Ptr to next or previous physical malloc_chunk.

* @brief Normalize a vector, returning @c safe if len is zero. */

Sets the buffer data and capacity. */

Tries resizing the buffer to contain *count* elements. If T is a POD type the new elements may not be initialized.

* @brief Write up to 8 bits at an arbitrary bit offset. * * The stored value is at most 8 bits, but can be stored at an offset of between 0 and 7 bi

-************************************************************** * Internal functions ****************************************************************

These spill a ton of globals throughout shader

ZSTD_isLegacy() : @return : > 0 if supported by legacy decoder. 0 otherwise. return value is the version. */

Compare packed single-precision (32-bit) floating-point elements in a and b, and store packed maximum values in dst. dst does not follow the IEEE Stan

Subtract packed 8-bit integers in b from packed 8-bit integers in a, and store the results in dst. https://www.intel.com/content/www/us/en/docs/intrin

Extracts bits [low,high]

Clears this buffer. */

Converts value in the range [0, 100) to a string.

* @brief SP float bit-interpreted as an integer. * * @param v The value to bitcast. * * @return The converted value. */

* @brief Return a integer value for a float vector, using truncation. */

See header for documentation. */

length_squared

GL/D3D hobbled non-pow2 mips by only supporting round down, not round up And then Metal followed OpenGL since it's the same hw and drivers. Round up a

This should always succeed since it follows a call to parse_number.

This should always succeed since it follows a call to parse_number.