fp4_kernel_hopper#

NVFP4 Fake Quantization Triton kernels requiring compute capability >= 8.9 (Hopper+).

These kernels use tl.float8e4nv which requires native FP8 hardware support.

Functions

fp4_fake_quant_block

FP4 fake quantization implementation using block-pointer tiling.
fp4_fake_quant_block(x, global_amax, block_size=16, tile_rows=16, tile_cols=64, num_warps=None, num_stages=None)#

FP4 fake quantization implementation using block-pointer tiling.

Parameters:

  • x (torch.Tensor) – Input tensor of shape (M, N) or higher.

  • global_amax (torch.Tensor) – Global maximum value for scaling: one value, or one per index of x’s first dimension (e.g. per expert of a stacked (E, N, K) weight), which quantizes each slice exactly as a separate call with its own amax would.

  • block_size (int) – Number of elements per FP4 block.

  • tile_rows (int, optional) – Row tile size. Defaults to 16.

  • tile_cols (int, optional) – Column tile size. Defaults to 64. Rounded up to the nearest multiple of block_size internally.

  • num_warps (int | None, optional) – Override for Triton warps. Autotuned when None.

  • num_stages (int | None, optional) – Override for pipeline stages. Autotuned when None.

Returns:

Fake-quantized tensor matching the input shape and dtype.

Return type:

torch.Tensor