Discussion in 'Article Discussion' started by CardJoe, 21 Jul 2010.
physical IP, Intel may have some serious competition on netbook markets. it's a shame ARM isn't a household name.
imagine, everyone knows what processor is in their smartphone that has way better battery life than their laptops. then an ARM based netbook will sell like hot cakes.
i really hope this will create a platform that allows manufacturers to easily create tablets based on this. current limitations is high R&D costs on developing a tablet with competitive performance, an ARM based SoC platform will help vastly.
Given TSCM's epic fail with 40mm, isn't this a risky plan?
I thought it was nVidia's fail. ATi is doing fine with TSMC's 40nm, at least they don't complain...
Ya it was mainly nVidia that epically failed during that 40nm process, so ya shouldnt be an issue at all.
cool go ARM!. By 2030 there may be a 3 dog race for desktop CPU's....well 2 architecture race. Anyone think x86 would disappear if ARM got into the desktop segment?
Seeing as there's not been an ARM processor powerful enough to rival even budget x86 processors in the desktop market (to my knowledge) then no. Running all my favorite games on a smartphone processor does not sound appealling.
Sloth is right, the RISC architecture is not aimed at pure performance (or backwards compatibility for that matter) it's aimed at power efficiency.
ARM architecture would only really take over when/if Intel/AMD hits a power wall and stops innovating.
Except, of course, that modern x86 hardware is RISC at heart anyway...
During the VCF, Sophie Wilson - original creator of the ARM architecture - fielded a question about relative performance with the statement that ARM and x86 are already at a "performance parity," which is to say that both architectures perform equally well.
The reason ARM *processors* (distinct from architecture) are slower is because, as you say, they are built with a low power consumption in mind. If anyone were to build a 3GHz ARM chip with a far higher power draw than usual and a massive cooling fan, Wilson argues that it would perform just as well - if not better - than a 3GHz x86 chip from Intel or AMD.
Such a processor doesn't exist *yet* - but could well appear in the future if ARM start making inroads in less-portable (i.e. smartbook, netbook, or even notebook) markets.
Granted, Sophie Wilson is hardly unbiased, but if anyone knows what they're talking about on this subject it's her.
well spotted, i meant processors.
IF the architecture has been designed to run at 3GHz, and IF the software has been re-written to be optimised, ARM processors running at same speed would win a x86.
problem of course is backwards compatibility and software optimisation. x86 is indeed basically a RISC core wrapped in a more complex decoder. the decoder takes care of those problem and it's what made x86 so successful.
Separate names with a comma.