$0x00, DX MOVQ DX, R13 // y[3] * y[2] MOVQ 16(SI), R14 MOVQ 24(SI), AX MULQ R14 ADDQ AX, R13 ADCQ $0x00, DX MOVQ DX, CX XORQ R15, R15 // *2 ADDQ R9, R9 ADCQ R10, R10 ADCQ R11, R11 ADCQ R12, R12 ADCQ R13, R13 ADCQ CX, CX ADCQ $0x00, R15 // Missing products MOVQ (SI), AX MULQ AX MOVQ AX, R8 MOVQ DX, R14 MOVQ 8(SI), AX MULQ AX ADDQ R14, R9 ADCQ AX, R10 ADCQ $0x00, DX MOVQ DX, R14 MOVQ 16(SI), AX MULQ AX ADDQ R14, R11 ADCQ AX, R12 ADCQ $0x00, DX MOVQ DX, R14 MOVQ 24(SI), AX MULQ AX ADDQ R14, R13 ADCQ...

millerRabinIteration(mr, b) if err != nil { // b was rejected. continue } if result == millerRabinCOMPOSITE { return false } iterations-- if iterations == 0 { return true } } } // productOfPrimes is the product of the first 74 primes higher than 2. // // The number of primes was selected to be the highest such that the product fit // in 512 bits, so to be usable for 1024 bit RSA keys. // // Higher values cause fewer Miller-Rabin tests of composites (nothing can help // with the final test on the actual prime) but...

            neural-network-focused hardware features such as AVX512_VNNI,
            AVX512_BF16, AMX, etc.
            ([Intel Cascade Lake](https://www.intel.com/content/www/us/en/products/platforms/details/cascade-lake.html)
            and newer CPUs.)
            *   [Example performance speedups.](https://medium.com/intel-analytics-software/leverage-intel-deep-learning-optimizations-in-tensorflow-129faa80ee07)