Dear Dr. Walker,
Thank you for your suggestion, but unfortunately I cannot upgrade to
amber12 at least for now.
Btw, I managed to re-install the cuda-4.*, and it turned out that it works.
Thank you.
Chinh
On Fri, Apr 26, 2013 at 12:10 PM, Ross Walker <ross.rosswalker.co.uk> wrote:
> Hi Chinh,
>
> AMBER 11 does not support CUDA 5.0 so yes you will need CUDA 4.2. That
> said the AMBER 11 GPU code is no longer supported, it would be too
> difficult to keep this updated in addition to the AMBER 12 code. I would
> highly suggest that if you can you upgrade to AMBER 12 and you will find
> that everything works much better.
>
> Firsly it supports CUDA 5.0, is faster, has more features, a better
> precision model and has multiple fixes related to corner cases causing
> issues when run on GPUs.
>
> All the best
> Ross
>
>
>
> On 4/25/13 9:00 PM, "Chinh Su Tran To" <chinh.sutranto.gmail.com> wrote:
>
> >Hi,
> >
> >I encountered the same problem* "cudaMemcpyToSymbol: SetSim copy to Sim
> >failed"* when i ran pmemd. And I found the solution from Dr. Walker as
> >below.
> >I have to admit that I am very new to CUDA and all installation stuff, so
> >please help me with some suggestion.
> >My information:
> >- Cuda-5.0, nvcc release 5.0, amber11, ubuntu 12.10
> >- GPU GeForce GTX580
> >
> >Does that mean I need to reinstall the nvcc v4.2? Do I have to start from
> >scratch? If yes, could you please instruct me how to do that?
> >
> >It was really nightmare to deal with "conflict" of gcc, g++, and gfortran
> >while compiling pmemd cuda in amber11!!
> >
> >Please help. Thank you.
> >Chinh
> >
> >
> >On Thu, Jan 10, 2013 at 6:23 AM, Ross Walker <ross.rosswalker.co.uk>
> >wrote:
> >
> >> Use NVCC v4.2 - Cuda 5.0 is not currently supported by the release
> >>version
> >> of the code as of Jan 9th 2013.
> >>
> >>
> >>
> >> On 1/9/13 2:02 PM, "Mohammad Ashraf Bhuiyan" <akasheee.gmail.com>
> wrote:
> >>
> >> >Hi,
> >> >I tried to run the GPU benchmarks JAC_production_NPT and NVE and
> >> >Cellulose_Production_NPT and NVE using CUDA 5 on Nvidia K20x. But it
> >>stuck
> >> >at some point, and after I cancel the run by cntrl+c, it says:
> >> >
> >> >cudaMemcpyToSymbol: SetSim copy to Sim failed
> >> >
> >> >
> >> >The last part of the "mdout" file says:
> >> >
> >>
> >>>------------------------------------------------------------------------
> >>>--
> >> >------
> >> > 3. ATOMIC COORDINATES AND VELOCITIES
> >>
> >>>------------------------------------------------------------------------
> >>>--
> >> >------
> >> >
> >> >
> >> >
> >> > begin time read from input coords = 6.000 ps
> >> >
> >> >
> >> > Number of triangulated 3-point waters found: 7023
> >> >
> >> > Sum of charges from parm topology file = -11.00000006
> >> > Assuming uniform neutralizing plasma
> >> >
> >>
> >>>........................................................................
> >>>..
> >> >....
> >> >Is it expected?
> >> >
> >> >The K20x GPU is active, I can run other cuda program on it.
> >> >The device query says:
> >> >Device 0: "Tesla K20Xm"
> >> > CUDA Driver Version / Runtime Version 5.0 / 5.0
> >> > CUDA Capability Major/Minor version number: 3.5
> >> > Total amount of global memory: 5760 MBytes
> >>(6039339008
> >> >bytes)
> >> > (14) Multiprocessors x (192) CUDA Cores/MP: 2688 CUDA Cores
> >> > GPU Clock rate: 732 MHz (0.73 GHz)
> >> > Memory Clock rate: 2600 Mhz
> >> > Memory Bus Width: 384-bit
> >> > L2 Cache Size: 1572864 bytes
> >> > Max Texture Dimension Size (x,y,z) 1D=(65536),
> >> >2D=(65536,65536), 3D=(4096,4096,4096)
> >> > Max Layered Texture Size (dim) x layers 1D=(16384) x 2048,
> >> >2D=(16384,16384) x 2048
> >> > Total amount of constant memory: 65536 bytes
> >> > Total amount of shared memory per block: 49152 bytes
> >> > Total number of registers available per block: 65536
> >> > Warp size: 32
> >> > Maximum number of threads per multiprocessor: 2048
> >> > Maximum number of threads per block: 1024
> >> > Maximum sizes of each dimension of a block: 1024 x 1024 x 64
> >> > Maximum sizes of each dimension of a grid: 2147483647 x 65535 x
> >> >65535
> >> > Maximum memory pitch: 2147483647 bytes
> >> > Texture alignment: 512 bytes
> >> > Concurrent copy and kernel execution: Yes with 2 copy
> >>engine(s)
> >> > Run time limit on kernels: No
> >> > Integrated GPU sharing Host Memory: No
> >> > Support host page-locked memory mapping: Yes
> >> > Alignment requirement for Surfaces: Yes
> >> > Device has ECC support: Enabled
> >> > Device supports Unified Addressing (UVA): Yes
> >> > Device PCI Bus ID / PCI location ID: 132 / 0
> >> > Compute Mode:
> >> > < Default (multiple host threads can use ::cudaSetDevice() with
> >> >device
> >> >simultaneously) >
> >> >
> >> >deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 5.0, CUDA
> >>Runtime
> >> >Version = 5.0, NumDevs = 1, Device0 = Tesla K20Xm
> >> >
> >> >
> >> >Best Regards
> >> >
> >> >Ashraf
> >> >
> >> >--------------------------------------------------
> >> >M Ashraf Bhuiyan, PhD
> >> >_______________________________________________
> >> >AMBER mailing list
> >> >AMBER.ambermd.org
> >> >http://lists.ambermd.org/mailman/listinfo/amber
> >>
> >>
> >>
> >> _______________________________________________
> >> AMBER mailing list
> >> AMBER.ambermd.org
> >> http://lists.ambermd.org/mailman/listinfo/amber
> >>
> >_______________________________________________
> >AMBER mailing list
> >AMBER.ambermd.org
> >http://lists.ambermd.org/mailman/listinfo/amber
>
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Apr 25 2013 - 23:00:02 PDT
