SGI (NASDAQ:SGI), a global leader in HPC and data center solutions, today announced it will report financial results for the Company’s fourth quarter and fiscal 2010 on Tuesday, August 17, 2010 after the close of the market. A live webcast of the earnings conference call will be held that day at 2:00 p.m. PT (5:00 p.m. ET).

The public is invited to listen to a live web cast of the call on the Investor Relations section of the Company’s website at investors.sgi.com. A replay of the web cast will be available approximately two hours after the conclusion of the call and will remain available until the next earnings call. An audio replay of the conference call will also be made available approximately two hours after the conclusion of the call. The audio replay will remain available for five days and can be accessed by dialing (706) 645-9291 or (800) 642-1687 and entering the confirmation code: 90230680.

SGI claim:

• SGI Altix UV 1000, with 256 cores and 128 JVMs, beat ScaleMP, its nearest competitor, by 82 percent in Java performance with throughput of 12,665,917 BOPS using Oracle JRockit.
• SGI Altix UV 1000, with 256 cores, beat Fujitsu/Sun, its nearest competitor, by 60 percent with a single Oracle Java HotSpot JVM performance of 2,818,350 BOPS.

As Darth Vader says – “Impressive. Most impressive.” – but lets look at those figures in more detail. A 256 core Altix UV 1000 running 128 JVMs? Who on earth would buy a massive single system image machine, with massive shared memory performance, and then carve it up with hundreds of JVMs – which are pegged to cores?

This is nonsense. Far better to pick a smaller machine – say, an Altix UV 100 – which would much more realistically be used for this sort of task. Or an Altix XE cluster, which would give both good parallelism and also high availability.

Joerg Mollenkamp has even more details, comparing these results to those of a Sun T5440. The price difference is extreme, leading to some embarrassing price/performance comparisons, which also highlights how meaningless these sort of benchmark results really are.

There’s no doubt those are some cracking results from the Altix UV 1000. But they’re meaningless, pie in the sky figures, and I can safely say anyone who ordered one of those to run Java apps on would be laughed at, and then fired on the spot.

A far more meaningful result I’d like to see from SGI would be pitching the Altix UV where it really needs to go – into the corporate data centre. Let’s see some data warehouse figures using Oracle and Sybase for big data workloads that can take advantage of all that fast, shared memory.

SGI need to move out of the pure HPC play for their big kit, and add in more sales to big business. (They’ve need to do this since the first big Origins came out, but that’s a rant for another day). This means playing to the kits strength, and posting JVM benchmarks like this accomplishes nothing apart from opening them up to some – very valid – criticism.

Hpctraining.com provides coursework in the following areas:

• System Administration
• Network Administration
• Cluster Administration
• Storage Administration
• Visualization
• Applications/Software Development

One of the key things from my point of view is that the courses aren’t just on-site ones – there are also an increasing number of e-learning courses. For people wanting to break into HPC there are two main barriers – not being able to play with HPC-class kit, and not having the time to learn up about it. Computer based courses and simulators are the way to go for this.

hpctraining.com isn’t an SGI only effort, though. The partner list includes:

• Adaptive Computing
• Atempo
• CAPS
• Intel
• LSI
• Novell
• Octality
• Oracle University
• Panasas
• Platform Computing
• Red Hat
• SGI
• Spectralogic

There’s a lot on offer here and I think it’s a great initiative for existing or budding HPC people. I’m also pleased to see SGI are still offering IRIX courses.

Day-one support for the new AMD Opteron 6000 series platforms is offered across SGI’s entire design-to-order server portfolio, including CloudRack and Rackable scale-out servers and SGI® InfiniteStorage servers. The ICE Cube modular data center also supports AMD Opteron processors for the first time.

It’s not just the old Rackable gear which is getting some AMD love – the ‘proper’ SGI product line is also getting Opterons:

As part of SGI’s increased commitment to AMD processor support, SGI expects to release AMD Opteron processor-based configurations of its Altix® ICE high performance computing (HPC) clusters and Octane III personal supercomputer later this year. Similarly, the SGI HPC cluster software stack will also be available on the AMD Opteron platform for the first time.

