fix algorithm assigning values in testsuite

avoid a division by zero which seems to only occur for op=prod and
datatype=half, since the maximum exponent is small (15) and can exceed
the number of ranks.

verifiable/verifiable.cu

@@ -14,9 +14,6 @@
 
 #include "rccl/rccl.h"
 
-
-#define RCCL_BFLOAT 1
-
 #if NCCL_VERSION_CODE >= NCCL_VERSION(2,10,0) && RCCL_BFLOAT16 ==1
   #define HAVE_ncclBfloat16 1
 #else
@@ -124,7 +121,7 @@ namespace {
     return Y(x);
   }
   template<>
-  __host__ __device__ half castTo<__half>(float x) {
+  __host__ __device__ __half castTo<__half>(float x) {
     return __float2half(x);
   }
   #if RCCL_BFLOAT16 == 1
@@ -425,7 +422,7 @@ __host__ __device__ void genSumXY(
   // Let s be the number of ranks per partition. This is either rn/pn as we
   // intended, or y/p_sum if that's smaller to prevent overshooting our target y.
   uint32_t s = y/p_sum < rn/pn ? y/p_sum : rn/pn;
-  x = r/s < pn ? 1 + r/s : 0; //  First s*pn ranks contribute partition index +1.
+  x = (s != 0 && r/s < pn) ? 1 + r/s : 0; //  First s*pn ranks contribute partition index +1.
   x += r == rn-1 ? y - s*p_sum : 0; // Last rank contributes discrepancy.
 }
 }