Hi Dave,
> By comparison, a C1060 Tesla gpu (in SPDP mode) gives 11.9 ns/day, and
> a
> C2050 gives 20.7 ns/day. (These are numbers from Ross' web page; my
> machines
> give about 12.6 ns/day for S1070 and 18.6 ns.day for C2050; this means
> you
> should probably expect some variaton in GPU speed depending on details
> of the
> machine configuation you have.)
It just occurred to me that this difference is approximately 10% which is
what NVIDIA were estimating would be the difference if ECC was turned off.
ECC is smart on the C2050 such that if you turn it off those transistors can
be used for computation instead. My C2050 had a pre-release bios and
initially didn't work properly and we assumed it was a problem with ECC so
we turned ECC off. I assume ECC comes on by default so is probably turned on
on your cards.
Can you try rooting into the machine and typing:
nvidia-smi -r
Which will indicate whether ECC is enabled
If ECC is enabled, try
nvidia-smi -g 0 --ecc-config=0
Then see if it makes a difference to the benchmarks.
All the best
Ross
/\
\/
|\oss Walker
---------------------------------------------------------
| Assistant Research Professor |
| San Diego Supercomputer Center |
| Adjunct Assistant Professor |
| Dept. of Chemistry and Biochemistry |
| University of California San Diego |
|
http://www.rosswalker.co.uk |
http://www.wmd-lab.org/ |
| Tel: +1 858 822 0854 | EMail:- ross.rosswalker.co.uk |
---------------------------------------------------------
Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read every day, and should not be used for urgent or sensitive issues.
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Wed Jul 21 2010 - 12:30:03 PDT
