multiplied.analysis package#

Submodules#

multiplied.analysis.context module#

This module will exclusively provide contexts to open multiple files in batches, providing a unified interface to track across multiple .parquet truth table files.

These contexts should allow for a simple “with xyz() as abc” to accommodate other analysis functions.

Use pyarrow.dataset when working with multiple files

multiplied.analysis.extract module#

multiplied.analysis.extract.pq_extract_bits(
path: str,
bits: list[int],
stages: list[int],
) → DataFrame[source]#

Return a DataFrame of specified bits across multiple stages from .parquet

multiplied.analysis.extract.pq_extract_formatted_all(path: str) → DataFrame[source]#

Return DataFrame of all formatted strings from .parquet

multiplied.analysis.extract.pq_extract_formatted_stages(
path: str,
stages: list[int],
) → DataFrame[source]#

Return DataFrame of formatted strings across multiple stages from .parquet

multiplied.analysis.extract.pq_extract_stages(
path: str,
*,
stages: list[int] = [],
) → DataFrame[source]#

Return a DataFrame of specified stages from .parquet

Parameters:

  • path (str) – Path to Multiplied-generated .parquet file

  • stages (list[int]) – List of stages to extract

Returns:

A subset of the original .parquet table

Return type:

pd.DataFrame

multiplied.analysis.heatmap module#

multiplied.analysis.heatmap.df_global_3d_heatmap(
path: str,
title: str,
df: DataFrame,
*,
dark=False,
) → None[source]#

Export image of 3d plot with heatmap for each stage stacked along the x-axis

Parameters:

  • path (str) – Path to save the heatmap, ending with a chosen image format.

  • title (str) – Title of the heatmap.

  • df (pd.DataFrame) – DataFrame containing the data to be plotted. All available columns are used.

  • dark (bool, optional) – Whether to use a dark theme for the plot, by default False.

Return type:

None

multiplied.analysis.heatmap.df_global_heatmap(
path: str,
title: str,
df: DataFrame,
*,
dark=False,
) → None[source]#

Export image of pyplot of global heatmap

Parameters:

  • path (str) – Path to save the heatmap, ending with a chosen image format.

  • title (str) – Title of the heatmap.

  • df (pd.DataFrame) – DataFrame containing the data to be plotted. All available columns are used.

  • dark (bool, optional) – Whether to use a dark theme for the plot, by default False.

Return type:

None

multiplied.analysis.heatmap.df_stage_bound_heatmap(
path: str,
df: DataFrame,
stages: list[int],
bound: list[tuple[int, int]],
) → None[source]#

Export pyplot heatmap of bounding box region across stages

multiplied.analysis.heatmap.df_stage_heatmap(
path: str,
df: DataFrame,
stages: list[int],
) → None[source]#

Export pyplot heatmap for each selected stage

multiplied.analysis.search module#

multiplied.analysis.search.df_early_bitwidth_cout(
df: DataFrame,
) → DataFrame[source]#

Return rows which carry past operand width during partial product generation

multiplied.analysis.search.df_late_bitwidth_cout(
df: DataFrame,
) → DataFrame[source]#

Return rows which carry past operand width during reduction

Module contents#