So many lols

What does this code do?

370     __m128 row0, row1, row2, row3;
371     __m128 tmp0, tmp1, tmp2, tmp3;
372 
373     
374     row0 = _mm_load_ps( mat2[0] );
375     row1 = _mm_load_ps( mat2[1] );
376     row2 = _mm_load_ps( mat2[2] );
377     row3 = _mm_load_ps( mat2[3] );
378 
379      
380     
381     tmp0 = _mm_unpacklo_ps( row0, row1 );
382     tmp2 = _mm_unpacklo_ps( row2, row3 );
383     tmp1 = _mm_unpackhi_ps( row0, row1 );
384     tmp3 = _mm_unpackhi_ps( row2, row3 );
385 
386     
387     row0 = _mm_movelh_ps( tmp0, tmp2 );
388     row1 = _mm_movehl_ps( tmp2, tmp0 );
389     row2 = _mm_movelh_ps( tmp1, tmp3 );
390     row3 = _mm_movehl_ps( tmp3, tmp1 );
391 
392     
393     _mm_store_ps( mat2[0], row0 );
394     _mm_store_ps( mat2[1], row1 );
395     _mm_store_ps( mat2[2], row2 );
396     _mm_store_ps( mat2[3], row3 );
397 
398     
399     
400     ret[0][0] = mul_asm(mat1[0], mat2[0]);
401     ret[0][1] = mul_asm(mat1[0], mat2[1]);
402     ret[0][2] = mul_asm(mat1[0], mat2[2]);
403     ret[0][3] = mul_asm(mat1[0], mat2[3]);
404     ret[1][0] = mul_asm(mat1[1], mat2[0]);
405     ret[1][1] = mul_asm(mat1[1], mat2[1]);
406     ret[1][2] = mul_asm(mat1[1], mat2[2]);
407     ret[1][3] = mul_asm(mat1[1], mat2[3]);
408     ret[2][0] = mul_asm(mat1[2], mat2[0]);
409     ret[2][1] = mul_asm(mat1[2], mat2[1]);
410     ret[2][2] = mul_asm(mat1[2], mat2[2]);
411     ret[2][3] = mul_asm(mat1[2], mat2[3]);
412     ret[3][0] = mul_asm(mat1[3], mat2[0]);
413     ret[3][1] = mul_asm(mat1[3], mat2[1]);
414     ret[3][2] = mul_asm(mat1[3], mat2[2]);
415     ret[3][3] = mul_asm(mat1[3], mat2[3]);
416 

Yeah, this is for a class. Guess which one? (6.172)