Shared-scale block conversion instances

Shared-scale block quantization requires an explicit shared exponent. No DefaultQuantizer instance is installed.

@[instance_reducible]

instance FloatLib.Floats.ExecFloat.SharedScale.Conversion.quantizer {lanes : ℕ} : Quantizer (SharedScale lanes) (Vector ℚ lanes)

Shared-scale destinations support quantization only with an explicit exponent context.

There is intentionally no DefaultQuantizer instance.