“Power8”, it’s starting to
“Power8”, it’s starting to piss ME off!
FTFY!

ISAs! ISAs! ISAs!

No what needs to be done is
No what needs to be done is just erasing the whole thing, and you are the one who set yourself up for the duck jokes, that and your potshots, at any other ISA you fill that threatens you masters edicts. You with your accusing anyone who even mentions another ISA, or RISC based microarchitecture, as a hater. And ARM, and MIPS were also mentioned, but the Power8s are coming, and they are licensable IP that can be fabbed at cost, for Google(Google has a license), and not cost $4000.00 a pop! This scares the hell out of the hand that feeds you, the one that fears competition in the marketplace most of all. Those Google licensed Power8s, and ARMs(Custom made for Google ARMv8 ISA based server chips), are going to make the high end x86 products one hell of a lot cheaper, for those that will need them, and maybe the Tyan produced ATX motherboards for POWER8, will lead to Power8 based high end gaming, at AMD style low prices. So MIPS ARM Power8 to you, and you’re the one that is screaming that x86 sucks, CISC has a hard time competing against RISC, and by the laws of physics less transistors will always use less power, than more transistors. You are fixated on a single ISA/microarchitecture to the exclusion of all others, sounds like an agenda to me. The mobile market was made around non legacy software, and thus is not dependent on CISC, or any ISA, and the entire computing market has the SDK, tool chains, libraries, and APIs to utilize any ISA, and ISA change without any problems, just wait until the MIPS SOCs, start competing with the ARM SOCs, software porting is just a complier flag away! Don’t look under your bed at night, because there will be Multiple alternative ISAs that scare the Hell out of you, ISAs in the server room, Tablets, phones, laptops, and IOT, that have no ties to x86 or legacy software concerns. Look at the licensed IP market, and fear, CISC single ISA fandroid!

So sorry Scott, this is a
So sorry Scott, this is a place for conversation. However this power 8 person keeps doing these paragraphs long rants and it’s starting to drive us all crazy. If at least he or she logged in with a alias It would be better, but time and time again it’s anon “power 8, power 8 power 8, x86 sucks ” it’s starting to piss people off.

Try to be more appropriate,

Try to be more appropriate, folks. This comment and the one it was replying to.

Your a champion duck wang
Your a champion duck wang twanger, you love them curlicues, and CISC, but the OEMs know the history of chipzilla’s market grabs, and the extra market benefits of the RISC family.

There is already an ISA of the IOT, and no amount market tactics will let x86 weasel in, and MIPS will be coming up along side ARM, and chips will sale like pork bellies from the farm!

Free! as in free from monopolies! ISAs ISAs!!!

Exactly! I want the chamber!
Exactly! I want the chamber!

You’re a champion asshole
You’re a champion asshole

From Intel to the device
From Intel to the device OEMs, the mobile Tablet/Phone SOC pod thingy comes with a brown envelope stuffed with contra revenue! And hidden in that pod thingy, is a nose ring, should the OEM let Intel inside of too large a percentage of the OEM’s mobile device portfolio!

Oh good, power8 person.
Oh good, power8 person.

The BOM(b) from the ARM(RISC)
The BOM(b) from the ARM(RISC) based industry is, that Intel has problems competing in power usage, and price, with ARM chips that are 2 process nodes behind. We’re not talking about the Reference design ARM SOCs from ARM holdings (though they compete on power usage with Intel too), as much as talking about the Custom wider order designs, like Apple’s A7, and soon to be released A8. Those custom RISC designs that just happen to be engineered to execute the ARMv8 ISA, and even with Intel’s fab process node lead, are giving Intel a hard time even getting into the mainstream mobile devices market. Just Imagine a custom ARM core fabbed on a equivalent process node, or even a 20nm process node, and that CISC is just not going to be able to cope with ARM ISA based custom RISC designs at 20nm, even with that 14nm lead. Intel can be commended for keeping CISC alive for so long and pushing it so far, but in the end on the lowest power, equivalent process node, more CISC transistors will not be able to beat the Less transistors required for RISC instruction sets, and RISC does not even need to be on the smallest node to prove the point.

Intel is going to have server SKU(high margin) problems going forward, once those Chinese, and others’, Licensed Power8s begin to hit the market beginning in 2015. Intel’s FINFET is going to be needed to compete with the mobile RISC designs, that are still on planar process IC fabrication, and even Intel will not be able to get its tick tock moving fast enough, sub 14nm, to make up for the RISC designs inherent advantages, in a mobile market that is not dependent on legacy CISC software. Others will be using FINFET, and the costs of going smaller than 14nm is going to tax the entire chip fabrication industry’s infrastructure, and even Intel is going to find it economically unviable without slowing the pace of the process node shrinks, the sales volumes will just not be there to justify the extra costs of frequent process node shrinks. Look for those 2 year tick tocks to become 3 and 4 years, in order realistically amortize the R&D and fab infrastructure costs.

The BOM(b) for the ARM based parts will continue to be pricing based on razor thin margins, that no amount of contra revenue will be able to temporarily overcome.

(b) Bill of Materials