Sun, Fujitsu juice entry Sparc box

Sun Microsystems and its Sparc chip and server partner, Fujitsu, have have upgraded the entry Sparc Enterprise M3000 rack server with a peppier Sparc64-VII chip. The M3000 was not part of the original rollout of midrange and high-end Sparc Enterprise servers that the two companies brought to market in April 2007 with dual-core …

[Anonymous Coward]

I will not buy SPARC64 until I see a public roadmap

Does the new minor chip speed finally get the core performance back up to the dual core chip?

Not that it matters as SPARC and Itanium are end of life in our mind.

The current roadmap shows the SPARC64 will cease to exist after the current generation. This is just like UltraSPARC V. Deja Vu all over again. There are a few minor tweaks but the chip will not be substantially improved and will never be on 45nm.

We have put SPARC64 on the divest list and are moving everything to either Nehalem or Power.

The best I heard was Fujitsu might be able to have 45nm in 2012, but that will only be an HPC chip.

Allison

[Neil Davis]

Re: Anonymous Coward.

Allison,

Sign a NDA with Sun and they will gladly show you the SPARC roadmap. Roadmaps from Sun are never public....

[Anonymous Coward]

NDA roadmap is a BS roadmap

I don't believe/trust/give credibility to roadmaps that are not public.

It's really Fujitsu's roadmap anyways that is the question.

http://www.fujitsu.com/global/services/computing/server/sparcenterprise/technology/protection/roadmap.html

There is NO SPARC64 VIII.

There is only a SPARC64VIIIfx high performance chip which is a huge chip

[Matt Bryant]

RE: Neil Davis

".....Sign a NDA with Sun and they will gladly show you the SPARC roadmap....." And what would be the point given that Sun are on the point of being borged by Oracle, which means everything could change in less than six months.

/Point, laugh, repeat.

[David Halko]

Anonymous does not "believe/trust/give credibility" to his/her own statements

Anonymous Coward posts, "The current roadmap shows the SPARC64 will cease to exist after the current generation."

Neil Davis posts, "Sign a NDA with Sun and they will gladly show you the SPARC roadmap. Roadmaps from Sun are never public."

Anonymous Coward posts, "I don't believe/trust/give credibility to roadmaps that are not public."

Smells like FUD. Let's think about this exchange...

If Anonymous Coward does not "believe/trust/give credibility to roadmaps that are not public" and "roadmaps from Sun are never public" - then Anonymous Coward does not "believe/trust/give credibility" to any of his/her own statements regarding "the current roadmap shows the SPARC64..."

If Anonymous Coward does not "believe/trust/give credibility" to his/her own statements, then no one else should, either.

Best to get an NDA to listen to the vendor, instead of believing an Anonymous Coward who does not believe himself/herself.

[Anonymous Coward]

nice try at humor

David....it was Neil who said Sun roadmaps are not public....not AC/Allison

nice try though....find us a Fujitsu roadmap that has a SPARC64 VIII that does not say "uncommitted"

Allison

[Anonymous Coward]

Pricing

Just a note ---

I know of no project I have been on where the Sun/Oracle/Fujitsu servers were sold for anywhere near the list price.

[Anonymous Coward]

@Anon can't read --- nice try at humor

David typed --- Neil Davis posts, "Sign a NDA with Sun and they will gladly show you the SPARC roadmap. Roadmaps from Sun are never public

Anon typed --- David....it was Neil who said Sun roadmaps are not public

Duh! David clearly typed it was Neil - can't read, I guess

[David Halko]

A little reason.

Anonymous Coward posts, "We have put SPARC64 on the divest list..."

Neil Davis posts, "Sign a NDA with Sun and they will gladly show you the SPARC roadmap. Roadmaps from Sun are never public."

Anonymous Coward posts, "I don't believe/trust/give credibility to roadmaps that are not public."

Anonymous Coward posts, "find us a Fujitsu roadmap that has a SPARC64 VIII that does not say 'uncommitted'"

Several points of reason here:

1) I don't remember someone saying to Intel that they were going to drop HP when they did not commit to building systems on the Intel Pentium Pro socket/platform from Intel. Such a statement is clearly as silly as Sun not committing to support a socket profile for 5 years.

