// <signedNumber>    ::= <number> | <sign><number>
// <suffix>          ::= <binarySI> | <decimalExponent> | <decimalSI>
// <binarySI>        ::= Ki | Mi | Gi | Ti | Pi | Ei
//
//	(International System of units; See: http://physics.nist.gov/cuu/Units/binary.html)
//
// <decimalSI>       ::= m | "" | k | M | G | T | P | E
//
//	(Note that 1024 = 1Ki but 1000 = 1k; I didn't choose the capitalization.)
//

        available automatically, without requiring an explicit `with
        tf.device(“/gpu:0”)`) (Fixes #14133)
    *   `tf.GradientTape` has moved out of contrib.
*   `tf.keras`:
    *   Added the fashion mnist dataset.
    *   New data preprocessing functions: `image/random_brightness`,
        `sequence/TimeseriesGenerator`, and `text/hashing_trick`.
*   Accelerated Linear Algebra (XLA):

//go:build !purego

#include "textflag.h"

// SHA512 block routine. See sha512block.go for Go equivalent.
//
// The algorithm is detailed in FIPS 180-4:
//
//  https://csrc.nist.gov/publications/fips/fips180-4/fips-180-4.pdf
//
// Wt = Mt; for 0 <= t <= 15
// Wt = SIGMA1(Wt-2) + SIGMA0(Wt-15) + Wt-16; for 16 <= t <= 79
//
// a = H0
// b = H1
// c = H2
// d = H3
// e = H4
// f = H5

//go:build !purego

#include "textflag.h"

// SHA256 block routine. See sha256block.go for Go equivalent.
//
// The algorithm is detailed in FIPS 180-4:
//
//  https://csrc.nist.gov/publications/fips/fips180-4/fips-180-4.pdf

// The avx2-version is described in an Intel White-Paper:
// "Fast SHA-256 Implementations on Intel Architecture Processors"

			0xb2, 0x61, 0x12, 0xb8, 0x2a, 0x90, 0xb7, 0x2f},
		[]byte{0xa3, 0xc2, 0x60, 0xb1, 0x0b, 0xb7, 0x28, 0x6e}, // random
		[]byte{0x56, 0x73, 0x7d, 0xfb, 0xb5, 0xa1, 0xc3, 0xde}},
}

// NIST Special Publication 800-20, Appendix A
// Key for use with Table A.1 tests
var tableA1Key = []byte{
	0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
	0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,

#define Y1    V3
#define T0    V4
#define T1    V5

#define PL    V30
#define PH    V31
/*
 * https://choucroutage.com/Papers/SideChannelAttacks/ctrsa-2011-brown.pdf "Software Implementation of the NIST Elliptic Curves Over Prime Fields"
 *
 * A = X₁×Z₂²
 * B = Y₁×Z₂³
 * C = X₂×Z₁²-A
 * D = Y₂×Z₁³-B
 * X₃ = D² - 2A×C² - C³
 * Y₃ = D×(A×C² - X₃) - B×C³
 * Z₃ = Z₁×Z₂×C
 *

#define Y0    V2
#define Y1    V3
#define T0    V4
#define T1    V5

#define PL    V30
#define PH    V31
/*
 * https://delta.cs.cinvestav.mx/~francisco/arith/julio.pdf "Software Implementation of the NIST Elliptic Curves Over Prime Fields"
 *
 * A = X₁×Z₂²
 * B = Y₁×Z₂³
 * C = X₂×Z₁²-A
 * D = Y₂×Z₁³-B
 * X₃ = D² - 2A×C² - C³
 * Y₃ = D×(A×C² - X₃) - B×C³
 * Z₃ = Z₁×Z₂×C
 *

        ::= \"+\" | \"-\" <signedNumber>    ::= <number> | <sign><number> <suffix>          ::= <binarySI> | <decimalExponent> | <decimalSI> <binarySI>        ::= Ki | Mi | Gi | Ti | Pi | Ei\n\n\t(International System of units; See: http://physics.nist.gov/cuu/Units/binary.html)\n\n<decimalSI>       ::= m | \"\" | k | M | G | T | P | E\n\n\t(Note that 1024 = 1Ki but 1000 = 1k; I didn't choose the capitalization.)\n\n<decimalExponent> ::= \"e\" <signedNumber> | \"E\" <signedNumber> ```\n\nNo matter...

        ::= \"+\" | \"-\" <signedNumber>    ::= <number> | <sign><number> <suffix>          ::= <binarySI> | <decimalExponent> | <decimalSI> <binarySI>        ::= Ki | Mi | Gi | Ti | Pi | Ei\n\n\t(International System of units; See: http://physics.nist.gov/cuu/Units/binary.html)\n\n<decimalSI>       ::= m | \"\" | k | M | G | T | P | E\n\n\t(Note that 1024 = 1Ki but 1000 = 1k; I didn't choose the capitalization.)\n\n<decimalExponent> ::= \"e\" <signedNumber> | \"E\" <signedNumber> ```\n\nNo matter...

          ::= \"+\" | \"-\" <signedNumber>    ::= <number> | <sign><number> <suffix>          ::= <binarySI> | <decimalExponent> | <decimalSI> <binarySI>        ::= Ki | Mi | Gi | Ti | Pi | Ei\n  (International System of units; See: http://physics.nist.gov/cuu/Units/binary.html)\n<decimalSI>       ::= m | \"\" | k | M | G | T | P | E\n  (Note that 1024 = 1Ki but 1000 = 1k; I didn't choose the capitalization.)\n<decimalExponent> ::= \"e\" <signedNumber> | \"E\" <signedNumber>\n\nNo matter which of...