logos-storage/plonky2
1
0
Fork 0
mirror of https://github.com/logos-storage/plonky2.git synced 2026-01-06 07:43:10 +00:00
Code Issues Packages Projects Releases Wiki Activity

improve comments

Browse Source
This commit is contained in:
Dmitry Vagner 2022-10-28 02:01:04 -07:00
parent faeb6e0bb0
commit a8ecdc0f69
2 changed files with 99 additions and 101 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
evm/src/cpu/kernel/asm/fields/fp12_mul.asm
View File

@ -71,7 +71,7 @@ ret_stack:
/// h = fg + sh(f'g') /// h = fg + sh(f'g')
/// h' = (f+f')(g+g') - fg - f'g' /// h' = (f+f')(g+g') - fg - f'g'
/// ///
/// memory offsets [ind' = ind+6] /// memory pointers [ind' = ind+6]
/// {in0: f, in0: f', in1: g, in1':g', out: h, out': h'} /// {in0: f, in0: f', in1: g, in1':g', out: h, out': h'}
/// ///
/// f, f', g, g' consist of six elements on the stack /// f, f', g, g' consist of six elements on the stack

198
evm/src/cpu/kernel/asm/fields/fp6_macros.asm
View File

@ -1,31 +1,26 @@
// cost: 6 loads + 6 dup/swaps + 5 adds = 6*4 + 6*1 + 5*2 = 40 // cost: 6 loads + 6 dup/swaps + 5 adds = 6*4 + 6*1 + 5*2 = 40
%macro load_fp6 %macro load_fp6
// stack: offset // stack: ptr
DUP1 DUP1 %add_const(4)
%add_const(4) // stack: ind4, ptr
// stack: ind4, offset
%mload_kernel_general %mload_kernel_general
// stack: x4, offset // stack: x4, ptr
DUP2 DUP2 %add_const(3)
%add_const(3) // stack: ind3, x4, ptr
// stack: ind3, x4, offset
%mload_kernel_general %mload_kernel_general
// stack: x3, x4, offset // stack: x3, x4, ptr
DUP3 DUP3 %add_const(2)
%add_const(2) // stack: ind2, x3, x4, ptr
// stack: ind2, x3, x4, offset
%mload_kernel_general %mload_kernel_general
// stack: x2, x3, x4, offset // stack: x2, x3, x4, ptr
DUP4 DUP4 %add_const(1)
%add_const(1) // stack: ind1, x2, x3, x4, ptr
// stack: ind1, x2, x3, x4, offset
%mload_kernel_general %mload_kernel_general
// stack: x1, x2, x3, x4, offset // stack: x1, x2, x3, x4, ptr
DUP5 DUP5 %add_const(5)
%add_const(5) // stack: ind5, x1, x2, x3, x4, ptr
// stack: ind5, x1, x2, x3, x4, offset
%mload_kernel_general %mload_kernel_general
// stack: x5, x1, x2, x3, x4, offset // stack: x5, x1, x2, x3, x4, ptr
SWAP5 SWAP5
// stack: ind0, x1, x2, x3, x4, x5 // stack: ind0, x1, x2, x3, x4, x5
%mload_kernel_general %mload_kernel_general
@ -33,121 +28,121 @@
%endmacro %endmacro
// cost: 6 loads + 6 pushes + 5 adds = 6*4 + 6*1 + 5*2 = 40 // cost: 6 loads + 6 pushes + 5 adds = 6*4 + 6*1 + 5*2 = 40
%macro load_fp6(offset) %macro load_fp6(ptr)
// stack: // stack:
PUSH $offset PUSH $ptr %add_const(5)
%add_const(5) // stack: ind5
%mload_kernel_general %mload_kernel_general
// stack: x5 // stack: x5
PUSH $offset PUSH $ptr %add_const(4)
%add_const(4) // stack: ind4, x5
%mload_kernel_general %mload_kernel_general
