Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 01 Sep 2007 11:53:21 +0300
From:      Stefan Lambrev <stefan.lambrev@moneybookers.com>
To:        JoaoBR <joao@matik.com.br>
Cc:        Roman Divacky <rdivacky@freebsd.org>, freebsd-current@freebsd.org, pluknet <pluknet@gmail.com>, =?UTF-8?B?QmrDtnJuIEvDtm5pZw==?= <bkoenig@cs.tu-berlin.de>
Subject:   Re: Adding k9 and k10 to bsd.cpu.mk
Message-ID:  <46D92881.4000004@moneybookers.com>
In-Reply-To: <200708312308.25372.joao@matik.com.br>
References:  <-3502020561049594852@unknownmsgid>	<200708312120.31912.joao@matik.com.br>	<20070901014323.GA41683@dragon.NUXI.org> <200708312308.25372.joao@matik.com.br>

next in thread | previous in thread | raw e-mail | index | archive | help
Hello,

JoaoBR wrote:
> On Friday 31 August 2007 22:43:23 David O'Brien wrote:
>   
>> On Fri, Aug 31, 2007 at 09:20:30PM -0300, JoaoBR wrote:
>>     
>>> On Friday 31 August 2007 21:07:10 David O'Brien wrote:
>>>       
>>>>> opterons are not easy but it is already kind of advanced cpu so
>>>>> could be
>>>>>           
>>>> Why are Opteron's any harder?
>>>>         
>>> because all of them are 64bit but some older ones are not SSE3 capable, <
>>> 250 I guess now but 252 is but not 100% sure
>>>       
>> It's not Opteron model # specific - but silicon revision specific.
>> There are rev C0 model 250's, along with rev CG, and rev E.
>>
>>     
>
>    less than rev.E support SEE3 ? Do you have a spec/link for that?
>
>   
>> Same for athlon64 - older ones don't support SSE3, newer ones do.
>>
>>     
>
> like I said before 'older ones' is kind of lame def
>
>   
>>> people 'kind of familiarly' with reading manuals and specs are already
>>> having difficulties here so imagin an average user (unkndefspec) who
>>> likes to optimize his kernel (his cpu's kernel of course :) )
>>>
>>> so as far as there are a cpu options for a freebsd kernel they should
>>> be understandable so it might be worse thinking well before doing
>>> (=less support = less questions = less problemas)
>>>       
>
>
>
>   
>> BTW, the AMD offically sanctioned spelling for GCC 'march' is "amdfam10"
>>     
>
>
> I guess you agree without any objections that that this 'is 100% userfriendly' 
> and could by add as '100% userfriendly' and so then I agree as well ... 
> but  ... will they print it on the box in order to see it ??? guess not ;)
>
>   
Sorry, but if GCC know the new amd processors as amdfam10 (with alias 
barcelona), but the FreeBSD alias for them is "athlon64-E"
how this is more user friendly ?

Same question to k8-sse3, opteron-sse3, athlon64-sse3.

Please take a look at this link : 
http://gcc.gnu.org/onlinedocs/gcc/i386-and-x86_002d64-Options.html#i386-and-x86_002d64-Options

As it is may be not a problem to have more aliases (not listed in gcc 
manuals), it sounds really bad idea to not have the same names as GCC.

-- 

Best Wishes,
Stefan Lambrev
ICQ# 24134177




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?46D92881.4000004>