On Sun, 27 Sep 2015 15:19:41 -0500
Bruce Dubbs <bruce.du...@gmail.com> wrote:

> so the patch doesn't make any difference for timing, at least with my 
> i7-5820K CPU.

  Bruce,

Thanks for letting us know. Given the comments in that make bug thread,
I'm pretty sure there *was* a problem here with regard to the performance
of fork() versus vfork(). I guess they fixed that in more recent kernels.

... Yep, from late 2010:

http://stackoverflow.com/questions/4259629/what-is-the-difference-between-fork-and-vfork

"With most current kernels, however, including Linux, the primary benefit
 of vfork has disappeared because of the way fork is implemented. Rather
 than copying the whole image when fork is executed, copy-on-write
 techniques are used."


Maybe the folks with make performance problems were the ones running
older kernels.


On Mon, 28 Sep 2015 00:27:24 +0100
Ken Moffat <zarniwh...@ntlworld.com> wrote:

> For the kernel, if I stop and repeat sometimes it builds ok and
> soemtimes it doesn't.  I think that is the definition of "flakey
> hardware".


  Ken,

In that case, I bet it is either a thermal or a power supply problem -
the -j4 pulls more power and makes more CPU heat. If there isn't
any/enough thermal grease on the CPU, the heat sink is not mounted
correctly, the CPU core voltage is not correct (or clock speed too
high), or the CPU fan isn't moving enough air over the right parts
(and also through the case), problems like this can happen under
high load. Ditto for a power supply or power connectors that cannot
handle the demanded current. Check that heat sink, check the core
voltage and setting. Also, where the power connector goes into the
motherboard - look for heat discoloration or other signs of burning.

Well, I'm glad to know there doesn't seem to be an overall problem
with AMD and Linux or else, as much as I do like AMD, I would have
to avoid buying an AMD-based system.


  Cheers,

  Mike

-- 
http://lists.linuxfromscratch.org/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Reply via email to