diff --git a/evm/src/cpu/kernel/asm/curve/bn254/curve_arithmetic/final_exponent.asm b/evm/src/cpu/kernel/asm/curve/bn254/curve_arithmetic/final_exponent.asm index 85b1d639..f8e48807 100644 --- a/evm/src/cpu/kernel/asm/curve/bn254/curve_arithmetic/final_exponent.asm +++ b/evm/src/cpu/kernel/asm/curve/bn254/curve_arithmetic/final_exponent.asm @@ -53,18 +53,18 @@ second_exp: final_exp: // stack: val, retdest - %stack (val) -> (val, 12, val) - // stack: val, 12, val, retdest + %stack (val) -> (val, 0, val) + // stack: val, 0, val, retdest %move_fp254_12 - // stack: 12, val, retdest {12: sqr} + // stack: 0, val, retdest {0: sqr} %stack () -> (1, 1, 1) - // stack: 1, 1, 1, 12, val, retdest + // stack: 1, 1, 1, 0, val, retdest + %mstore_kernel_bn254_pairing(12) %mstore_kernel_bn254_pairing(24) - %mstore_kernel_bn254_pairing(36) - %mstore_kernel_bn254_pairing(48) - // stack: 12, val, retdest {12: sqr, 24: y0, 36: y2, 48: y4} + %mstore_kernel_bn254_pairing(36) + // stack: 0, val, retdest {0: sqr, 12: y0, 24: y2, 36: y4} %stack () -> (64, 62, 65) - // stack: 64, 62, 65, 12, val, retdest {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: 64, 62, 65, 0, val, retdest {0: sqr, 12: y0, 24: y2, 36: y4} %jump(power_loop_4) /// After computing the powers @@ -87,49 +87,49 @@ final_exp: /// return y * y2 * y1 * y0 custom_powers: - // stack: val, retdest {24: y0, 36: y2, 48: y4} - %stack () -> (24, 60, make_term_1) - // stack: 24, 60, make_term_1, val, retdest {24: y0, 36: y2, 48: y4} + // stack: val, retdest {12: y0, 24: y2, 36: y4} + %stack () -> (12, 48, make_term_1) + // stack: 12, 48, make_term_1, val, retdest {12: y0, 24: y2, 36: y4} %jump(inv_fp254_12) make_term_1: - // stack: val, retdest {36: y2, 48: y4, 60: y0^-1} - %stack () -> (36, 48, 48, make_term_2) - // stack: 36, 48, 48, make_term_2, val, retdest {36: y2, 48: y4, 60: y0^-1} + // stack: val, retdest {24: y2, 36: y4, 48: y0^-1} + %stack () -> (24, 36, 36, make_term_2) + // stack: 24, 36, 36, make_term_2, val, retdest {24: y2, 36: y4, 48: y0^-1} %jump(mul_fp254_12) make_term_2: - // stack: val, retdest {36: y2, 48: y4 * y2, 60: y0^-1} - %stack () -> (36, 48, 48, make_term_3) - // stack: 36, 48, 48, make_term_3, val, retdest {36: y2, 48: y4 * y2, 60: y0^-1} + // stack: val, retdest {24: y2, 36: y4 * y2, 48: y0^-1} + %stack () -> (24, 36, 36, make_term_3) + // stack: 24, 36, 36, make_term_3, val, retdest {24: y2, 36: y4 * y2, 48: y0^-1} %jump(mul_fp254_12) make_term_3: - // stack: val, retdest {36: y2, 48: y4 * y2^2, 60: y0^-1} - %stack () -> (60, 48, 48, final_power) - // stack: 60, 48, 48, final_power, val, retdest {36: y2, 48: y4 * y2^2, 60: y0^-1} + // stack: val, retdest {24: y2, 36: y4 * y2^2, 48: y0^-1} + %stack () -> (48, 36, 36, final_power) + // stack: 48, 36, 36, final_power, val, retdest {24: y2, 36: y4 * y2^2, 48: y0^-1} %jump(mul_fp254_12) final_power: - // stack: val, retdest {val: y , 36: y^a2 , 48: y^a1 , 60: y^a0} + // stack: val, retdest {val: y , 24: y^a2 , 36: y^a1 , 48: y^a0} %frob_fp254_12_3 - // stack: val, retdest {val: y_3, 36: y^a2 , 48: y^a1 , 60: y^a0} - %stack () -> (36, 36) + // stack: val, retdest {val: y_3, 24: y^a2 , 36: y^a1 , 48: y^a0} + %stack () -> (24, 24) %frob_fp254_12_2_ POP - // stack: val, retdest {val: y_3, 36: (y^a2)_2, 48: y^a1 , 60: y^a0} - PUSH 48 + // stack: val, retdest {val: y_3, 24: (y^a2)_2, 36: y^a1 , 48: y^a0} + PUSH 36 %frob_fp254_12_1 POP - // stack: val, retdest {val: y_3, 36: (y^a2)_2, 48: (y^a1)_1, 60: y^a0} - %stack (val) -> (36, val, val, penult_mul, val) - // stack: 36, val, val, penult_mul, val, retdest {val: y_3, 36: (y^a2)_2, 48: (y^a1)_1, 60: y^a0} + // stack: val, retdest {val: y_3, 24: (y^a2)_2, 36: (y^a1)_1, 48: y^a0} + %stack (val) -> (24, val, val, penult_mul, val) + // stack: 24, val, val, penult_mul, val, retdest {val: y_3, 24: (y^a2)_2, 36: (y^a1)_1, 48: y^a0} %jump(mul_fp254_12) penult_mul: - // stack: val, retdest {val: y_3 * (y^a2)_2, 48: (y^a1)_1, 60: y^a0} - %stack (val) -> (48, val, val, final_mul, val) - // stack: 48, val, val, final_mul, val, retdest {val: y_3 * (y^a2)_2, 48: (y^a1)_1, 60: y^a0} + // stack: val, retdest {val: y_3 * (y^a2)_2, 36: (y^a1)_1, 48: y^a0} + %stack (val) -> (36, val, val, final_mul, val) + // stack: 36, val, val, final_mul, val, retdest {val: y_3 * (y^a2)_2, 36: (y^a1)_1, 48: y^a0} %jump(mul_fp254_12) final_mul: - // stack: val, retdest {val: y_3 * (y^a2)_2 * (y^a1)_1, 60: y^a0} - %stack (val) -> (60, val, val) - // stack: 60, val, val, retdest {val: y_3 * (y^a2)_2 * (y^a1)_1, 60: y^a0} + // stack: val, retdest {val: y_3 * (y^a2)_2 * (y^a1)_1, 48: y^a0} + %stack (val) -> (48, val, val) + // stack: 48, val, val, retdest {val: y_3 * (y^a2)_2 * (y^a1)_1, 48: y^a0} %jump(mul_fp254_12) @@ -164,156 +164,156 @@ final_mul: /// y0 *= acc power_loop_4: - // stack: i , j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: i , j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP1 ISZERO - // stack: break?, i , j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: break?, i , j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jumpi(power_loop_4_end) - // stack: i , j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: i , j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %sub_const(1) - // stack: i-1, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: i-1, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP1 %mload_kernel_code(power_data_4) - // stack: abc, i-1, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: abc, i-1, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP1 %lt_const(100) - // stack: skip?, abc, i-1, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: skip?, abc, i-1, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jumpi(power_loop_4_b) - // stack: abc, i-1, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: abc, i-1, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %sub_const(100) - // stack: bc, i-1, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} - %stack () -> (48, 48, power_loop_4_b) - // stack: 48, 48, power_loop_4_b, bc, i-1, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: bc, i-1, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} + %stack () -> (36, 36, power_loop_4_b) + // stack: 36, 36, power_loop_4_b, bc, i-1, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP8 - // stack: sqr, 48, 48, power_loop_4_b, bc, i-1, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: sqr, 36, 36, power_loop_4_b, bc, i-1, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jump(mul_fp254_12) power_loop_4_b: - // stack: bc, i, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: bc, i, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP1 %lt_const(10) - // stack: skip?, bc, i, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: skip?, bc, i, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jumpi(power_loop_4_c) - // stack: bc, i, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: bc, i, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %sub_const(10) - // stack: c, i, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} - %stack () -> (36, 36, power_loop_4_c) - // stack: 36, 36, power_loop_4_c, c, i, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: