Register Units

isa_manim.isa_objects.reg_unit presents one register or a group of registers. It can be used for general-purpose registers, vector registers, and matrix registers.

One example of a single register is shown below:

Source code: test_one_dim_reg.py

One example of a list of registers is shown below:

Source code: test_two_dim_reg.py

As shown in the above figures, one register unit contains the following Manim objects: - reg_rect presents the register. - The width of reg_rect presents the bit width of the register. - The height of reg_rect presents the number of registers or the row count of one matrix register. If there is only one register, the height is 1.0. - label_text_list contains the labels that present the name of the register. The vertical distance between the position of labels is 1.0. The right boundary of labels is close to the left boundary of reg_rect.

As the green dot shown in the above figures, the origin point of one register unit is located in the centre position of the first row in reg_rect, which is different from the geometry centre position of this object. It is suggested to use function shift rather than function move_to to change the location of the register.

Function get_elem_center returns the position of one specified element. The bit width of the specified element can be different from the element bit width of the vector.

Register units can maintain values of the register in the format of one single value (for GPRs), a 1-D array of values (for vector registers), or a 2-D array of values (for matrix registers). Function get_elem_value returns the value of one specified element, while function set_elem_value modifies the value of one specified element.

Thus, arguments for all methods are designed to support matrix registers. If creating or operating on vector registers, nreg must be 1 in the constructor function and reg_idx is ignored in other methods. If creating or operating on general-purpose registers, elements must be 1 in the constructor function and index is ignored in other methods.

RegUnit

Bases: VGroup

Object for one register.

Attributes:

Name	Type	Description
name_text_list	List[Text]	List of label text objects.
reg_rect	Rectangle	List of register rectangle objects.
reg_name_list	List[str]	List of register names.
reg_color	Color	Color of register rectangle and labels.
reg_font_size	int	Font size of register labels.
reg_value_format	str	Format string for output values. Inherented by element units.
reg_value	Union[Any, List[Any], List[List[Any]], None]	Register value. Inherented by element units.
reg_width	int	Bit width of register.
reg_count	int	Number of registers, or row count of matrix registers.
elem_count	int	Number of elements.
elem_width	int	Bit width of elements.

require_serialization = False class-attribute instance-attribute

Animation related with this object does not need to be serialized.

__init__(name_list, color, width, elements, nreg, value, font_size, value_format)

Constructor a register.

• When construct a scalar register, elements is 1 and nreg is 1.
• When construct a vector register, elements is the number of elements, and nreg is 1.
• When construct a matrix register or a list of registers, elements is the number of elements and nreg is the number of registers or the row count of matrix registers.

Parameters:

Name	Type	Description	Default
name_list	List[str]	Register names which could be a string or a list of string.	required
color	Color	Color of register and label.	required
width	int	Width of register, in bits.	required
elements	int	Number of elements in one register.	required
nreg	int	Number of registers, used to create matrix registers or a list of registers.	required
value	Union[Any, List[Any], List[List[Any]], None]	Value of this element. None is provided if not necessary.	required
font_size	int	Font size of label.	required
value_format	str	Format string for output values. Inherented by element units.	required

__repr__()

Return a string for debugging.

Returns:

Type	Description
str	A string for debugging.

__str__()

Return a string for debugging.

Returns:

Type	Description
str	A string for debugging.

get_elem_pos(index, reg_idx, offset, elem_width)

Return the center position of one specified element.

The element is specified by index and reg_idx. In general, index and reg_idx should within the scope of the register. If index or reg_idx beyond the scope of the register, this function returns one element specified as below:

• The actual index to access element is index % elem_count
• The actual row index to access element is (reg_idx + index // elem_count) % reg_count.

The width to index elements is determined by the construtor function. However, it is not possible to operate on only a part of the element. For example, one 128 bit vector has eight 16-bit elements.

• The 4-th element [79:64] is accessed if index is 4, offset is 0 and elem_width is 16.
• Lower half of the 4-th element [71:64] is accessed if index is 4, offset is 0 and elem_width is 8.
• Higher half of the 4-th element [79:72] is accessed if index is 4, offset is 8 and elem_width is 8.

Parameters:

Name	Type	Description	Default
index	int	Index of elements.	required
reg_idx	int	Index of register.	required
offset	int	Offset of lower bits.	required
elem_width	int	Width of element in bits.	required

Returns:

Type	Description
ndarray	Position of the specified element.

get_elem_value(index, reg_idx)

Return the value of one specified element. Return None if the value of this register is not specified in the constructor function.

• Return self.elem_value for scalar registers.
• Return self.elem_value[index] for vector registers.
• Return self.elem_value[reg_idx][index] for a list of registers or matrix registers.

The element is specified by index and reg_idx. In general, index and reg_idx should within the scope of the register. If index or reg_idx beyond the scope of the register, this function returns one element specified as below:

• The actual index to access element is index % elem_count
• The actual row index to access element is (reg_idx + index // elem_count) % reg_count.

Parameters:

Name	Type	Description	Default
index	int	Index of elements.	required
reg_idx	int	Index of register.	required

Returns:

Type	Description
Union[Any, List[Any], List[List[Any]], None]	Value of the specified element.

get_placement_height()

Return the height of this object for placement. The height is ceil to an integer.

Returns:

Type	Description
int	The height of this object.

get_placement_mark()

Return the marker of this object for placement, which is 2.

Returns:

Type	Description
int	Marker of this object.

get_placement_width()

Return the width of this object for placement. The width is ceil to an integer.

Returns:

Type	Description
int	The width of this object.

set_elem_value(value, index, reg_idx)

Modify the value of one specified element.

• Modify self.elem_value for scalar registers.
• Modify self.elem_value[index] for vector registers.
• Modify self.elem_value[reg_idx][index] for a list of registers or matrix registers.

The element is specified by index and reg_idx. In general, index and reg_idx should within the scope of the register. If index or reg_idx beyond the scope of the register, this function returns one element specified as below:

• The actual index to access element is index % elem_count
• The actual row index to access element is (reg_idx + index // elem_count) % reg_count.

If the value of this register is not specified in the constructor function (self.elem_value is None), and this register is one vector register, matrix register, or a group of registers, one 1-D/2-D array of values is created with None elements.

Parameters:

Name	Type	Description	Default
index	int	Index of elements.	required
reg_idx	int	Index of register.	required
value	Any	Value of this element.	required

set_placement_corner(row, col)

Set the position of object by the left-up corner position. Move object to the specified position.

Parameters:

Name	Type	Description	Default
row	int	Vertical ordinate of left-up corner.	required
col	int	Horizontal ordinate of left-up corner.	required