Closes #30663

Description

GLSL in WebGL2 supports the following functions for packing floating-point values:

uint packUnorm2x16(vec2 v);
uint packSnorm2x16(vec2 v);
uint packHalf2x16(vec2 v);

vec2 unpackUnorm2x16(uint v);
vec2 unpackSnorm2x16(uint v);
vec2 unpackHalf2x16(uint v);

WGSL has the equivalent set of float-packing functions:

fn pack2x16snorm(e: vec2<f32>) -> u32
fn pack2x16unorm(e: vec2<f32>) -> u32
fn pack2x16float(e: vec2<f32>) -> u32

fn unpack2x16snorm(e: u32) -> vec2<f32>
fn unpack2x16unorm(e: u32) -> vec2<f32>
fn unpack2x16float(e: u32) -> vec2<f32>

This PR adds the GLSL-style functions to TSL

Additional Notes
WGSL supports additional packing functions beyond the ones mentioned in this PR. However, WebGL2 doesn't support them. This PR only adds the TSL functions that can be directly turned into an intrinsic in both.