sounds great, didn't I read that ARM was backing off on server though (hoping it was a bad article).
edit: thanks, you all nailed it. I remember now and it was Qualcomm. Whew, I'm glad as I was bummed that this meant there was a good step for a piece of tech that would be DoA. Really wish Qualcomm wasn't abandoning based on how it appeared they were making progress as well to be a good competitor (and competition has been great where applied).
Not Arm, but Qualcomm. It was also a strange and sudden decision. I don't think they really wanted to do it, but might have been forced to do it by the circumstances: anti-trust lawsuits, Apple dumping them, Broadcom attempting a hostile takeover, U.S. government pressuring them not to sell to Broadcom no matter what, etc.
One or all of these may have forced Qualcomm to cut its losses. Either way, I hope they get to sell the Centriq line to some other company that can provide competition to both Cavium and Intel/AMD/IBM in the server space. We need more than one Arm chip to compete in the space, otherwise Cavium's chip could also be seen as some kind of outlier and be ignored by the industry.
> One or all of these may have forced Qualcomm to cut its losses.
Yeah, I think the LBO really scared Qualcomm. Their royalties fund a lot of the other work in the company and I think they've learned that it also makes them a target. A domestic LBO could happen and then there'd be no escape hatch. They pledged to cut costs when Broadcom announced their intentions, this may be a part of that.
I don't think of them selling the server chip business as "cutting its losses" -- instead I would call it narrowing its focus back to businesses closer to its core competency. The Centriq launch with Amberwing seemed to be relatively positive for a first design release. They didn't turn the industry upside-down but they probably achieved most of what they set out to do. It will take some time to convince server customers to switch, recompile their code and get their vendors to recompile their code.
> We need more than one Arm chip to compete in the space, otherwise Cavium's chip could also be seen as some kind of outlier and be ignored by the industry.
Agreed. There's momentum that's been built [1,2] and I'd hate to see it flub now.
I hope they do something with it though. Whether sell it to someone who can be a good caretaker to it, share lessons and designs with other companies that could then build upon it, spinning it off, whatever they choose. Just abandoning it outright seems like an unfortunate move. I won't say bad, because I'm sure they have their reasons, but unfortunate in that it seemed like it was on track to be viable.
Microsoft's Surface Pro line, to me, was kind of an "eh, that's nice I guess" up until the Pro 3 at which point it seemed like a platform that had hit its stride. The stuff I was reading on Amberwing seemed like it wasn't a silver bullet, but what it did work for, it worked very well when comparing price vs performance so one could hope that with time, they could make it even better.
edit: thanks, you all nailed it. I remember now and it was Qualcomm. Whew, I'm glad as I was bummed that this meant there was a good step for a piece of tech that would be DoA. Really wish Qualcomm wasn't abandoning based on how it appeared they were making progress as well to be a good competitor (and competition has been great where applied).