cudaMemcpyToSymbol do not copy data

Accelerated Computing CUDA CUDA Programming and Performance
1

I want to use constant memory which will be accessed by all threads across all of my kernels.

The declaration is something like this

extern constant float smooth [8 * 1024];

I am copying data to this variable using

cudaMemcpyToSymbol(“smooth”, smooth_local, smooth_size, 0, cudaMemcpyHostToDevice);

smooth_size = 7K bytes

It was giving me incorrect output

but when I run it in -deviceemu mode and tried to print the contents of both these variables inside the kernel, I was getting all zeroes for smooth and smooth_local was correct.

I tried printing the output just after cudaMemcpyToSymbol still it was giving me 0’s.

Can you anyone throw light on my problem?

2

You must declare GPU constant memory using the Cuda constant qualifier. I am pretty certain if you check the error status with cudaGetLastError after the cudaMemcpyToSymbol, you will find you are gettin an undefined symbol error.

3

sorry I missed the __ before and after constant

I m using this

extern constant float smooth [2 * 1024];

now when I am printing with cudaGetLastError , I m getting “invalid device symbol”.

4

When I removed “extern” it works.

constant float smooth [2 * 1024];