| 1 |
Am Sun, 09 Dec 2012 20:24:58 +0100 |
| 2 |
schrieb Florian Philipp <lists@×××××××××××.net>: |
| 3 |
|
| 4 |
> Am 09.12.2012 04:51, schrieb Michael Mol: |
| 5 |
> > On Sat, Dec 8, 2012 at 10:25 PM, Grant <emailgrant@×××××.com> wrote: |
| 6 |
> >> It seems like ARM processors will destroy x86 before too long. Does anyone |
| 7 |
> >> think this won't happen? |
| 8 |
> > |
| 9 |
> > It's looking promising. Not that I have a horse in the race, but I |
| 10 |
> > very much like ARM's low power consumption. The way I see it, they're |
| 11 |
> > only a short list of features away from obliterating x86: |
| 12 |
> > |
| 13 |
> > * I'd like to see fast division. |
| 14 |
> > I keep hearing about how this or that is slow because of ARM's lack of |
| 15 |
> > strong division. |
| 16 |
> > |
| 17 |
> > * I'd like to see a modern baseline of strong instructions. |
| 18 |
> > x86 kept continually improving in a very fragmented way, but there |
| 19 |
> > were, from time to time, baseline collections of feature sets you |
| 20 |
> > could expect all processors to have. i386 represented one. i686 |
| 21 |
> > represented one. Currently, it's x86_64, which implies not only a |
| 22 |
> > 64-bit flattened address space and a departure from real mode, but |
| 23 |
> > also a collection of SIMD instruction sets and other features |
| 24 |
> > developed between the release of the Pentium Pro and AMD's Hammer |
| 25 |
> > architecture. |
| 26 |
> > |
| 27 |
> > ARM just feels...fragmented. And I don't have the impression I could |
| 28 |
> > write my code assuming the availability of SIMD (presuming I use |
| 29 |
> > things like OpenMP to expand my code to leverage it, rather than |
| 30 |
> > writing processor-specific code. Though OpenCL could very well |
| 31 |
> > alleviate that issue.) |
| 32 |
> > |
| 33 |
> |
| 34 |
> +1 with regard to fragmentation. What I especially despise is the lack |
| 35 |
> of a common boot infrastructure. If I'm not mistaken, it is still |
| 36 |
> impossible to make a kernel that boots on all (or at least a large |
| 37 |
> subset of all) ARM platforms [1]. |
| 38 |
|
| 39 |
> [1] https://lwn.net/Articles/496400/ |
| 40 |
|
| 41 |
I'm no embedded developer, but I read recently that the first batch of |
| 42 |
multi-platform ARM support has been merged in Linux 3.7: |
| 43 |
|
| 44 |
http://www.h-online.com/open/features/Kernel-Log-Coming-in-3-7-Part-5-CPU-and-platform-code-1758293.html |
| 45 |
|
| 46 |
-- |
| 47 |
Marc Joliet |
| 48 |
-- |
| 49 |
"People who think they know everything really annoy those of us who know we |
| 50 |
don't" - Bjarne Stroustrup |
Archives