classify_matrix_type

C++ Function Reference

1 Signature

MatrixType BigDataStatMeth::classify_matrix_type(hsize_t N, hsize_t M, hsize_t K)

2 Description

Classify matrix type based on dimensional characteristics.

3 Parameters

N (hsize_t): Number of rows in result matrix
M (hsize_t): Number of columns in result matrix
K (hsize_t): Inner dimension for matrix multiplication

4 Returns

MatrixType Classification result for optimization strategy selection

5 Details

Analyzes matrix dimensions to determine the optimal computational strategy. Classification considers both absolute size and aspect ratios to identify rectangular vs. square-like matrices and their relative sizes.

6 Caller Graph

7 Source Code

Implementation

File: inst/include/Utilities/Utilities.hpp • Lines 580-596

inline MatrixType classify_matrix_type(hsize_t N, hsize_t M, hsize_t K) {
        hsize_t min_dim = std::min({N, M, K});
        hsize_t max_dim = std::max({N, M, K});
        double ratio = (double)max_dim / min_dim;
        
        if (min_dim < 1000 && ratio > 100) {
            return RECTANGULAR_EXTREME;
        } else if (max_dim < 10000 && ratio < 5) {
            return SQUARE_SMALL;
        } else if (max_dim < 100000 && ratio < 5) {
            return SQUARE_LARGE;  
        } else if (ratio < 5) {
            return SQUARE_EXTREME;
        } else {
            return RECTANGULAR_EXTREME;  // Default for other rectangular cases
        }
    }

8 Usage Example

#include "BigDataStatMeth.hpp"

// Example usage
auto result = classify_matrix_type(...);

---
title: "classify_matrix_type"
subtitle: "C++ Function Reference"
---

## Signature

```cpp
MatrixType BigDataStatMeth::classify_matrix_type(hsize_t N, hsize_t M, hsize_t K)
```

## Description

Classify matrix type based on dimensional characteristics.

## Parameters

- `N` (`hsize_t`): Number of rows in result matrix
- `M` (`hsize_t`): Number of columns in result matrix
- `K` (`hsize_t`): Inner dimension for matrix multiplication

## Returns

MatrixType Classification result for optimization strategy selection

## Details

Analyzes matrix dimensions to determine the optimal computational strategy. Classification considers both absolute size and aspect ratios to identify rectangular vs. square-like matrices and their relative sizes.

## Caller Graph

::: {.diagram-section}
<object type="image/svg+xml" data="images/namespace_big_data_stat_meth_aa38a6e468f7c44a736e9dee6fbf4dd98_icgraph.svg" class="class-diagram">
  <img src="images/namespace_big_data_stat_meth_aa38a6e468f7c44a736e9dee6fbf4dd98_icgraph.svg" alt="Function dependencies"/>
</object>
:::

## Source Code

::: {.callout-note collapse="true"}
### Implementation

**File:** `inst/include/Utilities/Utilities.hpp` • **Lines 580-596**

```cpp
inline MatrixType classify_matrix_type(hsize_t N, hsize_t M, hsize_t K) {
        hsize_t min_dim = std::min({N, M, K});
        hsize_t max_dim = std::max({N, M, K});
        double ratio = (double)max_dim / min_dim;
        
        if (min_dim < 1000 && ratio > 100) {
            return RECTANGULAR_EXTREME;
        } else if (max_dim < 10000 && ratio < 5) {
            return SQUARE_SMALL;
        } else if (max_dim < 100000 && ratio < 5) {
            return SQUARE_LARGE;  
        } else if (ratio < 5) {
            return SQUARE_EXTREME;
        } else {
            return RECTANGULAR_EXTREME;  // Default for other rectangular cases
        }
    }
```
:::

## Usage Example

```cpp
#include "BigDataStatMeth.hpp"

// Example usage
auto result = classify_matrix_type(...);
```