Skip to content

new VDF opt #829

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
vird wants to merge 1 commit into
base: master
Choose a base branch
from
Open

new VDF opt #829

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 15 additions & 5 deletions apps/arweave/c_src/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,17 +14,19 @@ else
CXXFLAGS ?= -O3
endif

# _mm_crc32_u32 support
CFLAGS += -msse4.2
CXXFLAGS += -msse4.2
UNAME_SYS := $(shell uname -s)
ifeq ($(UNAME_SYS), Linux)
# _mm_crc32_u32 support
CFLAGS += -msse4.2
CXXFLAGS += -msse4.2
endif

ERTS_INCLUDE_DIR ?= $(shell erl -noshell -eval 'io:format("~ts/erts-~ts/include/", [code:root_dir(), erlang:system_info(version)]).' -s init stop)
ERL_INTERFACE_INCLUDE_DIR ?= $(shell erl -noshell -eval 'io:format("~ts", [code:lib_dir(erl_interface, include)]).' -s init stop)
ERL_INTERFACE_LIB_DIR ?= $(shell erl -noshell -eval 'io:format("~ts", [code:lib_dir(erl_interface, lib)]).' -s init stop)

# System type and C compiler/flags.

UNAME_SYS := $(shell uname -s)
ifeq ($(UNAME_SYS), Darwin)
OSX_CPU_ARCH ?= x86_64
# nix systems may not have sysctl where uname -m will return the correct arch
Expand Down Expand Up @@ -84,12 +86,20 @@ RX4096_OBJECTS = $(addsuffix .o, $(basename $(RX4096_SOURCES)))
RXSQUARED_OBJECTS = $(addsuffix .o, $(basename $(RXSQUARED_SOURCES)))
VDF_OBJECTS = $(addsuffix .o, $(basename $(VDF_SOURCES)))

UNAME_SYS := $(shell uname -s)
# NOTE tabs here will cause build fail
ifeq ($(UNAME_SYS), Linux)
$(C_SRC_DIR)/vdf/vdf_exp_x86.o: CXXFLAGS += -msha
endif
ifeq ($(UNAME_SYS), Darwin)
$(C_SRC_DIR)/vdf/vdf_exp_arm.o: CXXFLAGS += -march=armv8-a+crypto
$(C_SRC_DIR)/vdf/vdf_exp_arm2.o: CXXFLAGS += -march=armv8-a+crypto
endif
ifeq ($(UNAME_SYS), Darwin)
VDF_ARM_ASM_OBJ = $(C_SRC_DIR)/vdf/sha256-armv8.o
VDF_OBJECTS += $(VDF_ARM_ASM_OBJ)
$(VDF_ARM_ASM_OBJ): $(C_SRC_DIR)/vdf/sha256-armv8.S
@echo "Assembling ARM64 specific file: $<"
clang -O3 -arch arm64 -c $(C_SRC_DIR)/vdf/sha256-armv8.S -o $(VDF_ARM_ASM_OBJ)
endif

# Verbosity.
Expand Down
2 changes: 1 addition & 1 deletion apps/arweave/c_src/vdf/ar_vdf_nif.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,7 +51,7 @@ static int vdf_load(ErlNifEnv* env, void** priv, ERL_NIF_TERM load_info) {
{
int val = 0; size_t len = sizeof(val);
if (sysctlbyname("hw.optional.arm.FEAT_SHA256", &val, &len, NULL, 0) == 0 && val != 0) {
vdf_sha2_exp_ptr = vdf_sha2_exp_arm;
vdf_sha2_exp_ptr = vdf_sha2_exp_arm2;
return 0;
}
}
Expand Down
268 changes: 268 additions & 0 deletions apps/arweave/c_src/vdf/sha256-armv8.S
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,268 @@
.text
.globl _sha256_block_vdf_order
.align 6
_sha256_block_vdf_order:
stp x29,x30,[sp,#-16]!
add x29,sp,#0
adr x3,K2564

ld1 {v4.16b,v5.16b,v6.16b,v7.16b},[x1]
ld1 {v0.4s,v1.4s},[x0]

ld1 {v16.4s},[x3],#16
add v16.4s,v16.4s,v4.4s
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v17.4s},[x3],#16
add v17.4s,v17.4s,v5.4s
orr v2.16b,v0.16b,v0.16b
.long 0x5e114020 //sha256h v0.16b,v1.16b,v17.4s
.long 0x5e115041 //sha256h2 v1.16b,v2.16b,v17.4s
orr v20.16b,v0.16b,v0.16b
orr v21.16b,v1.16b,v1.16b
ld1 {v18.4s,v19.4s},[x0]

