Hacker News Reader

Comment by imaginenore | Hacker News Reader

The price is the most impressive. $27 in bulk. $100 buys you 60 cores.

I would love to see some Blender render benchmark running on it.

The $27 price is for a dual core chip, not a 16 core one. If Avoton prices are anything to go by then expect higher core count chips to be considerably more expensive. Intel is always careful to make sure Atom products don't cannibalize their higher margin Core CPUs so don't expect these to ever be a win from a CPU perf-per-dollar standpoint.

Intel Atoms are utterly useless for that. I'd wager they are useless for most anything that isn't memcpy-bound, so I wonder what makes them put 16 of them into one package.

Per core they might be useless, but that's a lot of cores. So per dollar it might actually be good.

The intel Atom C2750 keeps par with my AMD Athlon II X4 640.

Fun fact: I was running Blender in non-GPU mode on far lower hardware than that almost 15 years ago.

For memory-bound apps are they sort of useful? For instance, A packet forwarding system that reads a packet, does 3 main-memory lookups, then forwards or drops?

Now I'm wondering how one of the 16-core chips would handle VyOS...

Won't be any faster than what you have today.

Right now I've got a little EdgeRouter X - it would probably be faster.

Hard to compare against the Cavium Octeon. x86 general purpose vs MIPS64 with dedicated offload hardware...really depends on the use-case. I have an ERLite and AFAICT it competes with stuff way above it price-wise, e.g. https://store.netgate.com/ADI/RCC-VE-2440-board.aspx

I'd rather run PFSense, but it's not worth an extra $200+ for just my home use.

If you do it right, you don't have to the lookups in main memory.

We have a c2758 doing 12mpps l3 routing (full BGP table) with 50 ACLs on one of these: https://netgate.com/products/xg-2758-1u.html

The c3000 prototypes are due in a couple weeks.

Well you have up to 16 cores fighting over one 64 bit memory bus.. so this is mostly aimed at packet-processing and filer workloads I'd say.

Reply on news.ycombinator.com