Facebook is looking for an ARM SoC server hardware engineer
facebook.com
facebook.com
Some possibilities:
- FB is designing some non-compute datacenter ARM devices. For example, environment monitoring, or a control plane processor for a network switch. (Most likely IMO.)
- FB is designing/bringing up mobile ARM devices, and the references to the datacenter are a smokescreen to keep the project under wraps. Even if they aren't taking the lead on the design, it's very useful to have engineers at this level eg to audit/collaborate on the work of a partner like HTC. Personally I think this is only a matter of time, if it hasn't long since begun.
- FB is building ARM servers. Edit: Originally wrote "very unlikely", but perhaps more plausible than first thought given FB's public statements on the matter (eg [1]).
[1] http://www.zdnet.com/facebook-stretches-arm-chips-in-datacen...
[0]: http://h17007.www1.hp.com/us/en/enterprise/servers/products/... [1]: https://cards.linaro.org/browse/CARD-327
> Why do you think that third situation unlikely?
Using a non-Intel architecture for general-purpose compute slices seems like, at minimum, a huge toolchain investment.Things are probably better now than they were five years ago (eg there seems to be LLVM support), but even assuming toolchain parity we're talking about having two flavors of each binary to maintain; a new constraint to consider when scheduling jobs; etc. (Edit: The HHVM ARM target is interesting, and mitigates some of these).
It's not impossible, but IMO pretty unlikely compared to the other more "traditional" possibilities.
There is quite a lot of assembly code in open source stuff, there was a survey here for the arm64 porting https://wiki.linaro.org/LEG/Engineering/OPTIM/Assembly although much should be removed... (the 64 bit port could be what Facebook is interested in and thats way less supported...)
You're definitely right there is a large software maintenance cost, but that has to be weighed against any perf/watt savings (Anandtech's Calxeda review shows even early systems have potential http://www.anandtech.com/show/6757/calxedas-arm-server-teste...), being able to put pricing pressure on Intel, and not being tied to Intel as a single supply source.
Also some reporting on Calxeda, who is ostensibly working on ARM servers: http://semiaccurate.com/2013/01/16/facebook-goes-arm-with-ca...
How wrong was i. Consistently one of the best feeds in my feedly these days.
I think though that it seems HN readers don't read semiaccurate because HN seems to be a software focused site, so hardware links probably don't get as many votes or people don't think to post them. Or maybe people just aren't as interested in hardware here?
Honestly, FB is doing an awful lot of development and building around a product of declining value, as opposed to innovating on that product.
The product itself is a vertical silo of information which makes one feel completely constrained.
No matter how many DCs they build (with US data access/control compliance), no matter their supporting of the OpenCompute platform, their attempts to emulate mobile innovation... the core of FB is a vertical column, groups and topics are utterly broken and they have eroded a lot of trust (what little remained) from their userbase.
Honestly, I just don't see any external value they can maintain.
Also, I suspect that work about wimpy cores was done for a reason.