2034          ; mapped                 ; 2032 2032 2032 #1.1  TRIPLE PRIME
2035          ; valid                  ;      ; NV8    # 1.1  REVERSED PRIME
2036          ; mapped                 ; 2035 2035     # 1.1  REVERSED DOUBLE PRIME
2037          ; mapped                 ; 2035 2035 2035 #1.1  REVERSED TRIPLE PRIME
2038..203B    ; valid                  ;      ; NV8    # 1.1  CARET..REFERENCE MARK

which is matrix inversion * tf.round now uses banker's rounding (round to even)
semantics to match NumPy. * `tf.split` now takes arguments in a reversed order
and with different keywords. In particular, we now match NumPy order as
`tf.split(value, num_or_size_splits, axis)`. * `tf.sparse_split` now takes
arguments in reversed order and with different keywords. In particular we now
match NumPy order as `tf.sparse_split(sp_input, num_split, axis)`. NOTE: we have

hardware support for AES-GCM. // // Each input is zero-padded to 128-bit before being absorbed. func ghash(out, H *[gcmBlockSize]byte, inputs ...[]byte) { // productTable contains the first sixteen powers of the key, H. // However, they are in bit reversed order. var productTable [16]gcmFieldElement // We precompute 16 multiples of H. However, when we do lookups // into this table we'll be using bits from a field element and // therefore the bits will be in the reverse order. So normally one // would...

hardware support for AES-GCM. // // Each input is zero-padded to 128-bit before being absorbed. func ghash(out, H *[gcmBlockSize]byte, inputs ...[]byte) { // productTable contains the first sixteen powers of the key, H. // However, they are in bit reversed order. var productTable [16]gcmFieldElement // We precompute 16 multiples of H. However, when we do lookups // into this table we'll be using bits from a field element and // therefore the bits will be in the reverse order. So normally one // would...