ARM seems to be completely ignoring the sales downturn that almost every single component manufacturer has seen in this quarter, as well as previous ones, turning in on increase of 14% on revenue and 24% on profit in Q1 of 2015. As The Register points out that equates to 450 chips selling every second, something even automated stock trading algorithms have to be impressed by. Royalty revenue increased by 31% thanks to Mali, regardless of Apple's decision not to use that chip in their iPhone 6. You can expect to see more news on ARM from us in the near future and you can expect the news to be good for their investors and users.
"The first three months of 2015 have been good to ARM, which saw revenues of $348.2m and pre-tax profits of $120.5m in the first quarter, with 3.8 billion ARM-based chips shipped – or more than 450 chips per second."
Here is some more Tech News from around the web:
- Google pulls plug on YouTube for older iPads, iPhones, smart TVs @ The Register
- How to Find Your Linux Version or Distro Release, and Why It Matters @ Linux.com
- SSL bug hits 1,000 iPhone, iPad apps including Microsoft and Yahoo titles @ The Inquirer
- 'No iOS Zone' Wi-Fi zero-day bug forces iPhones, iPads to crash and burn @ The Register
- New atomic clock won’t lose a second in 15 billion years @ Extremetech
How about a breakdown of
How about a breakdown of sales, chips made using ARM Holdings’ reference designs verses the chips that are fully custom and designed to run the ARMv8a instruction set, like Apples A8/X and Nvidia’a custom Denver, and others that are made by companies that only license the ARMv8 instruction set architecture(Top Tier architectural Licensees). AMD is one of those Top Tier architectural Licensees, and its future K12 should be the one to give Apple some competition among the ARM ISA based industry players with the resources to roll their own microarchitectures that are engineered to run the ARMv8a ISA.
Jim Keller’s K12 should be the one to make Apple worry, what with the rumors that AMD’s custom K12 ARMv8 running microarchitecture is getting SMT capabilities. Apple’s A8/X product are very wide superscalar designs 6 wide/6 IPC, Nvidia’s Denver is a little wider, at 7+ IPCs(see Nvidia’s Hot Chips presentation), so an AMD K12 custom extra wide order superscalar design, and SMT to take advantage of more than one processor thread per core would be able to take advantage of any execution stall/delays on the core’s other thread, and keep the core’s execution ports/pipelines operating at as close to 100% as possible. This is how Intel’s hyper-threading(SMT) keeps its individual core’s execution resources at as close to full utilization as possible, and why non hyper-threaded x86 cores are at a disadvantage in processing efficiency.
I’m very interested in just what AMD’s K12 is going to be, even more so than AMD’s Zen, as it will be AMD’s K12 with SMT that will beat anything x86 from Intel, or x86 from AMD itself for that matter, in power usage(lower power usage) and performance in a tablet form factor. AMD will not have to worry about getting its x86 tablet SKUs down towards the power usage metrics of the ARM RISC designs, as AMD will have a custom RISC ARMv8a ISA running design of its own. AMD’s custom ARM K12 will give AMD a Tablet SKU that even Apple’s ecosystem will have trouble matching. unless Apple’s P.A. semiconductor engineers begin working SMT into Apple’s custom A series designs.
I’m expecting good things from Jim Keller’s K12 design team, and a K12 based APU with AMD’s graphics/HSA that will give the PowerVRs in Apple’s A series SOC some competition, and Apple some serious headaches. For sure any OEM’s in the tablet market who do not have the engineering resources of Apple, or Nvidia/others, will be very interested in AMD’s K12, and K12 will get the design wins if it can compete in Apple’s A series performance metrics.
I wonder why Intel hasn’t
I wonder why Intel hasn’t gone and bought out ARM yet? That would be the easiest way to gain a foothold in the mobile world.
