Larrabee is dead! Intel has announced one of the most ambitious graphics ventures’ demise on a blog post by Bill Kircos, Intel’s Director of product and technology . Intel didn’t announce it explicitly though, but fFortunately, Ryan Smith at Anandtech has deciphered it for us.
Four years ago rumours surfaced that ‘a shadowy organization called Larrabee Development Group‘ had launched itself to do the unthinkable in the High-end graphics chip industry. It had decided to take head on the two big crocodiles of the pond, namely NVidia and AMD. However, Job vacancy postings on Intel’s website did little to hide that Larrabee was an Intel venture. After a few months Intel proudly announced launch of Larrabee, a multi-core processor design which was supposed to compete with other GPGPU based (General Purpose Graphic Processing Unit) future products from competitors NVIDIA and AMD. Larrabee was supposed to be something like a combination of GPU/CPU. A chip that would have a full programmability of a CPU and throughput computing feature of a GPU. (see image below)
However, Intel’s foray into producing a Multi-core GPGPU Chip didn’t bother Nvidia and AMD much. On the contrary their strategies were not at all affected by Intel’s apparent plans. Both of them rather steadily went ahead with their strategy of Integrated Graphics.
After missing their initial product launch deadlines, last December Intel delayed the Graphic Processor launch and decided to downsize Larrabee to Software Development Platform. The product was now slated to arrive sometime in 2010. However Bill Kircos in his blog at Intel’s website has discretely written, “We will not bring a discrete graphics product to market, at least in the short-term.”
This means, Larrabee – the product is not coming to shelf. This case brings us back to the issue of abandoning innovation that was discussed here at Creatologue in the light of Microsoft’s abandonment of Courier. However, Larrabee’s case is quite peculiar. Larrabee, if it would have been launched would have been quite a radically innovative product. As is shown in the image below.
What Intel has done by abandoning Larrabee is effectively imitating its competitors by focusing on Integrated Graphics rather than thinking about a CPU/GPU hybrid.
It would really be interesting to explore when a company might abandon a radical innovation project? One obvious reason was that it wasn’t giving expected results. But then, in an uncharted technology category, it will always be difficult to estimate performance. On the other hand, while its competitors in the graphics processing segment, have a strong presence, especially NVIDIA being stronger in gaming segment, overall Intel is incomparably bigger than these competitors. Given it’s size Intel shouldn’t worry about economies of scale or scope. The only other probable expectation is shift in strategic importance of the innovation (Agarwal, Barry and Tripsas, 2005). Why such a touted innovation lost its importance before it could see the light of the day, only time will tell.
Agarwal, Rajshree; Bayus, Berry & Tripsas, Mary. 2005. ‘Abandoning Innovation in an Emerging Industry. ‘ Working Paper.