Server and chip maker Fujitsu is starting to ship the Sparc-based supercomputer server nodes that are part part of Japan's "Kei" 10 petaflops supercomputer. As El Reg previously reported, back in January 2006 Japan's Ministry of Education, Culture, Sports, Science and Technology (MEXT) hired indigenous server makers NEC, Hitachi …
SPARC64 is dead
What I learned last week at Oracle Open World.
The Fujitsu booth was very evasive about answering the question of:
Will there ever be a SPARC64 VIII?
"VIII is already shipping for high performance computing"
OK, but will there ever be a commercial version for the M-class?
"SPARC64VIII is still in discussion with Oracle"
The Oracle booth was a lot more informative.
=> The future M-Class systems will have T processors in them. The current M-class systems are at the end of their life cycle anyways.
Bottom line: SPARC64 is dead and Oracle might slap an M on future T boxes where they have only 4 cores
Here's hoping it's make commercial.
Well, it's about bloody time there was a SPARC monster in the top 10, Sun never had the resources to do it seriously and it's taken Fujitsu with Japanese government backing (much like the US has backed the IBM Power based supers in the past) to create a next gen SPARC cpu with the grunt and scalability to do it.
I sincerely hope this CPU see's the light of day in the enterprise space, too much investment has gone into it for that not to happen.....in fact I would put money on a spin off of this CPU coming to market soon.
What OS is it running? Or is it not known yet? If it's Solaris 11 then we can be pretty damn sure we will see boxes based on this cpu in the field in the next couple of years.
The OS will be Linux, and managed by software stack from Platform Computing
SPARC64 is DEAD
Here is the official SPARC64 roadmap.
Clearly SPARC64 is DEAD
@ Sales Drones
Just because future M series will have T series CPU's in them, after Rock technologies are merged into the T series processors, does not mean the M series platforms get a 4 core T series (that we have today) stuck into a socket, which such a statement infers. Oracle roadmapped faster single threads and fewer threads in future T processors.
If SPARC64 VIII is shipping today, SPARC64 is not dead.
If Fujitsu has not yet updated a SPARC64 roadmap image where a current shipping processor is not included, it does not mean SPARC64 is dead.
If Fujitsu is still in discussion with Oracle for Oracle adoption of the latest SPARC64, then that is the fact, nothing more, nothing less, certainly not dead.
When a shipping processor in a particular market segment is called dead by opposing sales drones, you know those sales drones are Drinking the Kool-Aid, and are probably not long for life in their jobs.
SPARC64 VIIIfx is not the same as SPARC64 VIII....that chip was only made for the Japense government and has nothing to do with commercial systems to run Solaris.
"....that chip was only made for the Japense government and has nothing to do with commercial systems to run Solaris...." Yes, but Fujitsu are talking with Oracle now about that new chip into Solaris servers.
Not much talk
Oracle has been "talking to Oracle" since February.
I think Gartner has SPARC VIII at .10 probability
- August - Roadmap has T chips replacing SPARC64 in the M-class
- September - Oracle world and no word, not even the 3GHz 12MB cache chip
Face it SPARC revenue is down 50% and the return does not justify the investment, especially since Larry has no interest in selling OEM gear and the rumors of .5 core factor change for existing and prior systems didn't happen because the profit lose does not justify the action.
It's a shame that Oracle does not get it that only the chip is made by Fujitsu for systems in the US.