CI: fix cuda 9.2 build

Workaround following error:

```
/usr/lib/gcc/x86_64-linux-gnu/5/include/avx512fintrin.h(9220): error #167: argument of type "const void *" is incompatible with parameter of type "const float *"
100/usr/lib/gcc/x86_64-linux-gnu/5/include/avx512fintrin.h(9231): error #167: argument of type "const void *" is incompatible with parameter of type "const float *"
101/usr/lib/gcc/x86_64-linux-gnu/5/include/avx512fintrin.h(9244): error #167: argument of type "const void *" is incompatible with parameter of type "const double *"
```

.gitlab-ci.yml

@@ -12,6 +12,11 @@ cuda92:
   image: registry.gitlab.com/hzdr/crp/alpaka-group-container/alpaka-ci-cuda92-gcc:1.4
   variables:
     CUPLA_CXX: "g++-6"
+    # Workaround requires to enable debug build:
+    # GCC-5.5 has broken avx512vlintrin.h in Release mode with NVCC 9.X
+    #   https://gcc.gnu.org/bugzilla/show_bug.cgi?id=76731
+    #   https://github.com/tensorflow/tensorflow/issues/10220
+    CMAKE_BUILD_TYPE: "Debug"
     CUPLA_BOOST_VERSIONS: "1.65.1 1.66.0 1.67.0 1.68.0 1.69.0 1.70.0 1.71.0 1.72.0 1.73.0"
   extends: .base_cuda