2) You can get your own NDA if you are really a customer and see the roadmap. Clearly, you are not a customer, divesting hardware, or you are such a low level analyst that you don't have access to the information.

3) Providing an NDA road map of SPARC64 that you do not consider credible is a waste of time your time and the providers time.

Serious now... An anonymous poster, making silly statements, who appears to be a low level analyst (or someone who can't tell the truth), who asks for information which has been clearly communicated that he will not consider credible... is not going to make a significant decision regarding platform architecture.

[asdf]

all a bunch of sales drones

Nice to see the usual suspect sales people from the large Unix vendors all flaming each other anon. The fact is from here on out SPARC will die a slow quiet lingering death ala Itanium (where are those four core 45 nm Itaniums Matt?). I know my current large employer who are bunch of Sun fanbois are migrating all their production software from SPARC to x86. In fact that is what I was working on this morning.

[David Halko]

RE: all a bunch of sales drones

> Nice to see the usual suspect sales people from the large Unix vendors all flaming each other anon.

Yes, nice to see the anonymous suspect sales drone asdf!

Just got another SPARC about 2 weeks ago...

[Anonymous Coward]

IBM POWER6

I work at a large international Fortune500 telecom company and we have lots of IBM POWER P570, they are rock stable and fast as hell! As sysadmins, we love them. Our manager were recently convinced to try out Sun Niagara T5220 and on certain multi threaded workloads they were in fact, several times faster! We are now migrating some of our work load from IBM POWER to Sun Niagara. The list price on IBM P570 is $219.000 and we get seven Sun T5220 for the price of one P570! For that performance/price ratio, we can sadly not justify buying IBM POWER anymore. As Oracle 11gi is being prototyped and developed on Solaris because of DTrace (and then ported to other platforms), we think that Solaris will be the premiere platform for Oracle in the future. In this time of crisis we must save money and expensive machines like POWER have to go.

[Anonymous Coward]

T5220 = ($46K * 8 * .5) * (1.22) = $224K for Oracle license

Quarter million for a two socket box with average utilization of 15%.

I think you should call you IBM rep and discuss TCO and roadmap comparisons.

[David Halko]

T5220 = ($46K * 1 * .5) * (1.22) = $28K for Oracle license

Anonymous Coward posts, "Quarter million for a two socket box with average utilization of 15%."

If your application is only going to consume 15% of the total platform capacity, then run the Oracle database in a Capped Zone using 1 CPU at 100% utilization. (Unlike other operating systems, Solaris is very capable of running at 100% utilization without application failure.)

Your application has another 7 cores to play with! If you application can not leverage the capacity, then consolidate some more servers onto the platform and save some more money!

Anonymous Coward posts, "I think you should call you IBM rep and discuss TCO and roadmap comparisons"

Obviously, talking to the IBM rep would know more than an Anonymous person posting dubious information. He could have saved such a customer an order of magnitude by suggesting a Capped Solaris 10 Container and allowed them to spend $28K instead of $224K!!!

The free CPU capped hard partitioning, included in Solaris 10, gives Solaris customers many more options than the common reader may be aware of.

http://netmgt.blogspot.com/2009/03/partitioning-oracle-licensing-terms.html

Someone suggests paying for CPU capacity that they are not using, like the situation you posted, they you should kick them out the door... find someone who: went to University, knows how to read [the Oracle licensing guidelines] and understands basic operating system features.

Oracle will use lots of CPU threads, depending on how you tune it. We have applications which easily spread to utilize 24 threads with little tuning, but for the cost reduction, it is sometimes advantageous to reduce your footprint, so instead of using 64 threads, you leverage only 8, and run at a higher utilization. This is a fine example.

[Anonymous Coward]

use only 1/8th of a chip for Oracle and pay $28K?

Sun's answer to Oracle pricing is a joke.

It's bad enough the T processors were never make for "Data workloads" but to suggest you should only use one 1/8th of a chip to reduce the Oracle pricing is insane.

The core is a UltraSPARCII core with Keep It Simple architecture...basically they ripped out everything that would help a database.

If you want to write an interesting reply answer the question about how T2+ will not be completely eliminated by Linux and Nehalem EX. (let alone Larry)

Did you light your candle and say goodbye to Sun yet?