No mention of AMD support for the Altix, which is odd. When Silicon Graphics first said they were dropping the MIPS Origins and moving to Intel processors, the first thought was – why not AMD? AMD had a credible NUMA connect – Hypertransport – whereas Intel’s x86 offerings were still stuck with legacy bus interconnects. Itanium was too much of a wild card – but SGI drank the cool aid and embarked on a painful path.

You just need to have a look at Cray, who have managed a successful transition to AMD cores, and done pretty well out of it, to see what might have been. Cray clearly had the better idea – migrate to AMD and Hypertransport, plugging it into their own NUMA interconnect, and then drop in Intel x86 chips when they finally mature.

Nehalem is the long overdue x86 with a sensible NUMA interconnect, and Cray are well positioned to take advantage of the manufacturing scale. SGI’s use of AMD Opterons seems long overdue, and the timing is odd now that AMD seem to be struggling to keep Opterons performing well against the new Nehalems.

Could SGI be hedging their bets, opening up customer choice with AMD at the low end, and seeing how things pan out before plugging Opteron in to the high end Altix? Or are they treading carefully with Intel to secure higher performing Nehalem Xeon chips for the high end?

The Origin 400 looks like a solid bit of kit. Up to a total of 6 blades can be fitted, each one able to take two Intel Xeon quad- or six-core 5500 or the just announced 5600 series. 96gb of RAM per blade for a total of 576GB in the chassis, and space for up to 14 2.5″ SAS drives. The spec sheet also says it supports up to two ‘redundant Ethernet switch modules’, each with 10 gigabit ethernet ports – each blade is specced with 2 GigE ports, so I’m wondering if these ‘switch modules’ are for exporting the storage as NAS.

The specs on each CPU blade sound suspiciously like those in the Octane III – are they interchangeable?

You can view the full press release here and the Origin 400 product pages here.

Over at InsideHPC, John West has managed to dig up some more details – it should be a 6u form factor box, with 6 dual socket blades, and integrated storage presented as NAS.

The name choice is an interesting one. SGI are clearly wanting to capitalise on some of their brand name power – a bit of a risky strategy, though. The Octane 3 was a disappointment in many ways, and the Origin line were also rock solid, scalable bits of kit. I know several people who still have Origin 200 clusters and NUMA stacks in play today, so with that sort of longevity and scalability, re-using the brand name could seriously backfire on SGI if they push out a mediocre product.

Having said that, the SMB push is a welcome direction from SGI. Now that Oracle have finally consumed Sun, we’ve already been seeing Sun’s competent x86 line vanishing from their site – clearly there’s demand for well engineered solutions in this space that go beyond the HP/Dell approach of stacks of ‘nothing special’ 1u and 2u servers.

Through a friend I knew someone high up in COPAN who left a while ago to take up a similar position at 3PAR. COPAN had some clever ideas and some neat products, which sadly wasn’t enough to keep the company going in the current climate.

What’s odd about this is – why have SGI bought the remains? MAID is a clever idea and slots quite nicely into people’s existing HSM offerings. (HSM – Hierarchical Storage Management – is all about long term archival storage, and nothing to do with the appalling series of films from Disney which my kids seem to love so much. I weep for the future etc. etc.).

MAID was a perfect fit for a client I recently worked with who were doing digital imaging archiving – but unfortunately it’s a niche product and the vendors involved seem to miss out on the large bids. (In this case they went with standard RAID offerings from HDS). If SGI are going to try and punt pure MAID storage they’ll be facing similar issues.

SGI do have some solid HSM products – but they also have a very diversified product portfolio, and the company seems to still be struggling to find a unified identity for it’s solutions. Adding another bundle of products into the mix seems pretty risky, especially with the fallout from the latest bankruptcy and the takeover still rumbling on.

Do SGI really have that many customers saying “Hey, I’d like to buy some Altix ICE/Altix CE/Altix UV kit, but you don’t have any MAID storage on the proposal, so we’ll be going to IBM/HP/Dell?”