c, i, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} + %stack () -> (24, 24, power_loop_4_c) + // stack: 24, 24, power_loop_4_c, c, i, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP8 - // stack: sqr, 36, 36, power_loop_4_c, c, i, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: sqr, 24, 24, power_loop_4_c, c, i, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jump(mul_fp254_12) power_loop_4_c: - // stack: c, i, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: c, i, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} ISZERO - // stack: skip?, i, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: skip?, i, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jumpi(power_loop_4_sq) - // stack: i, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} - %stack () -> (24, 24, power_loop_4_sq) - // stack: 24, 24, power_loop_4_sq, i, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: i, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} + %stack () -> (12, 12, power_loop_4_sq) + // stack: 12, 12, power_loop_4_sq, i, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP7 - // stack: sqr, 24, 24, power_loop_4_sq, i, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: sqr, 12, 12, power_loop_4_sq, i, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jump(mul_fp254_12) power_loop_4_sq: - // stack: i, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: i, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} PUSH power_loop_4 - // stack: power_loop_4, i, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: power_loop_4, i, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP5 DUP1 - // stack: sqr, sqr, power_loop_4, i, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: sqr, sqr, power_loop_4, i, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jump(square_fp254_12) power_loop_4_end: - // stack: 0, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: 0, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} POP - // stack: j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} - %stack () -> (48, 48, power_loop_2) - // stack: 48, 48, power_loop_2, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} + %stack () -> (36, 36, power_loop_2) + // stack: 36, 36, power_loop_2, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP6 - // stack: sqr, 48, 48, power_loop_2, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: sqr, 36, 36, power_loop_2, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jump(mul_fp254_12) power_loop_2: - // stack: j , k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: j , k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP1 ISZERO - // stack: break?, j , k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: break?, j , k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jumpi(power_loop_2_end) - // stack: j , k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: j , k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %sub_const(1) - // stack: j-1, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: j-1, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP1 %mload_kernel_code(power_data_2) - // stack: ab, j-1, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: ab, j-1, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP1 %lt_const(10) - // stack: skip?, ab, j-1, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: skip?, ab, j-1, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jumpi(power_loop_2_b) - // stack: ab, j-1, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: ab, j-1, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %sub_const(10) - // stack: b, j-1, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} - %stack () -> (36, 36, power_loop_2_b) - // stack: 36, 36, power_loop_2_b, b, j-1, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: b, j-1, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} + %stack () -> (24, 24, power_loop_2_b) + // stack: 24, 24, power_loop_2_b, b, j-1, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP7 - // stack: sqr, 36, 36, power_loop_2_b, b, j-1, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: sqr, 24, 24, power_loop_2_b, b, j-1, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jump(mul_fp254_12) power_loop_2_b: - // stack: b, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: b, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} ISZERO - // stack: skip?, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: skip?, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jumpi(power_loop_2_sq) - // stack: j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} - %stack () -> (24, 24, power_loop_2_sq) - // stack: 24, 24, power_loop_2_sq, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} + %stack () -> (12, 12, power_loop_2_sq) + // stack: 12, 12, power_loop_2_sq, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP6 - // stack: sqr, 24, 24, power_loop_2_sq, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: sqr, 12, 12, power_loop_2_sq, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jump(mul_fp254_12) power_loop_2_sq: - // stack: j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} PUSH power_loop_2 - // stack: power_loop_2, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: power_loop_2, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP4 DUP1 - // stack: sqr, sqr, power_loop_2, j, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: sqr, sqr, power_loop_2, j, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jump(square_fp254_12) power_loop_2_end: - // stack: 0, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: 0, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} POP - // stack: k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} - %stack () -> (36, 36, power_loop_0) - // stack: 36, 36, power_loop_0, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} + %stack () -> (24, 24, power_loop_0) + // stack: 24, 24, power_loop_0, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP5 - // stack: sqr, 36, 36, power_loop_0, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: sqr, 24, 24, power_loop_0, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jump(mul_fp254_12) power_loop_0: - // stack: k , sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: k , sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP1 ISZERO - // stack: break?, k , sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: break?, k , sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jumpi(power_loop_0_end) - // stack: k , sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: k , sqr {0: sqr, 12: y0, 24: y2, 36: y4} %sub_const(1) - // stack: k-1, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: k-1, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP1 %mload_kernel_code(power_data_0) - // stack: a, k-1, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: a, k-1, sqr {0: sqr, 12: y0, 24: y2, 36: y4} ISZERO - // stack: skip?, k-1, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: skip?, k-1, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jumpi(power_loop_0_sq) - // stack: k-1, sqr {12: sqr, 24: y0, 36: y2, 48: y4} - %stack () -> (24, 