Last week word started to surface about changes afoot at SGI. VizWorld (as far as I know) broke news from inside the company that the entire graphics division had been eliminated. This group was mostly focused on the VUE suite of efforts, but there was a little hardware in there too. By Monday I had talked to even more people, well enough placed to confirm the story. When I wrote my commentary on Monday I said that I had emailed the company and was waiting for official word.
I was told via email that official comment would come in the form of Barrenechea’s blog, and late last night a new post did appear. In 932 words Barrenechea manages to not mention VUE, whether the company has entirely left vis or what its future direction might be, and does not acknowledge any reorganization or elimination of employees, either to confirm or refute the rumor. This is a problem.
SGI, the brand that Barrenechea has adopted for his new company, made a promise to the HPC community in the form of a line of visualization technology that would serve as the hub of pixel delivery for a wide variety of customer solutions, local and remote, from a tremendous variety of sources. I sat in many briefings on these technologies, and the company was 100% committed to getting the technology out.
In his post Barrenechea mentions the question of trust
Our customers have depended, and can continue to depend on SGI to provide leading edge systems aimed at collaboration, remote visualization, and decision support.
According to the post Barrenechea’s company is now a believer in GPUs, and so that’s where their focus is going to be. He characterizes this change by SGI with the jaunty phrase “shift happens.”
The problem here is not that the company has changed direction. It is hard to argue that the old SGI didn’t need retooling, and in that process a “not doing list” will surely have to be executed. The problem is that the company is not communicating honestly with its customers or the community. Both current and former employees say that the entire division was eliminated and people fired. Major programs were considering how to incorporate the VUE technology into their technical direction and are now left wondering what to do, or whether the technology even exists. We know from public filings in the bankruptcy case that all of the vis IP (including VUE, presumably) was transferred into the holding company, but was does that actually mean?
SGI does have to change. They are probably supporting much too broad a mix of products right now for a company with their revenues. They are presumably investing big money in bringing UV to market. They don’t have clear messaging. They are losing people.
This is a lot of change, and there is more to come if the new company is to remain in business. But if SGI wants its customers to continue to “continue to depend on SGI” for solutions, it needs to be up front and open with its customers and what it’s doing, and why.
well said. Thanks for taking a hard look at this. SGI is in sad shape when it comes to rebuilding its corporate reputation – and this is exactly why.
I certainly agree with the view in this article, however, you figure that this would have just been a natural shift in the SGI strategy. If you think about it, even though SGI made its in-roads as a graphics company, they were slowly becoming irrelevant as that kind of expert as more commodity hardware began catching up. You can’t just sit on your architectures the way SGI did and hope to turn a profit because some day, someone might just buy hardware because of the SGI name. When I was doing computer graphics 5 years ago, the debate was within the community as to whether to go with an SGI or a regular PC. Almost certainly, people were just going to buy Boxx or Dell or HP simply because for the money, they can get in some cases, an order of magnitude, a better computer.
Shift does happen. It happened in such a way that SGI’s IP started becoming irrelevant and they couldn’t innovate fast enough to keep people interested. The only thing left really is the name. While its IP was worth enough for Rackable to change their name to SGI, is it really enough? I think the decision to drop the graphics division as it was came long overdue. The other graphics powerhouses in the industry that do just that as their core business now have real processing solutions. At least SGI recognizes this and is focusing now towards the utilization of COTS GPUs as part of their portfolio.
Chris – Yes, I agree this was probably well advised, as it was when they got out the first time in 2006. But to get back in in a big way, and then leave by sneaking out the back door without telling anyone is a bad way to behave. Well, says me anyway.