By: Aaron Spink (aaronspink.delete@this.notearthlink.net), August 27, 2014 2:21 pm
Room: Moderated Discussions
Maynard Handley (name99.delete@this.name99.org) on August 27, 2014 1:34 pm wrote:
> To recap:
> The argument made (enthusiastically by juanrga, more moderately by others) is that the
> ARM-64 server CPUs which we should see from various vendors will do well because they
> will offer a compelling performance/power advantage over their x64 competitors.
> The argument made by me is a variant of this which places substantially more stress
> on business issues --- the costs of the CPUs/SOCs to design and then manufacture, the
> costs for which they can be and are sold, the worry that (so far negatively profitable)
> Atoms will cannibalize Intel's higher end if they are improved too much.
>
> With this background, there is a fairly large review of NAS units at AnandTech today:
> http://www.anandtech.com/show/8404/seagate-intel-rangeley-nas-pro-4bay-review
> This is interesting to me in the context of this argument because my argument has long been that ARM will
> get its start in servers at the low-end, in things that many don't want to call servers, like NAS units.
>
As has been pointed out previously...
This is a market that ARM/MIPS/et al, was once dominate and pretty much the only game in town. Pretty much all the low end NAS boxes shipped with some form of ARM and/or MIPS processor. Since Intel released ATOM, there has been an increasing trend of x86 pushing the ARM processors to the lowest of the low end - basically the entry level NAS that you buy if the only thing you possibly care about is price.
AKA this is a market where ARM has been getting hammered for at least the last 5 years. It is following a trend currently that is opposed to your theory. There are several reasons for this.
Even a lowly ATOM CPU tends to provide a significantly higher performance in this workloads, even older Intel SATA chips tend to be much better than anything else on the market, and the software stacks for Intel IO devices tend to be significantly better than any software stack for ARM IO.
A large portion of this is due to the fact that the bread and butter market for Intel cares about things like IO performance, SATA performance, et al. The bread and butter market for ARM SOC has IO almost as an after thought: when was the last time you saw a performance comparison of ARM SOCs in IO? pretty much never right? that's because they basically don't do IO and the levels of IO they do do are generally quite pathetic.
> To recap:
> The argument made (enthusiastically by juanrga, more moderately by others) is that the
> ARM-64 server CPUs which we should see from various vendors will do well because they
> will offer a compelling performance/power advantage over their x64 competitors.
> The argument made by me is a variant of this which places substantially more stress
> on business issues --- the costs of the CPUs/SOCs to design and then manufacture, the
> costs for which they can be and are sold, the worry that (so far negatively profitable)
> Atoms will cannibalize Intel's higher end if they are improved too much.
>
> With this background, there is a fairly large review of NAS units at AnandTech today:
> http://www.anandtech.com/show/8404/seagate-intel-rangeley-nas-pro-4bay-review
> This is interesting to me in the context of this argument because my argument has long been that ARM will
> get its start in servers at the low-end, in things that many don't want to call servers, like NAS units.
>
As has been pointed out previously...
This is a market that ARM/MIPS/et al, was once dominate and pretty much the only game in town. Pretty much all the low end NAS boxes shipped with some form of ARM and/or MIPS processor. Since Intel released ATOM, there has been an increasing trend of x86 pushing the ARM processors to the lowest of the low end - basically the entry level NAS that you buy if the only thing you possibly care about is price.
AKA this is a market where ARM has been getting hammered for at least the last 5 years. It is following a trend currently that is opposed to your theory. There are several reasons for this.
Even a lowly ATOM CPU tends to provide a significantly higher performance in this workloads, even older Intel SATA chips tend to be much better than anything else on the market, and the software stacks for Intel IO devices tend to be significantly better than any software stack for ARM IO.
A large portion of this is due to the fact that the bread and butter market for Intel cares about things like IO performance, SATA performance, et al. The bread and butter market for ARM SOC has IO almost as an after thought: when was the last time you saw a performance comparison of ARM SOCs in IO? pretty much never right? that's because they basically don't do IO and the levels of IO they do do are generally quite pathetic.