Working on a grid of 40 nodes
1 次查看(过去 30 天)
显示 更早的评论
Hello. I would like to run a monte carlo simulation for snr=2:0.5:5 and 1000 iterations for every snr on my university grid.It has 40 nodes and each node has 4 cores. I set 1 worker per node and i made 35 tasks like:
t1 = createTask(j, @mcc, 2, {2,1,200,10000});
t2 = createTask(j, @mcc, 2, {2,201,400,10000});
t3 = createTask(j, @mcc, 2, {2,401,600,10000});
t4 = createTask(j, @mcc, 2, {2,601,800,10000});
t5 = createTask(j, @mcc, 2, {2,801,1000,10000});
t6 = createTask(j, @mcc, 2, {2.5,1,200,10000});
....
t35 = createTask(j, @mcc, 2, {5,801,1000,10000});
where
t = createTask(j, @mcc, 2, {snr,j_start,j_end,length of code});
Every node had 25% cpu usage.So i set 4 workers per node and i made 140 tasks like:
t1 = createTask(j, @mcc, 2, {2,1,50,10000});
t2 = createTask(j, @mcc, 2, {2,51,100,10000});
t3 = createTask(j, @mcc, 2, {2,101,150,10000});
t4 = createTask(j, @mcc, 2, {2,151,200,10000});
....
t140= createTask(j, @mcc, 2, {5,951,1000,10000});
Now, every node has 100%cpu usage.
My questions are:
1.)Is there any reason to use matlabpool somewhere in my program?
2.)Is there any better way to run my program?
0 个评论
回答(1 个)
另请参阅
类别
在 Help Center 和 File Exchange 中查找有关 MATLAB Parallel Server 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!