// stack: x4, x5 // stack: x4, x5
PUSH $offset PUSH $ptr %add_const(3)
%add_const(3) // stack: ind3, x4, x5
%mload_kernel_general %mload_kernel_general
// stack: x3, x4, x5 // stack: x3, x4, x5
PUSH $offset PUSH $ptr %add_const(2)
%add_const(2) // stack: ind2, x3, x4, x5
%mload_kernel_general %mload_kernel_general
// stack: x2, x3, x4, x5 // stack: x2, x3, x4, x5
PUSH $offset PUSH $ptr %add_const(1)
%add_const(1) // stack: ind1, x2, x3, x4, x5
%mload_kernel_general %mload_kernel_general
// stack: x1, x2, x3, x4, x5 // stack: x1, x2, x3, x4, x5
PUSH $offset PUSH $ptr
// stack: ind0, x1, x2, x3, x4, x5
%mload_kernel_general %mload_kernel_general
// stack: x0, x1, x2, x3, x4, x5 // stack: x0, x1, x2, x3, x4, x5
%endmacro %endmacro
// cost: 6 stores + 6 swaps/dups + 5 adds = 6*4 + 6*1 + 5*2 = 40 // cost: 6 stores + 6 swaps/dups + 5 adds = 6*4 + 6*1 + 5*2 = 40
%macro store_fp6 %macro store_fp6
// stack: offset, x0, x1, x2, x3, x4, x5 // stack: ptr, x0, x1, x2, x3, x4 , x5
SWAP5 SWAP5
DUP6 // stack: x4, x0, x1, x2, x3, ptr, x5
%add_const(4) DUP6 %add_const(4)
// stack: ind4, x4, x0, x1, x2, x3, offset, x5 // stack: ind4, x4, x0, x1, x2, x3, ptr, x5
%mstore_kernel_general %mstore_kernel_general
// stack: x0, x1, x2, x3, offset, x5 // stack: x0, x1, x2, x3, ptr, x5
DUP5 DUP5
// stack: ind0, x0, x1, x2, x3, offset, x5 // stack: ind0, x0, x1, x2, x3, ptr, x5
%mstore_kernel_general %mstore_kernel_general
// stack: x1, x2, x3, offset, x5 // stack: x1, x2, x3, ptr, x5
DUP4 DUP4 %add_const(1)
%add_const(1) // stack: ind1, x1, x2, x3, ptr, x5
// stack: ind1, x1, x2, x3, offset, x5
%mstore_kernel_general %mstore_kernel_general
// stack: x2, x3, offset, x5 // stack: x2, x3, ptr, x5
DUP3 DUP3 %add_const(2)
%add_const(2) // stack: ind2, x2, x3, ptr, x5
// stack: ind2, x2, x3, offset, x5
%mstore_kernel_general %mstore_kernel_general
// stack: x3, offset, x5 // stack: x3, ptr, x5
DUP2 DUP2 %add_const(3)
%add_const(3) // stack: ind3, x3, ptr, x5
// stack: ind3, x3, offset, x5
%mstore_kernel_general %mstore_kernel_general
// stack: offset, x5 // stack: ptr, x5
%add_const(5) %add_const(5)
// stack: ind5, x5 // stack: ind5, x5
%mstore_kernel_general %mstore_kernel_general
// stack: // stack:
%endmacro %endmacro
// cost: 6 stores + 6 pushes + 5 adds = 6*4 + 6*1 + 5*2 = 40 // cost: 6 stores + 6 pushes + 5 adds = 6*4 + 6*1 + 5*2 = 40
%macro store_fp6(offset) %macro store_fp6(ptr)
// stack: x0, x1, x2, x3, x4, x5 // stack: x0, x1, x2, x3, x4, x5
PUSH $offset PUSH $ptr
// stack: ind0, x0, x1, x2, x3, x4, x5
%mstore_kernel_general %mstore_kernel_general
// stack: x1, x2, x3, x4, x5 // stack: x1, x2, x3, x4, x5
PUSH $offset PUSH $ptr %add_const(1)
%add_const(1) // stack: ind1, x1, x2, x3, x4, x5
