-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi folks,
Our new diskless Intel cluster has 10 nodes with Xeon Phi's in them, and I'm wondering if anyone has had any luck with getting the extra bloat out of the RAM disk? We're using xCAT to manage the cluster. Basic problem is none of them support relocation: [root@old-barcoo-m knc]# rpm -qip *.rpm | grep -i relocat Name : intel-mic Relocations: (not relocatable) Name : intel-mic-cdt Relocations: (not relocatable) Name : intel-mic-flash Relocations: (not relocatable) Name : intel-mic-gdb Relocations: (not relocatable) Name : intel-mic-gpl Relocations: (not relocatable) Name : intel-mic-kmod Relocations: (not relocatable) Name : intel-mic-kmod Relocations: (not relocatable) Name : intel-mic-micmgmt Relocations: (not relocatable) Name : intel-mic-mpm Relocations: (not relocatable) Name : intel-mic-sysmgmt Relocations: (not relocatable) They chuck all their stuff onto the local filesystem so you get about 1GB of extra stuff on an image compared to a non-phi node (which only lacks the Phi related RPMs). The only one you need to deal with is the intel-mic-gpl RPM as that has the Linux distro for the Phi in it and so has the largest footprint (way over 700MB) and includes compilers, etc. All the best, Chris - -- Christopher Samuel Senior Systems Administrator VLSCI - Victorian Life Sciences Computation Initiative Email: sam...@unimelb.edu.au Phone: +61 (0)3 903 55545 http://www.vlsci.org.au/ http://twitter.com/vlsci -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iEYEARECAAYFAlGHO5QACgkQO2KABBYQAh/mWwCghMirN+KcSV18ryPx6oJnjjiU 7kgAn1jyAg7PwNvMbXKK4H/SMhBbnskv =Crg5 -----END PGP SIGNATURE----- _______________________________________________ Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
