GPU-HMMER (errors seen with CUDA 4.0) Compilation errors.

While compiling GPU-HMMER with CUDA 4.0, these are some of the warnings and errors which I can see. Can anybody kindly help? Thank you!

cuda_kernel.cu(288): warning: variable “progress” was set but never used

cuda_kernel.cu(1885): warning: non-void function “GPUtask” should return a value

cuda_kernel.cu(1621): warning: variable “cfree” was declared but never referenced

cuda_kernel.cu(1621): warning: variable “ctotal” was declared but never referenced

cuda_kernel.cu(1912): warning: variable “vStart” was declared but never referenced

cuda_kernel.cu(1913): warning: variable “vStop” was declared but never referenced

cuda_kernel.cu(1914): warning: variable “DMAstart” was declared but never referenced

cuda_kernel.cu(1915): warning: variable “DMAstop” was declared but never referenced

cuda_kernel.cu(1916): warning: variable “Seqstart” was declared but never referenced

alphabet.o:1:8: warning: null character(s) ignored
alphabet.o:1:18: warning: null character(s) ignored
alphabet.o:1:20: warning: null character(s) ignored
alphabet.o:1:22: warning: null character(s) ignored

alphabet.o(1): error: unrecognized token

alphabet.o(1): error: expected a declaration

alphabet.o(1): error: unrecognized token

alphabet.o(1): error: unrecognized token

Error limit reached.
100 errors detected in the compilation of “/tmp/tmpxft_00003db9_00000000-78_alphabet.cpp1.ii”.
Compilation terminated.

the command generating the CUDAPROGS is not supported by nvcc now, so the two CUDA file should be compiled to object file first and then linked into the target. Here is what I did:

{CUDAPROGS}: %: %.o {OBJS} {cuda_objs} {NVCC} -g -Xptxas -fastimul --maxrregcount=42 --ptxas-options=-v {DEFS} {CUDADEFS} {MYINCDIR} {CUDAINCDIR} {MYLIBDIR} -o @ {cuda_objs} {OBJS} {MYLIBS} {PTHREAD_LIBS} {LIBS} {CUDALIBS} -I/home/user/NVIDIA_GPU_Computing_SDK/C/common/inc/ -L/home/user/NVIDIA_GPU_Computing_SDK/C/lib/

cuda_objs = cuda_kernel.o cuda_hmmsearch.o
cuda_kernel.o:cuda_kernel.cu
{NVCC} -g -Xptxas -fastimul --maxrregcount=42 --ptxas-options=-v {DEFS} {CUDADEFS} {MYINCDIR} {CUDAINCDIR} {MYLIBDIR} -c cuda_kernel.cu -I${HOME}/NVIDIA_GPU_Computing_SDK/C/common/inc/ -L${HOME}/NVIDIA_GPU_Computing_SDK/C/lib/
cuda_hmmsearch.o:cuda_hmmsearch.c
{NVCC} {CFLAGS} {CPPFLAGS} {DEFS} {PTHREAD_CFLAGS} {MYINCDIR} -c cuda_hmmsearch.c

how to download CUDA-HMMER?? i’ve searched in google but i didn’t get any…