
Perform 4D tensor-matrix multiplication of two tensors:
$$
C[b][i][j][k] = \sum_{l=0}^{L-1} A[b][i][j][l] \cdot B[l][k]
$$

## Input
- Tensor $A$ of size $B \times I \times J \times L$
- Matrix $B$ of size $L \times K$

## Output
- Tensor $C$ of size $B \times I \times J \times K$

## Notes:
- All tensors $\text{A}$, $\text{B}$, and $\text{C}$ are stored in row-major order
- This problem is adapted from [KernelBench](https://github.com/ScalingIntelligence/KernelBench/blob/main/KernelBench/level1/11_4D_tensor_matrix_multiplication.py)

Rank	User	GFLOPS	Runtime (ms)	GPU	Language	Date
#1	ell-hol	35.95T	2.89	H100	CUDA C++	27 days ago
#2	sagarreddypatil	26.41T	14.54	H100	Triton	about 2 months ago
#3	Firojpaudel	5.65T	32.03	H100	CUDA C++	about 2 months ago
#4	a-hamdi	5.02T	36.44	H100	CUDA C++	2 months ago
#5	harmya	4.37T	40.75	H100	CUDA C++	2 months ago
#6	Firojpaudel	615.79G	403.35	T4	CUDA C++	about 2 months ago
#7	a-hamdi	568.51G	425.87	T4	CUDA C++	2 months ago
#8	harmya	493.91G	408.92	T4	CUDA C++	2 months ago
#9	ShlokVFX	392.09G	437.96	H100	CUDA C++	about 2 months ago