y-cruncher Multi-Precision Library
	A Multi-threaded Multi-Precision Arithmetic Library (Powered by y-cruncher)

(Last updated: March 3, 2016)

 

Globals

 

Miscellaneous global variables and datatypes used by the YMP library (as well as y-cruncher itself).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

---

Primitive Datatypes:

Type	Description	Notes
u32_t s32_t	32-bit integer
u64_t s64_t	64-bit integer
upL_t spL_t	Pointer Length	Usually the same size as size_t. It is used for lengths and sizes that fit into the memory address space. These are 32 or 64-bit integers.
ufL_t sfL_t	File Length	Same as upL_t and spL_t but for file lengths. Currently, these are always 64-bit integers. Guaranteed to be at least as large as the pointer length types.
uiL_t siL_t	Index Length	An integer type that is large enough to do all memory calculations. Currently, these are 64-bit integers on all targets.

 

 

---

Alignment:

const upL_t ALIGNMENT;

Description:

The minimum alignment for scratch buffers used for large multiplication.

 

In the past, this alignment was equal the SIMD size on the target processor. Now it simply returns the size of the cacheline. There is sufficient evidence showing that cacheline alignment benefits the internal algorithms.