Sun & Fujitsu Team On SPARC Chips & System

An anonymous reader writes "Sun and Fujitsu just announced a 20-year partnership to jointly develop SPARC based technology and systems. It looks like the long-predicted partnership that was hinted at earlier has finally come to pass in a much more comprehensive manner than I've heard anyone predict, i.e. not just chips, but a unified range of systems. My guess: Sun drops Ultrasparc III to provide the Throughput computing chips for the low end / web / network stuff, and takes up the Fujitsu provided SPARC64 chips for the high end and workstation market. Will this spark a new RISC renaissance for Sun and Fujitsu? Or is it a last gasp before Opteron / PowerPC / Itanium crush them? I for one will be interested to see what systems and processors come out of this. This could really revitalize the SPARC system market, especially if Sun's work on Throughput computing proves out."

Re:20 years?

[syphoon]

You were misled by the OP, but RTFA please. The press release said they're expanding their relationship that's already existed for 20 years. Not that they're announcing a 20 year partnership.

Re:20 years?

[Ratbert42]

NEW YORK, NY -- July 8, 1987 -- Sun Microsystems, Inc., introduced today the Sun-4 family of 10-MIPS supercomputing workstations and servers that give users the performance of a VAX 8800 system at one-tenth the price.
...
Sun also announced that it will license the new SPARC architecture... SPARC licensees announced today are Fujitsu Microelectronics, Cypress Semiconductor, and Bipolar Integrated Technology.
...

Re:What's actually going on here...

[turgid]

You run Solaris on SPARC processors. Solaris is highly multi-threaded from the ground up. It's extremely fine-grained. It also has some sophisticated algorithms for migrating threads to the most appropriate processor based on things like memory locality and load. (Forgive me if my terminology isn't terribly accurate, I'm not an OS kernel expert). The Niagra and ROCK processors are designed to execute highly multithreaded loads. Fujitsu SPARC64 is more traditional, in that it is designed for loads with fewer concurrent threads. By adopting Fujitsu's high-end gear, Sun gets performance on less thread-intesive loads too. Now Sun and Fujitsu have a horse for every course, so to speak. If I were HP trying to sell itanic boxes, and cranky old (soon to be exterminated) PA-RISC kit, I'd be very worried.

Re:Wonder how this fits into the free hardware

[hutkey]

no problem for Fujitsu [fujitsu.com]. they are customer oriented [fujitsu.com] people.

so they will not find it diffcult to adapt to the innovative "giving away hardware" teminology, i guess. they will concentrate more on what happens after customers get free hardware.

Re:Crush Fujitsu... maybe.

[thesupraman]

Sigh.

Please try to remember that entry to the 'Top 500' list is as much about your interconnect topology and technology as the capabilities of the processors used.

It is a measure of one, and exactly one benchmark, LINPACK [top500.org]

Machines which are not well suited to this benchmark, or do not have network technologies/topologies well matching linpacks requirements will perform poorly at it, but possibly very well for their chosen purpose.

Good examples of this are the WETA digital clusters used in parts of the LOTR films, which are great for rendering, but hampered seriously in their linpack result by their 100MBit standard ethernet connections.

Another good example of this is the Virginia Tech G5 cluster, which gets a LARGE boost from it's infiniband interconnects (well, it will when Apple finish giving them the new machines... eventually..).

Not that I am defending SPARC's rather lackluster performance these days, just making a rather important point.

Those SPARC boxes better get a LOT cheaper VERY fast if they intend to find any real home in HPC.

Re:What's actually going on here...

[Anonymous Coward]

>If I were HP trying to sell itanic boxes, and
>cranky old (soon to be exterminated) PA-RISC
>kit, I'd be very worried.

I'm sorry, based on what you posted, you seem
to know something of Sparc hardware and
software, but even if what you say is true,
it's still just another crack pipe dream.

Why? Because business want platforms that
a) Aren't expensive
b) Can run their software today

There really aren't that many apps written
today that have a mandatory need for
"some sophisticated algorithms for migrating threads to the most appropriate processor based on things like memory locality and load"

I really don't think Oracle does; Microsoft's
software doesn't (doesn't even run on sparc);
I don't think OpenOffice needs it. What's left?
Apache? Apache runs just fine on commodity
hardware.

