Quadro NVS 290

matrixvecmult executes kernels MatrixVectorMul1n which compute c = A * b
where A is stored in row major form and c and b are vectors

Using device Quadro NVS 290
CPU took 0.01701 s

Testing MatrixVectorMul1
 WorkGroupSize = 64 GlobalSize 100032
Finished kernel execution
  Average kernel execution time 0.22918
  Found 45542 different entries
  0 entries more than 0.001%

Testing MatrixVectorMul2
 WorkGroupSize = 64 GlobalSize 3840
Finished kernel execution
  Average kernel execution time 0.229667
  Found 45542 different entries
  0 entries more than 0.001%

Testing MatrixVectorMul3
 WorkGroupSize = 64 GlobalSize 3840
Finished kernel execution
  Average kernel execution time 0.142201
  Found 85800 different entries
  0 entries more than 0.001%

Testing MatrixVectorMul4
 WorkGroupSize = 64 GlobalSize 3840
Finished kernel execution
  Average kernel execution time 0.115059
  Found 82080 different entries
  0 entries more than 0.001%

Testing MatrixVectorMul5
 WorkGroupSize = 64 GlobalSize 3840
Finished kernel execution
  Average kernel execution time 0.115417
  Found 82202 different entries
  0 entries more than 0.001%