Loop_hw:
sub x2,x2,#1
// rev32 v4.16b,v4.16b
// rev32 v5.16b,v5.16b
// rev32 v6.16b,v6.16b
// rev32 v7.16b,v7.16b

ld1 {v4.16b,v5.16b},[x1]
orr v0.16b,v20.16b,v20.16b
orr v1.16b,v21.16b,v21.16b

.long 0x5e2828a4 //sha256su0 v4.16b,v5.16b
.long 0x5e0760c4 //sha256su1 v4.16b,v6.16b,v7.16b
ld1 {v16.4s},[x3],#16
.long 0x5e2828c5 //sha256su0 v5.16b,v6.16b
.long 0x5e0460e5 //sha256su1 v5.16b,v7.16b,v4.16b
ld1 {v17.4s},[x3],#16
add v16.4s,v16.4s,v6.4s
.long 0x5e2828e6 //sha256su0 v6.16b,v7.16b
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s
.long 0x5e056086 //sha256su1 v6.16b,v4.16b,v5.16b
ld1 {v16.4s},[x3],#16
add v17.4s,v17.4s,v7.4s
.long 0x5e282887 //sha256su0 v7.16b,v4.16b
orr v2.16b,v0.16b,v0.16b
.long 0x5e114020 //sha256h v0.16b,v1.16b,v17.4s
.long 0x5e115041 //sha256h2 v1.16b,v2.16b,v17.4s
.long 0x5e0660a7 //sha256su1 v7.16b,v5.16b,v6.16b
ld1 {v17.4s},[x3],#16
add v16.4s,v16.4s,v4.4s
.long 0x5e2828a4 //sha256su0 v4.16b,v5.16b
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s
.long 0x5e0760c4 //sha256su1 v4.16b,v6.16b,v7.16b
ld1 {v16.4s},[x3],#16
add v17.4s,v17.4s,v5.4s
.long 0x5e2828c5 //sha256su0 v5.16b,v6.16b
orr v2.16b,v0.16b,v0.16b
.long 0x5e114020 //sha256h v0.16b,v1.16b,v17.4s
.long 0x5e115041 //sha256h2 v1.16b,v2.16b,v17.4s
.long 0x5e0460e5 //sha256su1 v5.16b,v7.16b,v4.16b
ld1 {v17.4s},[x3],#16
add v16.4s,v16.4s,v6.4s
.long 0x5e2828e6 //sha256su0 v6.16b,v7.16b
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s
.long 0x5e056086 //sha256su1 v6.16b,v4.16b,v5.16b
ld1 {v16.4s},[x3],#16
add v17.4s,v17.4s,v7.4s
.long 0x5e282887 //sha256su0 v7.16b,v4.16b
orr v2.16b,v0.16b,v0.16b
.long 0x5e114020 //sha256h v0.16b,v1.16b,v17.4s
.long 0x5e115041 //sha256h2 v1.16b,v2.16b,v17.4s
.long 0x5e0660a7 //sha256su1 v7.16b,v5.16b,v6.16b
ld1 {v17.4s},[x3],#16
add v16.4s,v16.4s,v4.4s
.long 0x5e2828a4 //sha256su0 v4.16b,v5.16b
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s
.long 0x5e0760c4 //sha256su1 v4.16b,v6.16b,v7.16b
ld1 {v16.4s},[x3],#16
add v17.4s,v17.4s,v5.4s
.long 0x5e2828c5 //sha256su0 v5.16b,v6.16b
orr v2.16b,v0.16b,v0.16b
.long 0x5e114020 //sha256h v0.16b,v1.16b,v17.4s
.long 0x5e115041 //sha256h2 v1.16b,v2.16b,v17.4s
.long 0x5e0460e5 //sha256su1 v5.16b,v7.16b,v4.16b
ld1 {v17.4s},[x3],#16
add v16.4s,v16.4s,v6.4s
.long 0x5e2828e6 //sha256su0 v6.16b,v7.16b
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s
.long 0x5e056086 //sha256su1 v6.16b,v4.16b,v5.16b
ld1 {v16.4s},[x3],#16
add v17.4s,v17.4s,v7.4s
.long 0x5e282887 //sha256su0 v7.16b,v4.16b
orr v2.16b,v0.16b,v0.16b
.long 0x5e114020 //sha256h v0.16b,v1.16b,v17.4s
.long 0x5e115041 //sha256h2 v1.16b,v2.16b,v17.4s
.long 0x5e0660a7 //sha256su1 v7.16b,v5.16b,v6.16b
ld1 {v17.4s},[x3],#16
add v16.4s,v16.4s,v4.4s
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3],#16
add v17.4s,v17.4s,v5.4s
orr v2.16b,v0.16b,v0.16b
.long 0x5e114020 //sha256h v0.16b,v1.16b,v17.4s
.long 0x5e115041 //sha256h2 v1.16b,v2.16b,v17.4s

ld1 {v17.4s},[x3],#16
add v16.4s,v16.4s,v6.4s
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

add v17.4s,v17.4s,v7.4s
orr v2.16b,v0.16b,v0.16b
.long 0x5e114020 //sha256h v0.16b,v1.16b,v17.4s
.long 0x5e115041 //sha256h2 v1.16b,v2.16b,v17.4s

add v6.4s,v0.4s,v18.4s
add v7.4s,v1.4s,v19.4s

//64B

ld1 {v16.4s},[x3],#16
orr v2.16b,v6.16b,v6.16b
orr v0.16b,v6.16b,v6.16b
orr v1.16b,v7.16b,v7.16b
//.long 0x5e1040e0 //sha256h v0.16b,v7.16b,v16.4s
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3],#16
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3],#16
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3],#16
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3],#16
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3],#16
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3],#16
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3],#16
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3],#16
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3],#16
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3],#16
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3],#16
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3],#16
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3],#16
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3],#16
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