24, power_loop_0_sq) - // stack: 24, 24, power_loop_0_sq, k-1, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: k-1, sqr {0: sqr, 12: y0, 24: y2, 36: y4} + %stack () -> (12, 12, power_loop_0_sq) + // stack: 12, 12, power_loop_0_sq, k-1, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP5 - // stack: sqr, 24, 24, power_loop_0_sq, k-1, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: sqr, 12, 12, power_loop_0_sq, k-1, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jump(mul_fp254_12) power_loop_0_sq: - // stack: k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} PUSH power_loop_0 - // stack: power_loop_0, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: power_loop_0, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} DUP3 DUP1 - // stack: sqr, sqr, power_loop_0, k, sqr {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: sqr, sqr, power_loop_0, k, sqr {0: sqr, 12: y0, 24: y2, 36: y4} %jump(square_fp254_12) power_loop_0_end: - // stack: 0, sqr {12: sqr, 24: y0, 36: y2, 48: y4} - %stack (i, sqr) -> (24, sqr, 24, custom_powers) - // stack: 24, sqr, 24, custom_powers {12: sqr, 24: y0, 36: y2, 48: y4} + // stack: 0, sqr {0: sqr, 12: y0, 24: y2, 36: y4} + %stack (i, sqr) -> (12, sqr, 12, custom_powers) + // stack: 12, sqr, 12, custom_powers {0: sqr, 12: y0, 24: y2, 36: y4} %jump(mul_fp254_12) diff --git a/evm/src/cpu/kernel/asm/curve/bn254/curve_arithmetic/miller_loop.asm b/evm/src/cpu/kernel/asm/curve/bn254/curve_arithmetic/miller_loop.asm index ce43d4f1..87537bf2 100644 --- a/evm/src/cpu/kernel/asm/curve/bn254/curve_arithmetic/miller_loop.asm +++ b/evm/src/cpu/kernel/asm/curve/bn254/curve_arithmetic/miller_loop.asm @@ -55,6 +55,7 @@ miller_return: // stack: times, O, P, Q, out, retdest %stack (times, O: 2, P: 2, Q: 4, out, retdest) -> (retdest) // stack: retdest + %clear_line JUMP miller_one: @@ -281,3 +282,12 @@ after_add: // stack: (p1x - p2x)*qy_ %mstore_kernel_bn254_pairing(21) %endmacro + +%macro clear_line + %stack () -> (0, 0, 0, 0, 0) + %mstore_kernel_bn254_pairing(12) + %mstore_kernel_bn254_pairing(14) + %mstore_kernel_bn254_pairing(15) + %mstore_kernel_bn254_pairing(20) + %mstore_kernel_bn254_pairing(21) +%endmacro diff --git a/evm/src/cpu/kernel/asm/curve/bn254/field_arithmetic/degree_12_mul.asm b/evm/src/cpu/kernel/asm/curve/bn254/field_arithmetic/degree_12_mul.asm index 7c0538f4..ca32a30d 100644 --- a/evm/src/cpu/kernel/asm/curve/bn254/field_arithmetic/degree_12_mul.asm +++ b/evm/src/cpu/kernel/asm/curve/bn254/field_arithmetic/degree_12_mul.asm @@ -40,73 +40,73 @@ mul_fp254_12_1: // stack: f'g', g' , f', inA, inB, out %dup_fp254_6_0 // stack: f'g', f'g', g' , f', inA, inB, out - %store_fp254_6_sh(84) - // stack: f'g', g' , f', inA, inB, out {84: sh(f'g')} - %store_fp254_6(90) - // stack: g' , f', inA, inB, out {84: sh(f'g'), 90: f'g'} + %store_fp254_6_sh(60) + // stack: f'g', g' , f', inA, inB, out {60: sh(f'g')} + %store_fp254_6(66) + // stack: g' , f', inA, inB, out {60: sh(f'g'), 66: f'g'} DUP13 - // stack: inA, g' , f', inA, inB, out {84: sh(f'g'), 90: f'g'} + // stack: inA, g' , f', inA, inB, out {60: sh(f'g'), 66: f'g'} DUP15 - // stack: inB, inA, g' , f', inA, inB, out {84: sh(f'g'), 90: f'g'} + // stack: inB, inA, g' , f', inA, inB, out {60: sh(f'g'), 66: f'g'} %load_fp254_6 - // stack: g , inA, g' , f', inA, inB, out {84: sh(f'g'), 90: f'g'} + // stack: g , inA, g' , f', inA, inB, out {60: sh(f'g'), 66: f'g'} %stack (f: 6, x, g: 6) -> (g, x, f) - // stack: g', inA, g , f', inA, inB, out {84: sh(f'g'), 90: f'g'} + // stack: g', inA, g , f', inA, inB, out {60: sh(f'g'), 66: f'g'} %dup_fp254_6_7 - // stack: g,g', inA, g , f', inA, inB, out {84: sh(f'g'), 90: f'g'} + // stack: g,g', inA, g , f', inA, inB, out {60: sh(f'g'), 66: f'g'} %add_fp254_6 - // stack: g+g', inA, g , f', inA, inB, out {84: sh(f'g'), 90: f'g'} + // stack: g+g', inA, g , f', inA, inB, out {60: sh(f'g'), 66: f'g'} %stack (f: 6, x, g: 6) -> (g, x, f) - // stack: g, inA, g+g', f', inA, inB, out {84: sh(f'g'), 90: f'g'} + // stack: g, inA, g+g', f', inA, inB, out {60: sh(f'g'), 66: f'g'} PUSH mul_fp254_12_2 - // stack: mul_fp254_12_2, g, inA, g+g', f', inA, inB, out {84: sh(f'g'), 90: f'g'} + // stack: mul_fp254_12_2, g, inA, g+g', f', inA, inB, out {60: sh(f'g'), 66: f'g'} SWAP7 - // stack: inA, g, mul_fp254_12_2, g+g', f', inA, inB, out {84: sh(f'g'), 90: f'g'} + // stack: inA, g, mul_fp254_12_2, g+g', f', inA, inB, out {60: sh(f'g'), 66: f'g'} %load_fp254_6 - // stack: f, g, mul_fp254_12_2, g+g', f', inA, inB, out {84: sh(f'g'), 90: f'g'} + // stack: f, g, mul_fp254_12_2, g+g', f', inA, inB, out {60: sh(f'g'), 66: f'g'} %jump(mul_fp254_6) mul_fp254_12_2: - // stack: fg, g+g', f', inA, inB, out {84: sh(f'g'), 90: f'g'} - %store_fp254_6(96) - // stack: g+g', f', inA, inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: fg, g+g', f', inA, inB, out {60: sh(f'g'), 66: f'g'} + %store_fp254_6(72) + // stack: g+g', f', inA, inB, out {60: sh(f'g'), 66: f'g', 72: fg} %stack (x: 6, y: 6) -> (y, x) - // stack: f', g+g', inA, inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: f', g+g', inA, inB, out {60: sh(f'g'), 66: f'g', 72: fg} PUSH mul_fp254_12_3 - // stack: mul_fp254_12_3, f', g+g', inA, inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: mul_fp254_12_3, f', g+g', inA, inB, out {60: sh(f'g'), 66: f'g', 72: fg} SWAP13 - // stack: inA, f', g+g', mul_fp254_12_3, inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: inA, f', g+g', mul_fp254_12_3, inB, out {60: sh(f'g'), 66: f'g', 72: fg} %load_fp254_6 - // stack: f,f', g+g', mul_fp254_12_3, inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: f,f', g+g', mul_fp254_12_3, inB, out {60: sh(f'g'), 66: f'g', 72: fg} %add_fp254_6 - // stack: f+f', g+g', mul_fp254_12_3, inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: f+f', g+g', mul_fp254_12_3, inB, out {60: sh(f'g'), 66: f'g', 72: fg} %jump(mul_fp254_6) mul_fp254_12_3: - // stack: (f+f')(g+g'), inB, out {84: sh(f'g'), 90: f'g', 96: fg} - %load_fp254_6(96) - // stack: fg, (f+f')(g+g'), inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: (f+f')(g+g'), inB, out {60: sh(f'g'), 66: f'g', 72: fg} + %load_fp254_6(72) + // stack: fg, (f+f')(g+g'), inB, out {60: sh(f'g'), 66: f'g', 72: fg} %stack (x: 6, y: 6) -> (y, x) - // stack: (f+f')(g+g'), fg, inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: (f+f')(g+g'), fg, inB, out {60: sh(f'g'), 66: f'g', 72: fg} %dup_fp254_6_6 - // stack: fg, (f+f')(g+g'), fg, inB, out {84: sh(f'g'), 90: f'g', 96: fg} - %load_fp254_6(90) - // stack: f'g',fg, (f+f')(g+g'), fg, inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: fg, (f+f')(g+g'), fg, inB, out {60: sh(f'g'), 66: f'g', 72: fg} + %load_fp254_6(66) + // stack: f'g',fg, (f+f')(g+g'), fg, inB, out {60: sh(f'g'), 66: f'g', 72: fg} %add_fp254_6 - // stack: f'g'+fg, (f+f')(g+g'), fg, inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: f'g'+fg, (f+f')(g+g'), fg, inB, out {60: sh(f'g'), 66: f'g', 72: fg} %subr_fp254_6 - // stack: (f+f')(g+g') - (f'g'+fg), fg, inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: (f+f')(g+g') - (f'g'+fg), fg, inB, out {60: sh(f'g'), 66: f'g', 72: fg} DUP14 %add_const(6) - // stack: