Set correct range for OpVectorTimesScalar tests

Set correct range for OpVectorTimesScalar tests

The min[] and max[] values should be based on the result in double precision, not after it has been converted to fp16type.

Affects:

dEQP-VK.spirv_assembly.instruction.compute.float16.arithmetic_2.opvectortimesscalar
dEQP-VK.spirv_assembly.instruction.compute.float16.arithmetic_3.opvectortimesscalar
dEQP-VK.spirv_assembly.instruction.compute.float16.arithmetic_4.opvectortimesscalar
dEQP-VK.spirv_assembly.instruction.graphics.float16.arithmetic_2.opvectortimesscalar*
dEQP-VK.spirv_assembly.instruction.graphics.float16.arithmetic_3.opvectortimesscalar*
dEQP-VK.spirv_assembly.instruction.graphics.float16.arithmetic_4.opvectortimesscalar*

Components: Vulkan

VK-GL-CTS issue: 1532
Change-Id: I30810c8c954ee899bb05602b1f8805f82ac021b3