I could see this working if SGI take the COPAN technology and integrate it into their own products. I’m thinking of something along the lines of the old CXFS kit, which had a couple of Origin 300 heads in the top of the storage rack to handle CXFS metadata. It would require a chunk of engineering resource, and then an even bigger chunk of sales and marketing to shift the boxes – but could SGI pull it off?

The delays in Tukwila have hurt SGI’s NUMA sales, and the new Altix UV gives much better price/performance than Itanium could ever deliver. I predicted (Project Ultraviolet and the future of Itanium Altix) that we’d see a final Itanium Altix using Tukwila later this year, before the Itanium line was killed off.

With no product announcement from SGI to accompany the Intel fanfare, and with SGI’s Cyclone cloud offering accidentally offering a neat Itanium to x86 migration platform, it looks like we could finally be seeing the death of Itanium within SGI’s product line.

Silicon Graphics went there in the past with the R8000 MIPS CPU – it had the potential for massive performance, but only if you optimised the code and really knew what you were doing. That level of investment is always a niche game, and with the lower price and better performance for less effort from x86, Itanium was always going to struggle in the long term.

It’s just a shame SGI had to go bankrupt twice and shed a load of talented and skilled engineers to learn the lesson.

The offering is non virtualised (what’s called “single tenancy”) which addresses the main stumbling block to using cloud compute resources for HPC – the overhead of that virtualisation layer. The other stumbling block – how to actually get your data onto the cloud – is addressed by being able to ship drives of data direct to SGI, who will preload it into your compute instance for you.

On the hardware side, Cyclone offers a nice possibility of “try before buy” for compute customers, with SGI’s entire product range available, packed with some GPU and accelerator goodness:

The SGI technology at Cyclone’s core is comprised of some of the world’s fastest supercomputing hardware architectures, including SGI® Altix® scale-up, Altix® ICE scale-out and Altix® XE hybrid clusters, all based on Intel® Xeon® or Itanium® processors. The hybrid architecture offers either NVIDIA® Tesla GPUs or AMD FireStream GPU compute accelerators for floating point double precision workloads, and Tilera accelerators for integer workloads. High performance SGI InfiniteStorage systems are available for scratch space and long-term archival of customer data.

Itanium and x86 offerings would offer customers a great way to port their apps from Itanium and onto the new Altix UV platform. But I’m sure SGI would never have done that intentionally. Ahem.

On the software side, SGI will be pre-installing many commonly used technical computing applications:

With Cyclone’s SaaS (Software as a Service) model, SGI delivers access to leading-edge open source applications and best-of-breed commercial software platforms from top Independent Software Vendors (ISVs). Supported applications include: OpenFOAM, NUMECA, Acusolve, LS-Dyna, Gaussian, Gamess, NAMD, Gromacs, LAMMPS, BLAST, FASTA, HMMER, ClustalW and OntoStudio. SGI expects to add additional domains and applications partners over time

SGI are mixing it up with Penguin and NewServers, coming in at a higher price but arguably offering more value by pre-loading software, and enabling users to migrate to in-house SGI hardware later on down the line. Costs are also high compared to Amazon, but really, I can’t see anyone putting HPC or technical compute apps on Amazon’s offering.

You can read more in SGI’s press release here.

Revenue is up to over $90m but SGI still posted a loss of $23m. The last few quarters have been spent consolidating operations and reducing costs, so the loss isn’t as bad as it has been.

A quick aside on one of the quirks of the HPC market. Rules mean that SGI can’t book revenue from HPC installations until they’ve been qualified, and if it’s a multi-year deal they have to book those revenues across the year. This always makes the business look a lot more shaky than it is, which is why they also report non-GAAP figures, which get around this.

According to this, revenues were up at just over $150m with a profit of just over $2m. According to details on the call, SGI are gunning for over $500m in sales this year.

The new Altix UV systems have a lot of promise. Now that Oracle have completed their Sun merger, looking at Sun’s future roadmap their involvement in HPC – both at the hardware and software level – looks doubtful.

This leaves a nice gap for SGI to exploit, and at the lower end HP and Dell don’t have much to offer apart from clusters. With new products and a server market that’s starting to pick up, SGI are well positioned to do well this year.

You can read the full details here.