c++ - Cuda Thrust memory management -

i wrote code below:

 unsigned int* addrarray = (unsigned int*)malloc(sizeof(unsigned int)*datasetrows);  thrust::exclusive_scan(binarydataset,binarydataset+(sizeof(unsigned int)*datasetrows),addrarray);  free(addrarray); 

binarydataset unsigned int* type , have no problems allocating, processing data in part of memory , freeing it. addrarray causes problems. aftter running program following error:

*** glibc detected *** ./prog: free(): invalid next size (normal): 0x000000000180be10 *** , memory map.

this code wont generate error when use cudafree(addrarray) instead, later cause segmentation fault or glibc malloc memory corruption problem. example running this:

unsigned int* addrarray = (unsigned int*)malloc(sizeof(unsigned int)*datasetrows); thrust::exclusive_scan(binarydataset,binarydataset+(sizeof(unsigned int)*datasetrows),addrarray); const unsigned int compactarraysize = addrarray[datasetrows-1] - 1; printf("%u\n",compactarraysize) float* compactminorclass = (float*)malloc(sizeof(float)*datasetcols*compactarraysize); 

will generate segmentation fault when trying malloc memory on compactminorclass pointer(printf gives correct output, prefixsum calculated correctly , output saved in addrarray should be). build program set flags g++ , nvcc following:

gxxflags:= -o3 -wall -wextra -m64 -std=c++0x

nvccflags:= -xcompiler -wall -xcompiler -wextra -m64 -arch=sm_11

(i have geforce310m have set sm_11) question why running thrust function generates errors, why can't free memory allocated on host using free() ?

this not correct:

thrust::exclusive_scan(binarydataset,binarydataset+(sizeof(unsigned int)*datasetrows),addrarray); 

try instead:

thrust::exclusive_scan(binarydataset,binarydataset+(datasetrows),addrarray); 

the intention here pointer arithmetic, not byte arithmetic. , likewise in second code sample.

i'm not sure it's idea specify:

-std=c++0x 

but don't know sure it's source of of problems.

you should not cudafree on ordinary host pointer.