bugfix: fix fp4 quantization with 8x4 scale factor layout (#1611)

Author: cyx-6

csrc/nv_internal/tensorrt_llm/kernels/quantization.cuh

@@ -766,8 +766,9 @@ quantize_with_block_size(
   bool isSfSwizzledLayout = layout == QuantizationSFLayout::SWIZZLED_128x4 ||
                             layout == QuantizationSFLayout::SWIZZLED_8x4;
 
-  // The number of padded rows considering 128x4 SF layout.
-  int numPaddedRowsForSf = isSfSwizzledLayout ? PadUpFn(numRows, 128) : numRows;
+  // The number of padded rows considering 128x4 or 8x4 SF layout.
+  int rowTile = (layout == QuantizationSFLayout::SWIZZLED_128x4) ? 128 : 8;
+  int numPaddedRowsForSf = isSfSwizzledLayout ? PadUpFn(numRows, rowTile) : numRows;
   int numColsForSf = isSfSwizzledLayout ? PadUpFn(numPaddedCols, 4 * SF_VEC_SIZE) : numPaddedCols;
 
   // The number of threads in the column dimension。