Since Einstein@home doesn't release their source, the only constructive thing to do is suggest as may optimization methods as possible and hope that the project team picks them up and run with them.
Not really. It's the matter of optimization.
The transformation is right and acceptable. However as You notices it 'may'
lead to numeric instability.
In order to compute just like You want You have to turn off optimization and optimize the sourcecode itself.
Under gnuC it's -O0 - dunno how for M$.
Just so people know what optimization can do to your code, check out this little test I did:
For some reason the bloody forum won't let me post the code, basically it just added numbers in a loop of an undetermined length (read from stdin)...here's the results anyway:
Just so people know what optimization can do to your code, check out this little test I did:
For some reason the bloody forum won't let me post the code, basically it just added numbers in a loop of an undetermined length (read from stdin)...here's the results anyway:
I added two computers to my E@H cluster early this month just to see if what I thought I was seeing is true. The machines are identical P4 boxes (2.6GHz, HT, 512MB) but one is running Windows 2000 and the other runs Linux 2.6.9-1.667smp. The Linux box is much slower than the windows box. Any ideas as to why??
Linux: Measured FP 355.07, Interger 644.78
Last 5 work units:
CPS claim grant
Over Success Done 114,603.87 66.31 72.75
Over Success Done 103,363.96 59.81 72.78
Over Success Done 115,757.43 66.98 67.71
Over Success Done 98,435.62 56.96 72.80
Over Success Done 101,199.79 58.56 69.06
Win2K: Measured FP 1071.13, Interger 959.95
Last 5 work units:
CPS claim grant
Over Success Done 48,109.14 56.55 pending
Over Success Done 47,934.58 56.34 74.90
Over Success Done 48,194.14 56.65 72.88
Over Success Done 48,134.44 56.58 72.70
Over Success Done 47,992.80 56.41 75.54
Thanks,
Pat
I am running an AMD 1600 using tripple boot between W2K, Mandrake 10.1 and Amiga and receiving similar results. The Win version is roughly running 2 times the speed of the Mandrake.
I'm wondering if there's a code logic problem that the calculations are being done twice.
I'm wondering if there's a code logic problem that the calculations are being done twice.
Brian
No such problem. There would have to be literally dozens of threads over the last few months documenting this and with the reasons explained. Please do a search for wine+linux+client in thread titles and you'll get plenty of info.
Please learn to use the force ..errhhh.. search people :).
Since Einstein@home doesn't
)
Since Einstein@home doesn't release their source, the only constructive thing to do is suggest as may optimization methods as possible and hope that the project team picks them up and run with them.
Nah, just run the 'nix client
)
Nah, just run the 'nix client using Microsoft's Services For Unix.
http://www.microsoft.com/windowsserversystem/sfu/default.mspx
If you can compile it, you can run it.
RE: Today I found out that
)
Not really. It's the matter of optimization.
The transformation is right and acceptable. However as You notices it 'may'
lead to numeric instability.
In order to compute just like You want You have to turn off optimization and optimize the sourcecode itself.
Under gnuC it's -O0 - dunno how for M$.
RE: RE: Today I found
)
If optimization may cause incorrect results, it is good for [________]?
Please fill in. Thanks.
Just so people know what
)
Just so people know what optimization can do to your code, check out this little test I did:
For some reason the bloody forum won't let me post the code, basically it just added numbers in a loop of an undetermined length (read from stdin)...here's the results anyway:
$ gcc main.c -O0
$ echo 10000| ./a.out
1000.000000000158821 100.000000000014253 9.999999999999897 3700555018.500922203063965
$ gcc main.c -O3
$ echo 10000| ./a.out
999.999999999999886 100.000000000000014 10.000000000000000 3700555018.499999046325684
Just note the different results for different optimizations.
RE: Just so people know
)
What about the Microsoft compiler?
RE: Hi, I added two
)
I am running an AMD 1600 using tripple boot between W2K, Mandrake 10.1 and Amiga and receiving similar results. The Win version is roughly running 2 times the speed of the Mandrake.
I'm wondering if there's a code logic problem that the calculations are being done twice.
Brian
RE: I'm wondering if
)
No such problem. There would have to be literally dozens of threads over the last few months documenting this and with the reasons explained. Please do a search for wine+linux+client in thread titles and you'll get plenty of info.
Please learn to use the force ..errhhh.. search people :).
Cheers,
Gary.
RE: There would have to be
)
Surely it is time we had an FAQ on this topic - anyone up for writing it?
~~gravywavy
RE: RE: There would have
)
What would it say?
I know what little I have said about this topic is summed up as "it depends" ...