By what measure of performance? Are you comparing SIMD performance
on code with tight vectorizable loops (and a decent compiler)? Or
more typical workloads like compiling code?
On 2015-04-09 00:39, Mark Campbell wrote:
I think a cluster of RPi 2s would be worthwhile here--RPi 2 is now
1.5x - 6x (depending on task) faster than the original RPi, and looks
to be one of the more powerful embedded environments now. I read
someone do a benchmark with it, and they indicated that by using all 4
cores, it had the equivalent performance of roughly a high end Pentium
III or low end Pentium IV. The benchmark was by no means
comprehensive, but by comparison on the same benchmark, the RPi B+ was
roughly a Pentium II.
A cluster of those babies would make for a formidable build farm.
On Wed, Apr 8, 2015 at 7:10 PM, Gordan Bobic <[email protected]>
wrote:
Back in the day I considered running some QEMU emulated ARM VMs,
but the perforance on a high end x86 box was sufficiently poor
that a bunch of NFS-root Sheeva/Dream Plugs was considerably
more performance at a fraction of a power budget.
Don't get me wrong, it has it's uses, especially when you
don't have better alternatives, just prepare for the
performance to be quite poor.
Gordan
On 2015-04-09 00:02, Al Hopper wrote:
Very kool!
What are the specs on your ESXi host and:
- how many build VMs are you running?
- what kind of build times do you see?
- perhaps you could share the build times for a couple of
representative packages?
If this system were publicly accessible (by invitation), we'd have
a
redsleeve build farm!
Great idea...
On Wed, Apr 8, 2015 at 5:26 PM, Bjarne <[email protected]>
wrote:
Hi Gordan.
I am giving a little peep :)
I am currently setting up my own little build system on my VMware
ESXi host, using slaves with qemu running the Redsleeve image.
Lets see how it ends up :)
BR
Bjarne
On 08-04-2015 20:59, Gordan Bobic wrote:
RS EPEL is unmaintained. It was originally conributed by a
community member who afterwards didn't have time to maintain
it due to other commitments.
If anybody would like to step up as a maintainer for EPEL6
(and EPEL7, for that matter), please, do speak up.
Gordan
_______________________________________________
users mailing list
[email protected]
http://lists.redsleeve.org/mailman/listinfo/users [1] [1]
--
Al Hopper
Links:
------
[1] http://lists.redsleeve.org/mailman/listinfo/users [1]
_______________________________________________
users mailing list
[email protected]
http://lists.redsleeve.org/mailman/listinfo/users [1]
_______________________________________________
users mailing list
[email protected]
http://lists.redsleeve.org/mailman/listinfo/users [1]
--
--Mark
Links:
------
[1] http://lists.redsleeve.org/mailman/listinfo/users
_______________________________________________
users mailing list
[email protected]
http://lists.redsleeve.org/mailman/listinfo/users
_______________________________________________
users mailing list
[email protected]
http://lists.redsleeve.org/mailman/listinfo/users
