search

HiFiLES / HiFiLES-solver

High Fidelity Large Eddy Simulation Solver
Other
173 stars 131 forks source link

ACCELERATE_BLAS & nvcc #10

Closed RcktMan77 closed 10 years ago

RcktMan77 commented 10 years ago

The CUDA compiler (nvcc) doesn't recognize the -flax-vector-conversions compiler flag which is enabled by default when selecting the ACCELERATE_BLAS option.

jrbull commented 10 years ago

Zach,

Are you trying to compile for GPUs on your Mac? It should only be possible to compile the CPU version with that architecture. Regards, Jonathan

----- Original Message ----- From: "Zach Davis" notifications@github.com To: "HiFiLES/HiFiLES-solver" HiFiLES-solver@noreply.github.com Sent: Sunday, June 22, 2014 10:40:28 AM Subject: [HiFiLES-solver] ACCELERATE_BLAS & nvcc (#10)

The CUDA compiler (nvcc) doesn't recognize the -flax-vector-conversions compiler flag which is enabled by default when selecting the ACCELERATE_BLAS option.

Reply to this email directly or view it on GitHub: https://github.com/HiFiLES/HiFiLES-solver/issues/10

RcktMan77 commented 10 years ago

HI Jonathan,

Yes, I was compiling for GPUs on a Mac. I was able to get an executable that runs by manually removing the -flax-vector-conversions flag when nvcc was called during make and then continuing the make process.

Best Regards,

Zach

On Jun 23, 2014, at 12:58 PM, jrbull notifications@github.com wrote:

Zach,

Are you trying to compile for GPUs on your Mac? It should only be possible to compile the CPU version with that architecture. Regards, Jonathan

----- Original Message ----- From: "Zach Davis" notifications@github.com To: "HiFiLES/HiFiLES-solver" HiFiLES-solver@noreply.github.com Sent: Sunday, June 22, 2014 10:40:28 AM Subject: [HiFiLES-solver] ACCELERATE_BLAS & nvcc (#10)

The CUDA compiler (nvcc) doesn't recognize the -flax-vector-conversions compiler flag which is enabled by default when selecting the ACCELERATE_BLAS option.

Reply to this email directly or view it on GitHub: https://github.com/HiFiLES/HiFiLES-solver/issues/10 — Reply to this email directly or view it on GitHub.