Hi *, I’ve upgraded my desktop at work:
• from Kubuntu Hardy (8.04) to Debian unstable (sid), still i386 • from aranym_0.9.13-3.1 to whatever Debian sid carries (0.9.15-1) • from a 4x3.2 GHz AMD system to an 8x3.1 GHz Intel system My VMs, despite turning on the “performance” governor by writing into sysfs, have only half the speed of before (roughly 500 ipv 1000 pystones/sec, 70/70/61/65 instead of 192/192/192/90 BogoMIPS; the last VM to start was always a bit slower). In /proc/cpuinfo I have cpu MHz : 3060.000 for all eight host cores (now), though. I did have issues with governors in the past; even on hardy, the last VM to start was slower, and sometimes I had to start them several times to get the full BogoMIPS value. I do pin VMs to single CPU cores by using taskset with a first argument that has only a single bit set (0x1, 0x2, 0x4, 0x8). Does anyone else see that, and which of the changes are most likely responsible? (Debian also builds with hardening I think, which Kubuntu hardy didn’t do.) I’ve not yet done any real testing, but I can probably run a hardy chroot and compare (that keeps hardware and kernel but not the userspace); might be able to install hardy to an LV and boot it with its own kernel, too… the old mainboard was broken, so can’t easily test it… but maybe someone else here sees differences between versions on the same hardware, or across hardware with similar nominal speed. For the record: the new CPU is Intel(R) Core(TM) i7 CPU 950 @ 3.07GHz and the old one was an AMD X4 II or something like that. Ah. Before, I ran the hardy i386 -server kernel; now I run the sid amd64 kernel (with i386 userspace). bye, //mirabilos -- 15:39⎜«mika:#grml» mira|AO: "mit XFree86® wär’ das nicht passiert" - muhaha 15:48⎜<thkoehler:#grml> also warum machen die xorg Jungs eigentlich alles kaputt? :) 15:49⎜<novoid:#grml> thkoehler: weil sie als Kinder nie den gebauten Turm selber umschmeissen durften? -- ~/.Xmodmap wonders… -- To UNSUBSCRIBE, email to debian-68k-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/pine.bsm.4.64l.1305112200170.28...@herc.mirbsd.org
