Re: armel v4t vs v5

[Rogério Brito <rbrito@ime.usp.br>]

Hi.

On Apr 18 2012, Wookey wrote:
> +++ Martin Guy [2012-04-18 14:30 +0200]:
> > How big is the speed advantage for
> > v5 users when optimizing for v5 instead of v4t? Enough to warrant the
> > exclusion of our v4t parts?
> 
> I just asked and the answer seems to be 'some, but not much'. There
> are a few more useful instructions which should amount to a few
> percent speed improvement, but the gains are not exciting, which does
> suggest leaving it at v4t until really no-one cares anymore.

This is an interesting discussion, as trying to understand the
variety/diversity of ARM is like a big puzzle to people used to more
"regular" platforms like amd64 and i386.

In the same fashion as the question above, what would the benefits of
compiling some selected packages (like what Mike Thompson is doing for the
R-Pi) to an armel such as:

,----[ cat /proc/cpuinfo ]
| Processor	: Feroceon rev 0 (v5l)
| BogoMIPS	: 265.42
| Features	: swp half thumb fastmult edsp 
| CPU implementer	: 0x41
| CPU architecture: 5TEJ
| CPU variant	: 0x0
| CPU part	: 0x926
| CPU revision	: 0
| 
| Hardware	: Buffalo/Revogear Kurobox Pro
| Revision	: 0000
| Serial		: 0000000000000000
`----

Reading Wikipedia, it seems that the E in the "5TEJ" as reported by Linux
could be of some use (akin to MMX/SSE?)

BTW, talking about ARM, the very fact that ARMvx != ARMx is confusing to
newcomers and, if I understand it correctly, for Linux, what is important is
the thing with "v" (e.g., ARMv5TEJ).

BTW#2, is there any application in Debian (or is GCC actually able to
generate code to) use Jazelle?


Thanks for any guidance for this misguided soul,

-- 
Rogério Brito : rbrito@{ime.usp.br,gmail.com} : GPG key 4096R/BCFCAAAA
http://rb.doesntexist.org/blog : Projects : https://github.com/rbrito/
DebianQA: http://qa.debian.org/developer.php?login=rbrito%40ime.usp.br