Rosetta 3.8 GPU compilation

[Anonymous]

Hello,

I am having an error while compiling Rosetta with GPU options. I am using the following command to build Rosetta 3.8 with opencl GPU options:

scons mode=release extras=opencl bin

The log produced during the building of executables is as follows:

g++ -o build/src/release/linux/4.8/64/x86/gcc/5.4/opencl/basic/gpu/GPU.os -c -std=c++0x -ffor-scope -isystem external/boost_1_55_0/ -isystem external/ -isystem external/include/ -isystem external/dbio/ -pipe -Wall -Wextra -pedantic -Werror -Wno-long-long -Wno-strict-aliasing -march=core2 -mtune=generic -O3 -ffast-math -fno-finite-math-only -funroll-loops -finline-functions -finline-limit=20000 -s -Wno-unused-variable -Wno-unused-parameter -fPIC -DBOOST_ERROR_CODE_HEADER_ONLY -DBOOST_SYSTEM_NO_DEPRECATED -DBOOST_MATH_NO_LONG_DOUBLE_MATH_FUNCTIONS -DPTR_STD -DNDEBUG -DUSEOPENCL -Isrc -Iexternal/include -Isrc/platform/linux/64/gcc/5.4 -Isrc/platform/linux/64/gcc -Isrc/platform/linux/64 -Isrc/platform/linux -Iexternal/boost_1_55_0 -Iexternal/libxml2/include -Iexternal -Iexternal/dbio -I/usr/include -I/usr/local/include -I/opt/AMDAPP/include -I/opt/AMDAPP/include/CL -I/opt/AMDAPP/lib/x86 -I/opt/AMDAPP/lib/x86_64 src/basic/gpu/GPU.cc

src/basic/gpu/GPU.cc: In member function 'int basic::gpu::GPU::Init()':

src/basic/gpu/GPU.cc:283:22: error: '_cl_command_queue* clCreateCommandQueue(cl_context, cl_device_id, cl_command_queue_properties, cl_int*)' is deprecated [-Werror=deprecated-declarations]

dev.commandQueue = clCreateCommandQueue(context_, dev.device, profiling_ ? CL_QUEUE_PROFILING_ENABLE : 0, NULL);

^

In file included from /usr/include/CL/opencl.h:42:0,

from src/basic/gpu/GPU.hh:29,

from src/basic/gpu/GPU.cc:27:

/usr/include/CL/cl.h:1359:1: note: declared here

clCreateCommandQueue(cl_context                     /* context */,

^

src/basic/gpu/GPU.cc:283:22: error: '_cl_command_queue* clCreateCommandQueue(cl_context, cl_device_id, cl_command_queue_properties, cl_int*)' is deprecated [-Werror=deprecated-declarations]

dev.commandQueue = clCreateCommandQueue(context_, dev.device, profiling_ ? CL_QUEUE_PROFILING_ENABLE : 0, NULL);

^

In file included from /usr/include/CL/opencl.h:42:0,

from src/basic/gpu/GPU.hh:29,

from src/basic/gpu/GPU.cc:27:

/usr/include/CL/cl.h:1359:1: note: declared here

clCreateCommandQueue(cl_context                     /* context */,

^

src/basic/gpu/GPU.cc:283:113: error: '_cl_command_queue* clCreateCommandQueue(cl_context, cl_device_id, cl_command_queue_properties, cl_int*)' is deprecated [-Werror=deprecated-declarations]

dev.commandQueue = clCreateCommandQueue(context_, dev.device, profiling_ ? CL_QUEUE_PROFILING_ENABLE : 0, NULL);

^

In file included from /usr/include/CL/opencl.h:42:0,

from src/basic/gpu/GPU.hh:29,

from src/basic/gpu/GPU.cc:27:

/usr/include/CL/cl.h:1359:1: note: declared here

clCreateCommandQueue(cl_context                     /* context */,

^

src/basic/gpu/GPU.cc: In member function 'int basic::gpu::GPU::ExecuteKernel(const char*, int, int, int, ...)':

src/basic/gpu/GPU.cc:687:26: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]

arg.k_p = (void*)arg.i;

^

src/basic/gpu/GPU.cc:695:26: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]

arg.k_p = (void*)arg.i;

^

cc1plus: all warnings being treated as errors

scons: *** [build/src/release/linux/4.8/64/x86/gcc/5.4/opencl/basic/gpu/GPU.os] Error 1

scons: building terminated because of errors.

 

Does anyone know what this error means or (better) how can i solve it?

[Anonymous]

Those are all warnings emitting from the GPU code itself, or even the system headers.  Probably the simplest thing to do is turn off the warnings-as-errors option in your build settings.  Here’s how you turn it off.  https://www.rosettacommons.org/comment/7584#comment-7584

We use warnings-as-errors for the standard build, but the GPU code is probably not regularly tested to make sure it stays warning-free (and certainly isn’t tested against multiple compiler setups, all of which produce different warnings).

Don’t forget -j# when you compile for speed!    (Although defininitely use -j1 when making nice log files for us!)

[Anonymous]

Those are all warnings emitting from the GPU code itself, or even the system headers.  Probably the simplest thing to do is turn off the warnings-as-errors option in your build settings.  Here’s how you turn it off.  https://www.rosettacommons.org/comment/7584#comment-7584

We use warnings-as-errors for the standard build, but the GPU code is probably not regularly tested to make sure it stays warning-free (and certainly isn’t tested against multiple compiler setups, all of which produce different warnings).

Don’t forget -j# when you compile for speed!    (Although defininitely use -j1 when making nice log files for us!)

[Anonymous]

Those are all warnings emitting from the GPU code itself, or even the system headers.  Probably the simplest thing to do is turn off the warnings-as-errors option in your build settings.  Here’s how you turn it off.  https://www.rosettacommons.org/comment/7584#comment-7584

We use warnings-as-errors for the standard build, but the GPU code is probably not regularly tested to make sure it stays warning-free (and certainly isn’t tested against multiple compiler setups, all of which produce different warnings).

Don’t forget -j# when you compile for speed!    (Although defininitely use -j1 when making nice log files for us!)

[Author]

Posts