out', (f+f')(g+g') - (f'g'+fg), fg, inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: out', (f+f')(g+g') - (f'g'+fg), fg, inB, out {60: sh(f'g'), 66: f'g', 72: fg} %store_fp254_6 - // stack: fg, inB, out {84: sh(f'g'), 90: f'g', 96: fg} - %load_fp254_6(84) - // stack: sh(f'g') , fg, inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: fg, inB, out {60: sh(f'g'), 66: f'g', 72: fg} + %load_fp254_6(60) + // stack: sh(f'g') , fg, inB, out {60: sh(f'g'), 66: f'g', 72: fg} %add_fp254_6 - // stack: sh(f'g') + fg, inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: sh(f'g') + fg, inB, out {60: sh(f'g'), 66: f'g', 72: fg} DUP8 - // stack: out, sh(f'g') + fg, inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: out, sh(f'g') + fg, inB, out {60: sh(f'g'), 66: f'g', 72: fg} %store_fp254_6 - // stack: inB, out {84: sh(f'g'), 90: f'g', 96: fg} + // stack: inB, out {60: sh(f'g'), 66: f'g', 72: fg} %pop2 JUMP diff --git a/evm/src/cpu/kernel/asm/curve/bn254/field_arithmetic/inverse.asm b/evm/src/cpu/kernel/asm/curve/bn254/field_arithmetic/inverse.asm index 6214f385..ce4602a9 100644 --- a/evm/src/cpu/kernel/asm/curve/bn254/field_arithmetic/inverse.asm +++ b/evm/src/cpu/kernel/asm/curve/bn254/field_arithmetic/inverse.asm @@ -32,16 +32,19 @@ global inv_fp254_12: // stack: out, f^-1, inp, out, retdest %store_fp254_12 // stack: inp, out, retdest - %stack (inp, out) -> (inp, out, 72, check_inv_fp254_12) - // stack: inp, out, 72, check_inv_fp254_12, retdest + %stack (inp, out) -> (inp, out, 60, check_inv_fp254_12) + // stack: inp, out, 60, check_inv_fp254_12, retdest %jump(mul_fp254_12) check_inv_fp254_12: // stack: retdest - PUSH 72 + PUSH 60 %load_fp254_12 // stack: unit?, retdest %assert_eq_unit_fp254_12 // stack: retdest + PUSH 0 + // stack: 0, retdest + %mstore_kernel_bn254_pairing(60) JUMP %macro prover_inv_fp254_12 diff --git a/evm/src/cpu/kernel/tests/bn254.rs b/evm/src/cpu/kernel/tests/bn254.rs index 148847da..44589831 100644 --- a/evm/src/cpu/kernel/tests/bn254.rs +++ b/evm/src/cpu/kernel/tests/bn254.rs @@ -56,9 +56,9 @@ fn test_bn_mul_fp6() -> Result<()> { } fn run_bn_mul_fp12(f: Fp12, g: Fp12, label: &str) -> Vec { - let in0: usize = 200; - let in1: usize = 212; - let out: usize = 224; + let in0: usize = 100; + let in1: usize = 112; + let out: usize = 124; let mut stack = vec![ U256::from(in0), @@ -126,7 +126,7 @@ fn test_bn_frob_fp6() -> Result<()> { } fn run_bn_frob_fp12(f: Fp12, n: usize) -> Vec { - let ptr: usize = 200; + let ptr: usize = 100; let setup = InterpreterMemoryInitialization { label: format!("test_frob_fp254_12_{}", n), stack: vec![U256::from(ptr)], @@ -152,8 +152,8 @@ fn test_bn_frob_fp12() -> Result<()> { #[test] fn test_bn_inv_fp12() -> Result<()> { - let ptr: usize = 200; - let inv: usize = 212; + let ptr: usize = 100; + let inv: usize = 112; let mut rng = rand::thread_rng(); let f: Fp12 = rng.gen::>(); @@ -174,7 +174,7 @@ fn test_bn_inv_fp12() -> Result<()> { #[test] fn test_bn_final_exponent() -> Result<()> { - let ptr: usize = 200; + let ptr: usize = 100; let mut rng = rand::thread_rng(); let f: Fp12 = rng.gen::>(); @@ -254,8 +254,8 @@ pub const TWISTED_GENERATOR: TwistedCurve = { #[test] fn test_bn_miller() -> Result<()> { - let ptr: usize = 200; - let out: usize = 206; + let ptr: usize = 100; + let out: usize = 106; let inputs: Vec = vec![ CURVE_GENERATOR.x.val, CURVE_GENERATOR.y.val, @@ -282,8 +282,8 @@ fn test_bn_miller() -> Result<()> { #[test] fn test_bn_pairing() -> Result<()> { - let out: usize = 200; - let ptr: usize = 212; + let out: usize = 100; + let ptr: usize = 112; let inputs: Vec = vec![ CURVE_GENERATOR.x.val,