To use a simile, if your analysis is correct,
sun's os+hardware is like buying a $300,000
Feraris for a 5 mile daily commute.
Only an idiot with more money than common
sense would spend that much money when
they could spend $10,000 on a cheap and reliable
commuter car for the 5 mile daily commute
and spend the rest of the money on a house/vacation/girlfriends/etc.

other market

[millahtime]

They have another market in high end engineering desktops. For people who design chips and other detailed components and need to simulate them there is still a market for their work stations.

Now, the other chips are catching quick on this so they need to stay ahead or they could loose that market too.

Re:sun problem

[grigori]

Baloney. I've had power drop out and kill Sun machines and they just come right back. Are you talking about file system check? Just turn on logging in /etc/vfstab and even that goes away, just like the same reason you use ext3 instead of ext2 on Linux.

Re:sun problem

[larien]

I'm with grigori; never had a problem with power failures on a properly configured Sun box. The only server I know of which has a problem is the E150, which you can't power on from init 5 without either a keyboard or a screwdriver. The E150 is probably the worst put together piece of hardware I've ever seen and it astounds me that Sun released it.

Besides, power failures shouldn't happen; you should have UPS on all important servers so power failures shouldn't be a problem at all.

Re:20 years?

[Dj]

And you seem to not do your research.

http://www.wordiq.com/definition/Fujitsu [wordiq.com]

The company was established in 1935 under the name Fuji Tsshinki Seiz, a spinoff of the Fuji Electric company, this in turn being a joint venture between the Furukawa mining company and German conglomerate Siemens.

Or how about more obviously....

http://pr.fujitsu.com/en/profile/profile.html [fujitsu.com]

Fujitsu is a leading provider of customer-focused information technology and communications solutions for the global marketplace. Since Fujitsu's establishment in 1935, we have maintained a commitment to cutting-edge technological innovation and uncompromising product quality.

So only 50 years out there old chap. :)

Re:other market

[afidel]

They already lost that market. I know quite a few engineers that two years ago couldn't wait for the Opteron to ship so that they could get a cheap fast system to do design work on. They absolutly HAD to have 64bit support as their chip routing would often take 8-12GB of RAM to run in a single process but they were tired of paying SUN prices. When you can get a dual Opteron system with 16GB for less than the 16GB RAM upgrade from SUN you can see why they have lost the market.

Here's some more inside information

[Anonymous Coward]

First, check out my previous article [slashdot.org]. I must write anonymously because I currently work in the server division.

To put the matter simply, what killed Sun Microsystems is a pathetic engineering team in the microprocessor division. With the exception of the UltraSPARC I and II, all the other processors were poorly designed and managed. What is unique about Sun's microprocessor division is that the managers consistently and actively hired H-1B workers from Taiwan and India. Foreign engineers were the rule, not the exception.

Not so with Fujitsu. Admittedly Fujitsu had a similar problem with the SPARC64-I, SPARC64-II, SPARC64-III, and SPARC64-IV because all these processors were developed at HAL in Campbell, California. HAL also hired mainly foreign engineers, and previous generations of SPARC64 sucked. Then, Fujitsu became tired of this nonsense, shutdown HAL, and fired everyone. Fujitsu then developed the SPARC64-V entirely in-house, using only Japanese engineers. No foreign engineers.

There is a myth that, somehow, tech companies absolutely need H-1B workers. Well, now we have yet another example of why that myth is just a myth. SPARC64-V built by native engineers crush UltraSPARC III built by H-1B engineers.

To understand how pathetic Sun's microprocessor engineers are, we in the server division actually had servers ready to accept the new UltraSPARC III by the end of 1999. Unfortunately, the processor team was two years late. So, our test machines sat idle.

Note that the server division is not dominated by H-1B engineers. The server division and the microprocessor division are two different worlds: first world versus 3rd world.

I, for one, am glad that we are relying on Fujistu. Its processors are much better designed and built than Sun's own processors. I am glad that Fujitsu will soon OEM high-end servers to Sun. Sun will stop designing and building high-end servers in 2006. (I work on the low-end servers running x86.)

I simply do not see Niagara and Rock as the savior of the company because those designs are well-known public knowledge. Check out Professor Kunle's Hydra work: it is 70% of Niagara. Intel has now embraced the Hydra work and will produce an x86 chip based on Hydra.

Here's a dumb question: Which company will build the fastest, highest performance multi-core chip based on Hydra? Intel or Sun?