Sony bails on the Cell-BE? [Updated]

Joe over at Scalability.org points to an article in the Inquirer reporting yesterday that

SONY is planning to sell of [sic] the foundries in which it pokes its Cell processors, according to reports out of Japan.

As Joe points out, the Cell has potential for HPC, and he wonders about the viability of the market for non-GPU based accelerators

We know about where the price point needs to be (and nVidia is hitting it). Since Sony looks like it wants out, what does the future hold for Cell-BE? Even if we could get some of the monied classes to want to fund a company building accelerators, this sale would be sure to raise eyebrows on any company wanting to build Cell-BE based accelerators. I don’t see IBM doing it any time soon, nor Mercury.

GP-GPUs are hard to use right now, but the HPC community has already demonstrated that they’ll make big sacrifices if they can save enough money by doing it.

I agree with Joe that some kind of acceleration/co-processing technology is probably here to stay. The only way that non-GPU technologies will survive in the long run is if they can build enough market to be price-competitive with the video guys.

There is a model for this: ClearSpeed looks to have taken the first step when they licensed their HPC accelerator to BAE for use in its satellite (story and podcast covering this topic as well).

Update: Today Sony denied reports in the Japanese press that it’s looking to sell its Cell fabs.

Sony has been saying for months it’s exploring ways to streamline its chip business as it focuses on home entertainment devices. “But nothing concrete has been decided,” Sony spokesman Tomio Takizawa said.

Comments

  1. Might have been a cut and paste error, as the the original un-edited version on my site reads

    “a report noted that Sony may be trying to sell off …”

    Regardless, it is good to see Sony respond to this.