Sun CEO shows off Rock ahead of Fujitsu launch
Could an UltraSPARC chip arrive on time?
So, it's April 2007 and Sun Microsystems has just popped one of its 16-core Rock chips on CEO Jonathan Schwartz's desk.
Schwartz posted pictures of the Rock silicon on his blog, bragging that "the chips are running billions of instructions already".
Sun's customers must be encouraged by the Rock display, having suffered from about five years of delayed UltraSPARC chips. Servers based on the Rock family – Boulder and Pebble – should begin shipping in 2008.
“Rock will allow us to build coherent systems supporting 256 terabytes in a single software domain (plain vanilla OpenSolaris, no less),” Schwartz wrote. “That is an awful lot of RAM in a single system (and given the cost of memory nowadays, you'd want to post an armed guard next to that machine).
“Rock is 16 cores - we haven't said how many threads per core. Nor have we said why this chip heralds the golden age of effortless parallel programming, or how it brings fault tolerance to the masses. But stay tuned, I think we're planning on talking up both in the next few weeks.”
In the past, we've disclosed that Rock will run two threads per core, giving it 32 threads per chip. In addition, the 256TB of memory described by Schwartz would come via a 8-socket box that holds 512 DIMMs, depicted here. Our sources have indicated that Sun stopped work on systems any larger than the Platinum box.
Sun looks set next week to announce more details on its tie-up with Fujitsu around SPARC64. Those boxes will be on sale for about a year before Rock-based gear arrives, which means many Sun customers will pass on the Fujitsu transition.
The rumor mill says that Sun is in the market for a new chip production partner with TI declining to invest in a 32nm fab. The same rumor mill continues to tell us that Power6 is suffering from serious yield issues and that some redesign work is in order. You know where to reach me. ®
Sponsored: Benefits from the lessons learned in HPC