|
|
|
|
|
|
|
|
|
|
xen-users
Re: [Xen-users] DomU slower with smp?
Hi, Thanks for your answer. I've done as you said. Unfortunatelly no difference. it goes worse when I increase file size or threads number.
for: time iozone -t 40 -r 2m -s 100m -m -R DomU with four cpus (each from different core)
real 6m19.300s DomU with one cpu real 3m32.651s Dom0 real 2m37.418s
-- # xm info host : xxxx release : 2.6.18.8-xen version : #2 SMP Mon May 12 13:19:28 IST 2008
machine : x86_64 nr_cpus : 16 nr_nodes : 1 cores_per_socket : 4 threads_per_core : 1 cpu_mhz : 2393 hw_caps : bfebfbff:20100800:00000000:00000140:0004e3bd:00000000:00000001
total_memory : 65530 free_memory : 12 node_to_cpu : node0:0-15 xen_major : 3 xen_minor : 2 xen_extra : .1 xen_caps : xen-3.0-x86_64 xen-3.0-x86_32p hvm-3.0-x86_32 hvm-3.0-x86_32p hvm-3.0-x86_64
xen_scheduler : credit xen_pagesize : 4096 platform_params : virt_start=0xffff800000000000 xen_changeset : Fri May 02 14:23:07 2008 +0100 16888:9733f681d5c5 cc_compiler : gcc version 4.1.3 20080114 (prerelease) (Debian 4.1.2-19)
cc_compile_by : root cc_compile_domain : lan cc_compile_date : Fri May 9 14:41:37 IST 2008 xend_config_format : 4
Thanks in advance.
2008/5/14 Apparao, Padmashree K < padmashree.k.apparao@xxxxxxxxx>:
The SMP case may be suffering because of
context switching.
Can you run the SMP by affintizing the
vcpus to specific cores
Vcpu on core 1
Vcpu2 on core 2
Vcpu3 on core 3 etc
Thanks
-Padma
Hi,
I just made simple peformance test with iozone tool on xen 3.2.1.
time iozone -t 20 -r 2m -s 100m -m -R
and:
Dom0 - lvm(XFS) : real 1m35.670s
DomU - 2m46.963s
lvm(xfs)
cpus = "4-7"
vcpus=4
DomU - 1m53.772s
lvm(xfs)
cpus = "8"
vcpus=1
It seems DomU is much slower if using more than one cpu.
am I right?
--
Thanks
Janul
_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users
|
|
|
|
|