Leaky ReLU

Perform the Leaky ReLU (Leaky Rectified Linear Unit) activation function on an input matrix:

$$\text{C}[i][j] = \max(\alpha \cdot \text{A}[i][j], \text{A}[i][j])$$

where $\alpha$ is a small positive constant (e.g. 0.01)

The Leaky ReLU function is defined as:

$$f(x) = \begin{cases} x & \text{if } x > 0 \\ \alpha x & \text{if } x \leq 0 \end{cases}$$

Input:

• Matrix $\text{A}$ of size $M \times N$
• $\alpha$ value (slope for negative values)

Output:

• Matrix $\text{C}$ of size $M \times N$

Notes:

• Both matrices $\text{A}$ and $\text{C}$ are stored in row-major order
• This problem is adapted from KernelBench

Test Case Sizes

• 4096x4096, alpha=0.01
• 4096x4096, alpha=0.05
• 4096x4096, alpha=0.1
• 4096x4096, alpha=0.2
• 6144x4096, alpha=0.01
• 6144x4096, alpha=0.05
• 6144x4096, alpha=0.1
• 6144x4096, alpha=0.2