Intel would not be allowed to
Intel would not be allowed to by ARM, nor Samsung, Apple, Qualcomm, etc. Those companies already have CPU/SOC products, and Buying ARM Holdings is not necessary for Intel to get ARM designs of its own, Intel could just get an ARM architectural license of its own, and tape out an Intel custom microarchitecture of its own to run the ARMv8a ISA.
At one time Intel Had an ARM license of its own, after Intel bought StrongARM from Digital Equipment Corporation in 1997, and the StrongARM ran the ARMv4 instruction set architecture, then it was rebranded Xscale that ran the ARMv5 ISA, then Xscale was sold to Marvell.
You do realize that Instruction Sets are merely templates that the underlying microarchitecture is designed to run, None of Apple’s A7, or A8/X core designs are made by ARM Holdings, Apple only licensed the Instruction Set Architecture(ARMv8a) from ARM Holdings, and Apple had its P.A. semiconductor(CPU design company bought by Apple) engineers design a CPU core from scratch. Hell look at AMD, AMD’s x86 ISA is implemented using a different underlying microarchitecture than Intel’s underlying microarchitecture, Intel’s x86 microarchitecture uses SMT, while AMDs x86 microarchitecture does not currently utilize SMT, for one example and there are other differences. The standard x86 ISA(Not some extensions unique to Intel or AMD) will be able to run across any AMD or Intel CPU/SOC hardware.
Back to the question of Intel, or any company already in the CPU/SOC business being able to buy ARM holdings, the US Department of Justice Anti trust division, as well as the other concerned US/British government agencies would never approve just any company’s purchase of ARM Holdings, especially if that company was already producing CPUs/SOCs for the marketplace. The one major litmus test for any acquisition of one company by another is: Will the acquisition reduce competition in the marketplace. None of the current players in the CPU/SOC market could buy ARM Holdings, to remove the other players from the market, and not allow them licenses to the various ARM ISAs. If Intel/other CPU/SOC company were to try to buy ARM Holdings they would be stopped, and told to simply get a license from ARM Holdings.
ARM Holdings holds a special place in the industry in that ARM Holdings is the major supplier of ARM based CPU/ISA IP for most of the mobile market, so not anyone of the companies that are already licensing from ARM holdings would be allowed to buy ARM Holdings, nor would any company that already makes CPUs/SOCs of any ISA. Intel, and Apple/etc. are welcome to invest in ARM Holdings, but never a controlling interest level of investment, Apple Licenses the ARMV8a ISA from ARM Holdings, but designs its own microarchitecture to implement the ARMv8a ISA, Intel could do the same, hell Intel spent 7 billion dollars to pay OEMs to use its ATOM chips in tablets/Phones, Intel could have taped out a custom ARMv8a running microarchitecture for much less than 7 billion dollars. AMD’s K12 is going to be an all AMD custom microarchitecture designed to run the ARMv8a ISA.
Why would any company need to purchase ARM Holdings, when all ARM Holdings does to make money is license its reference designs, or license its ISA and lets the licensee design a custom microarchitecture to run the licensed ISA. Even IBM is getting into the licensing game with its power8 CPU designs, on a little bit more restrictive terms than ARM holdings does, but still non IBM companies are licensing Power8 designs, and producing non made by IBM power8s.
I would not be surprised in the least if Samsung got a Power8 license, Samsung, Globalfoundries and IBM have been in a technology sharing partnership for years, mostly concerned with processor chip fabrication technology sharing. And oh look, did not IBM pay GlobalFoundries to take over its commercial chip fab business. Why do you think that Samsung Fabs licensed the 14nm process node technology to Globalfoubdries, it has more to do with that technology partnership that both Samsung and GlobalFoundries had with IBM than other reasons, it’s most likely that it was the main reason. GlobalFoundries is going to be IBM’s fab partner, at least for IBM’s internal Power8/9 products, Samsung may get some of that third party Power8 fab business for the Chinese, and others.
you must have carpel tunnel
you must have carpel tunnel from writing these essays and stroking off ARM