Does it show that my cluster runs perfectly?
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
=?gb2312?B?194gueLPyA==?= zouguangxian at hotmail.comWed Aug 29 17:13:30 PDT 2001
- Previous message: meshing together multiple nodes without a switch
- Next message: Does it show that my cluster runs perfectly?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Hi,:)
yestoday ,I test this beowulf cluster by linpack and get some data.during
the testing ,i change the N so that the linpack can lead to the BEST
performance.for example,i have 2 nodes with PIII 933 (1 master node and 1
slave node),2*1G=2G memory.so the N can be 10000(if i let N=12000 or
N=15000,it will cause error :( ).i am sure the net is beyond the request
in this test.finally,the linpack output such result:
T/V N NB P Q Time Gflops
w11c2c4 10000 60 1 2 725.99 9.205e-1
in the abstract,the cluster can reach 0.933*2=1.866Gflops( i am not sure
whether it can be calculated in this way).it seems that the cluster works
so
badly.do it?
maybe you can teach me how to judge the performance of the cluster by
observing these data.maybe there is a stardard to do so.but i dont know.:(
thanks for your attention. ^_^
Regards
weck
_________________________________________________________________
Äú¿ÉÒÔÔÚ MSN Hotmail Õ¾µã http://www.hotmail.com/cn Ãâ·ÑÊÕ·¢µç×ÓÓʼþ
- Previous message: meshing together multiple nodes without a switch
- Next message: Does it show that my cluster runs perfectly?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
