FAILED to free megabyte array on device

Accelerated Computing HPC Compilers Legacy PGI Compilers
#1

hello , I am trying to use the Accelerator and but I get these error
call to cuMemcpyDtoH returned error 999: Unknown
CUDA driver version: 5000
and when I tried to get some info about my device using pgaccelinfo




C:\Users\ASUS>pgaccelinfo
CUDA Driver Version: 5000

Device Number: 0
Device Name: GeForce GT 630M
Device Revision Number: 2.1
Global Memory Size: 2147287040
Number of Multiprocessors: 2
Number of Cores: 64
Concurrent Copy and Execution: Yes
Total Constant Memory: 65536
Total Shared Memory per Block: 49152
Registers per Block: 32768
Warp Size: 32
Maximum Threads per Block: 1024
Maximum Block Dimensions: 1024, 1024, 64
Maximum Grid Dimensions: 65535 x 65535 x 65535
Maximum Memory Pitch: 2147483647B
Texture Alignment: 512B
Clock Rate: 1344 MHz
Current free memory: 2086117376
FAILED to free megabyte array on device

C:\Users\ASUS>pgaccelinfo
CUDA Driver Version: 5000

Device Number: 0
Device Name: GeForce GT 630M
Device Revision Number: 2.1
Global Memory Size: 2147287040
Number of Multiprocessors: 2
Number of Cores: 64
Concurrent Copy and Execution: Yes
Total Constant Memory: 65536
Total Shared Memory per Block: 49152
Registers per Block: 32768
Warp Size: 32
Maximum Threads per Block: 1024
Maximum Block Dimensions: 1024, 1024, 64
Maximum Grid Dimensions: 65535 x 65535 x 65535
Maximum Memory Pitch: 2147483647B
Texture Alignment: 512B
Clock Rate: 1344 MHz
Current free memory: 2086117376
FAILED to free megabyte array on device

#2

Hi ahmad adyleh,

The only time I’ve seen this problem is when using an older CUDA driver on a newer card. However, you have a new driver so I’m not sure what’s going on. Are you able to run a nvcc built CUDA C program?

  • Mat
#3

yes , even I can compile and run example like Matmul in fortran

#4

Hi ahmad adyleh,

I wonder if you are timing out. Since you’re running a display, Windows will kill your job after a few seconds so it doesn’t lock up your display.

What happens if you increase the size of you’re Matmul and thus increase the compute time?

  • Mat