Differences

This shows you the differences between two versions of the page.

--- en:multiasm:papc:chapter_6_4 [2025/05/15 11:35] – [Floating point values] ktokarz
+++ en:multiasm:papc:chapter_6_4 [2025/05/18 19:02] (current) – [Pointers] ktokarz
@@ Line 66: / Line 66: @@
 ===== Floating point values =====
-Floating point values store the data encoded for calculation on real numbers. Depending on the precision required for the algorithm, we can use different data sizes. Scalar data types are supported by the FPU (Floating Point Unit), offering single precision, double precision or double extended precision real numbers. In C/C++ compilers, they are referred to as float, double and long double data types, respectively. Vector (packed) floating-point data types can be processed by many SSE and AVX instructions, offering fast vector, matrix or artificial intelligence calculations. Vector units can process half precision, single precision and double precision formats. The 16-bit Brain Float format was introduced to calculate the dot scalar product to improve the efficiency of AI training and inference algorithms. Floating point data types are shown in figure {{ref>floattypes}} and described in table {{ref>tablefloattypes}}.
+Floating point values store the data encoded for calculation on real numbers. Depending on the precision required for the algorithm, we can use different data sizes. Scalar data types are supported by the FPU (Floating Point Unit), offering single precision, double precision or double extended precision real numbers. In C/C++ compilers, they are referred to as float, double and long double data types, respectively. Vector (packed) floating-point data types can be processed by many SSE and AVX instructions, offering fast vector, matrix or artificial intelligence calculations. Vector units can process half precision, single precision and double precision formats. The 16-bit Brain Float format was introduced to calculate the dot scalar product to improve the efficiency of AI training and inference algorithms. Floating point data types are shown in figure {{ref>floattypes}} and described in table {{ref>tablefloattypes}}. The table shows the number of bits used. In reality, the number of mantissa bits is assumed to be one bit longer, because the highest bit representing the integer part is always "1", so there is no need to store it (except for Double extended data format, where the integer bit is present).
 <figure floattypes>
@@ Line 75: / Line 75: @@
 <table tablefloattypes>
 <caption>Floating point data types</caption>
-^ Name              ^ Bits  ^ Mantissa  ^ Exponent  ^ Min value                                 ^ Max value                                 ^
+^ Name              ^ Bits  ^ Mantissa bits  ^ Exponent bits  ^ Min value                                 ^ Max value                                 ^
-| Double extended   | 80    | 64        | 15        | {{:en:multiasm:cs:float_ep_min.png?125}}  | {{:en:multiasm:cs:float_ep_max.png?125}}  |
+| Double extended   | 80    | 64             | 15             | {{:en:multiasm:cs:float_ep_min.png?125}}  | {{:en:multiasm:cs:float_ep_max.png?125}}  |
-| Double precision  | 64    | 52        | 11        | {{:en:multiasm:cs:float_dp_min.png?120}}  | {{:en:multiasm:cs:float_dp_max.png?115}}  |
+| Double precision  | 64    | 52             | 11             | {{:en:multiasm:cs:float_dp_min.png?120}}  | {{:en:multiasm:cs:float_dp_max.png?115}}  |
-| Single precision  | 32    | 23        | 8         | {{:en:multiasm:cs:float_sp_min.png?120}}  | {{:en:multiasm:cs:float_sp_max.png?110}}  |
+| Single precision  | 32    | 23             | 8              | {{:en:multiasm:cs:float_sp_min.png?120}}  | {{:en:multiasm:cs:float_sp_max.png?110}}  |
-| Half precision    | 16    | 10        | 5         | {{:en:multiasm:cs:float_hp_min.png?110}}  | {{:en:multiasm:cs:float_hp_max.png?100}}  |
+| Half precision    | 16    | 10             | 5              | {{:en:multiasm:cs:float_hp_min.png?110}}  | {{:en:multiasm:cs:float_hp_max.png?100}}  |
-| Brain Float       | 16    | 7         | 8         | {{:en:multiasm:cs:float_bf_min.png?120}}  | {{:en:multiasm:cs:float_bf_max.png?110}}  |
+| Brain Float       | 16    | 7              | 8              | {{:en:multiasm:cs:float_bf_min.png?120}}  | {{:en:multiasm:cs:float_bf_max.png?110}}  |
 </table>
-Single precision
-Double precision
+===== Floating point vector data types =====
-Double extended precision
+Floating point vectors are formed with single or double precision packed data formats. They are processed by SSE or AVX instructions in a SIMD approach of processing. A 128-bit packed data format can store four single-precision data elements or two double-precision data elements. A 256-bit packed data format can store eight single-precision values or four double-precision values. A 512-bit packed data format can store sixteen single-precision values or eight double-precision values. These packed data types are shown in figure {{ref>packedfloattypes}}. Instructions operating on 16-bit half-precision values or Brain Floats can use twice as many such elements simultaneously in comparison to single-precision data.
-Half precision
+It is worth mentioning that some instructions operate on a single floating-point value, using only the lowest elements of the operands.
-Brain Float
+<figure packedfloattypes>
+{{ :en:multiasm:cs:packed_floats.png?800 |Illustration of packed floating point data types}}
+<caption>Pcked floating point data types in x64 architecture</caption>
+</figure>
+===== Bit field data type =====
+A bit field is a data type whose size is counted by the number of bits it occupies. The bit field can start at any bit position in the fundamental data type and can be up to 32 bits long. MASM supports it with the RECORD data type. The bit field type is shown in figure {{ref>bitfieldtype}}.
+<figure bitfieldtype>
+{{ :en:multiasm:cs:bit_field_type.png?300 |Illustration of bit field data type}}
+<caption>The bit field data type</caption>
+</figure>
+===== Pointers =====
+Pointers store the address of the memory which contains interesting information. They can point to the data or the instruction. If the segmentation is enabled, pointers can be near or far. The far pointer contains the logical address (formed with the segment and offset parts). The near pointer contains the offset only. The offset can be 16, 32 or 64 bits long. The segment selector is always stored as a 16-bit number. Illustration of possible pointer types is shown in figure {{ref>pointertypes}}.
+<figure pointertypes>
+{{ :en:multiasm:cs:pointers_types.png?600 |Illustration of near and far pointers types}}
+<caption>The near and far pointers types</caption>
+</figure>
+<note>
+The offset is often the result of complex addressing mode calculations and is called an effective address.
+</note>

en/multiasm/papc/chapter_6_4.1747308933.txt.gz · Last modified: 2025/05/15 11:35 by ktokarz