%mstore_kernel_general %mstore_kernel_general
// stack: x2, x3, x4, x5 // stack: x2, x3, x4, x5
PUSH $offset PUSH $ptr %add_const(2)
%add_const(2) // stack: ind2, x2, x3, x4, x5
%mstore_kernel_general %mstore_kernel_general
// stack: x3, x4, x5 // stack: x3, x4, x5
PUSH $offset PUSH $ptr %add_const(3)
%add_const(3) // stack: ind3, x3, x4, x5
%mstore_kernel_general %mstore_kernel_general
// stack: x4, x5 // stack: x4, x5
PUSH $offset PUSH $ptr %add_const(4)
%add_const(4) // stack: ind4, x4, x5
%mstore_kernel_general %mstore_kernel_general
// stack: x5 // stack: x5
PUSH $offset PUSH $ptr %add_const(5)
%add_const(5) // stack: ind5, x5
%mstore_kernel_general %mstore_kernel_general
// stack: // stack:
%endmacro %endmacro
// cost: store (40) + i9 (9) = 49 // cost: store (40) + i9 (9) = 49
%macro store_fp6_sh(offset) %macro store_fp6_sh(ptr)
// stack: x0, x1, x2, x3, x4, x5 // stack: x0, x1, x2, x3, x4, x5
PUSH $offset PUSH $ptr %add_const(2)
%add_const(2) // stack: ind2, x0, x1, x2, x3, x4, x5
%mstore_kernel_general %mstore_kernel_general
// stack: x1, x2, x3, x4, x5 // stack: x1, x2, x3, x4, x5
PUSH $offset PUSH $ptr %add_const(3)
%add_const(3) // stack: ind3, x1, x2, x3, x4, x5
%mstore_kernel_general %mstore_kernel_general
// stack: x2, x3, x4, x5 // stack: x2, x3, x4, x5
PUSH $offset PUSH $ptr %add_const(4)
%add_const(4) // stack: ind4, x2, x3, x4, x5
%mstore_kernel_general %mstore_kernel_general
// stack: x3, x4, x5 // stack: x3, x4, x5
PUSH $offset PUSH $ptr %add_const(5)
%add_const(5) // stack: ind5, x3, x4, x5
%mstore_kernel_general %mstore_kernel_general
// stack: x4, x5 // stack: x4, x5
%i9 %i9
// stack: y5, y4 // stack: y5, y4
PUSH $offset PUSH $ptr %add_const(1)
%add_const(1) // stack: ind1, y5, y4
%mstore_kernel_general %mstore_kernel_general
// stack: y4 // stack: y4
PUSH $offset PUSH $ptr
// stack: ind0, y4
%mstore_kernel_general %mstore_kernel_general
// stack: // stack:
%endmacro %endmacro
@ -156,16 +151,17 @@
%macro i9 %macro i9
// stack: a , b // stack: a , b
DUP2 DUP2
// stack: b, a, b
DUP2 DUP2
// stack: a , b, a , b // stack: a , b, a , b
PUSH 9 PUSH 9 MULFP254
MULFP254 // stack: 9a , b, a , b
SUBFP254 SUBFP254
// stack: 9a - b, a , b // stack: 9a - b, a , b
SWAP2 SWAP2
// stack: b , a, 9a - b // stack: b , a, 9a - b
PUSH 9 PUSH 9 MULFP254
MULFP254 // stack 9b , a, 9a - b
ADDFP254 ADDFP254
// stack: 9b + a, 9a - b // stack: 9b + a, 9a - b
%endmacro %endmacro
@ -234,6 +230,8 @@
%endmacro %endmacro
// cost: 16 // cost: 16
// swap two fp6 elements with a stack term separating them
// (f: 6, x, g: 6) -> (g: 6, x, f: 6)
%macro swap_fp6_hole %macro swap_fp6_hole
// stack: f0, f1, f2, f3, f4, f5, X, g0, g1, g2, g3, g4, g5 // stack: f0, f1, f2, f3, f4, f5, X, g0, g1, g2, g3, g4, g5
SWAP7 SWAP7