growing more and more diverse in terms of CPU and GPU computing platforms, as well as their combinations.
The Way Intel Knights Corner and NVIDIA Kepler GPU modules are going to both be used in this upcoming supercomputer is good enough proof of that.
Oracle seems to be going for less of a hybridized design on the enterprise front, with its SPARC Supercluster mission-critical systems.
It intends to use SPARC T4 chips, which is said to be about five times as powerful as the T3, even though it has half the core number (8 instead of 16).
It comes with better per-core/per-thread performance and should reach clock frequencies of 3 GHz or more.
"In Q2, we will accelerate the trend [of transition away from selling low-margin commodity servers to selling high-margin engineering systems] when we introduce four brand-new engineered systems products,” said Larry Ellison, chief executive officer of Oracle, during a conference call with financial analysts.
“The largest of these new-engineered systems is the fault tolerant SPARC Supercluster, featuring our new SPARC T4 microprocessor that runs up to five times faster than a T3 microprocessor it replaces."
The SPARC T4-based supercluster is set to act as a strong rival to existing fault-tolerant systems, performance-wise.
"The SPARC Supercluster runs the Oracle Database applications faster and less expensively than anything available from IBM. Thus, we expect that many customers in our large SPARC Solaris installed base will be upgrading to SPARC Superclusters," added Mr. Ellison.
The company expects the highest demand to come from its existing customers, the ones wishing to upgrade their business critical systems.