Atom 330
32 bit profile (x86):
CHOST="i686-pc-linux-gnu"
CFLAGS="-march=prescott -O2 -pipe -fomit-frame-pointer"
CXXFLAGS="${CFLAGS}"
MAKEOPTS="-j5"
64 bit profile (amd64):
CHOST="x86_64-pc-linux-gnu"
CFLAGS="-march=nocona -O2 -pipe"
CXXFLAGS="${CFLAGS}"
MAKEOPTS="-j5"
On Thu, Jan 29, 2009 at 4:36 PM, Mario. <emef...@gmail.com> wrote:
> found it.
> http://stackoverflow.com/questions/110674/gcc-optimization-flags-for-intel-atom
>
> And list of m_arch flags is here:
> http://en.gentoo-wiki.com/wiki/Safe_Cflags/Intel#Atom_N270
>
> On Thu, Jan 29, 2009 at 4:29 PM, Eric H. Johnson
> <ejohn...@camalytics.com> wrote:
>> Alex,
>>
>>>> They are pretty ok, but it depends what you're after. <<
>>
>> Just the easiest way for a kernel neophyte like me to get smp support for
>> the Intel Atom 330. I know jmk has the same board and the guilty party as
>> far as getting me started on this. :) I just want to see if I can tweak a
>> few more microseconds out of the servo thread on this board.
>>
>>>> I am surprised you are running 2.6.26-generic on Hardy. I would expect
>> that to be 2.6.24 <<
>>
>> I think you are right, just poor memory. I had to reboot into doze to send
>> the email and typed it from what I thought I remembered.
>>
>>> I was basically able to get all of the steps for building on Debian
>>> Lenny to work down through 'make menuconfig' except that the
>>> CFLAGS_KERNEL value was not recognized. I used the source for kernel
>>> version 2.6.22 because it was the latest version having a patch file
>>> from rtai. I wasn't sure whether
>>> 2.6.24 needed a patch file or not, so just to get the procedure down I
>>> decided to go with the latest version that did have a patch.
>>
>>>> Since 2.6.24 the 2 architectures i386 and x86_64 have been merged into a
>> single arch called x86.
>> You can fin rtai patches for newer kernels in arch/x86/.. <<
>>
>> Ok, I will look there.
>>
>>>> I'm not sure that's right for Atom. Core2 refers to Core 2 or Core 2 Duo,
>> which might be way different than a dual core Atom. <<
>>
>> I did not find anything really definitive in a Google search, but found one
>> forum post which said to use core 2. That is pretty weak, but the best I
>> had.
>>
>>> The script make-kpkg did not exist on my system, so I went to the
>>> rtai-steps documentation and was able to do a "make all", "make
>>> modules", "make bzImage" and "make modules install" (see below).
>>> mkinitrd was not found but I saw that it has been replaced with
>>> mkinitramfs, which did appear to run properly.
>>
>>>> If it doesn't exist, then you install it with "sudo apt-get install
>> make-kpkg".
>> HOWEVER, the make-kpkg is the debian preffered way of building kernels,
>> Ubuntu prefers to do it differently.
>> There is a Ubuntu wiki showing how kernels are to be built. (This assumes
>> you want a distributable .deb package, if you only want to compile and
>> install the kernel, then the "make menuconfig, make modules, make
>> modules_install, make bzImage"-way is perfectly fine.
>> https://help.ubuntu.com/community/Kernel/Compile <<
>>
>> I do want to make a deb package, both because I am not building on the
>> machine with the Atom processor, and as noted above, at least one other
>> person is using the same board and could use it.
>>
>> I suppose that may be a problem. The development machine has a genuine Core
>> 2 duo processor, while the machine I am targeting it for uses the Atom 330.
>> I was assuming the same kernel would work for both systems.
>>
>> I will look at the Ubuntu wiki.
>>
>>>> One of the ideas is if the initrd doesn't hold modules which allow
>> mounting your / partition.
>> Maybe you have some more errors in the scrollback..
>> If that's the case, you need to put together an initrd (make sure you pass
>> the info how to load it from grub), or compile the needed bits into the
>> kernel. <<
>>
>> If by scrollback you mean looking back at what was outputted during
>> compiling, I checked that pretty carefully. There were a fair number of
>> warnings, but I did not see any errors. I will check it again.
>>
>> Thanks,
>> Eric
>>
>>
>>
>> ------------------------------------------------------------------------------
>> This SF.net email is sponsored by:
>> SourcForge Community
>> SourceForge wants to tell your story.
>> http://p.sf.net/sfu/sf-spreadtheword
>> _______________________________________________
>> Emc-developers mailing list
>> Emc-developers@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/emc-developers
>>
>
------------------------------------------------------------------------------
This SF.net email is sponsored by:
SourcForge Community
SourceForge wants to tell your story.
http://p.sf.net/sfu/sf-spreadtheword
_______________________________________________
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
