call magma function from device (NOT host)
Posted: Mon Mar 09, 2015 3:27 pm
Hi when I tried to call a magma function from a cuda kernel I got an error like this:
error: calling a __host__ function("magma_sposv_batched") from a __global__ function("invokeDeviceMagma") is not allowed
Does magma functions have a device version, like cublas has? I really want to invoke magma from device (i.e., inside a kernel) using dynamic parallelism.
Thanks!
Wei
error: calling a __host__ function("magma_sposv_batched") from a __global__ function("invokeDeviceMagma") is not allowed
Does magma functions have a device version, like cublas has? I really want to invoke magma from device (i.e., inside a kernel) using dynamic parallelism.
Thanks!
Wei