ld1 {v16.4s},[x3]
sub x3,x3,#128*4-48 // rewind
orr v2.16b,v0.16b,v0.16b
.long 0x5e104020 //sha256h v0.16b,v1.16b,v16.4s
.long 0x5e105041 //sha256h2 v1.16b,v2.16b,v16.4s

add v6.4s,v0.4s,v6.4s
add v7.4s,v1.4s,v7.4s

cbnz x2,Loop_hw

st1 {v6.4s,v7.4s},[x0]

ldr x29,[sp],#16
ret

K2564:
.long 0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
.long 0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
.long 0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
.long 0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
.long 0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
.long 0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
.long 0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
.long 0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
.long 0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
.long 0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
.long 0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
.long 0xd192e819,0xd6990624,0xf40e3585,0x106aa070
.long 0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
.long 0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
.long 0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
.long 0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
.long 0xC28A2F98,0x71374491,0xB5C0FBCF,0xE9B5DBA5 //64B
.long 0x3956C25B,0x59F111F1,0x923F82A4,0xAB1C5ED5
.long 0xD807AA98,0x12835B01,0x243185BE,0x550C7DC3
.long 0x72BE5D74,0x80DEB1FE,0x9BDC06A7,0xC19BF374
.long 0x649B69C1,0xF0FE4786,0x0FE1EDC6,0x240CF254
.long 0x4FE9346F,0x6CC984BE,0x61B9411E,0x16F988FA
.long 0xF2C65152,0xA88E5A6D,0xB019FC65,0xB9D99EC7
.long 0x9A1231C3,0xE70EEAA0,0xFDB1232B,0xC7353EB0
.long 0x3069BAD5,0xCB976D5F,0x5A0F118F,0xDC1EEEFD
.long 0x0A35B689,0xDE0B7A04,0x58F4CA9D,0xE15D5B16
.long 0x007F3E86,0x37088980,0xA507EA32,0x6FAB9537
.long 0x17406110,0x0D8CD6F1,0xCDAA3B6D,0xC0BBBE37
.long 0x83613BDA,0xDB48A363,0x0B02E931,0x6FD15CA7
.long 0x521AFACA,0x31338431,0x6ED41A95,0x6D437890
.long 0xC39C91F2,0x9ECCABBD,0xB5C9A0E6,0x532FB63C
.long 0xD2C741C6,0x07237EA3,0xA4954B68,0x4C191D76
.long 0 //terminator
12 changes: 0 additions & 12 deletions apps/arweave/c_src/vdf/vdf.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -48,18 +48,6 @@ struct vdf_sha_verify_thread_arg {
int checkpointIdx;
};

void long_add(unsigned char* saltBuffer, int checkpointIdx) {
unsigned int acc = checkpointIdx;
// big endian from erlang
for(int i=SALT_SIZE-1;i>=0;i--) {
unsigned int value = saltBuffer[i];
value += acc;
saltBuffer[i] = value & 0xFF;
acc = value >> 8;
if (acc == 0) break;
}
}

////////////////////////////////////////////////////////////////////////////////////////////////////
// SHA
////////////////////////////////////////////////////////////////////////////////////////////////////
Expand Down
14 changes: 14 additions & 0 deletions apps/arweave/c_src/vdf/vdf.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,18 @@
const int SALT_SIZE = 32;
const int VDF_SHA_HASH_SIZE = 32;

static inline void long_add(unsigned char* saltBuffer, int checkpointIdx) {
unsigned int acc = checkpointIdx;
// big endian from erlang
for(int i=SALT_SIZE-1;i>=0;i--) {
unsigned int value = saltBuffer[i];
value += acc;
saltBuffer[i] = value & 0xFF;
acc = value >> 8;
if (acc == 0) break;
}
}

#if defined(__cplusplus)
extern "C" {
#endif
Expand All @@ -14,10 +26,12 @@ extern "C" {
void vdf_sha2(unsigned char* saltBuffer, unsigned char* seed, unsigned char* out, unsigned char* outCheckpoint, int checkpointCount, int skipCheckpointCount, int hashingIterations);
void vdf_sha2_exp_x86(unsigned char* saltBuffer, unsigned char* seed, unsigned char* out, unsigned char* outCheckpoint, int checkpointCount, int skipCheckpointCount, int hashingIterations);
void vdf_sha2_exp_arm(unsigned char* saltBuffer, unsigned char* seed, unsigned char* out, unsigned char* outCheckpoint, int checkpointCount, int skipCheckpointCount, int hashingIterations);
void vdf_sha2_exp_arm2(unsigned char* saltBuffer, unsigned char* seed, unsigned char* out, unsigned char* outCheckpoint, int checkpointCount, int skipCheckpointCount, int hashingIterations);
bool vdf_parallel_sha_verify_with_reset(unsigned char* startSaltBuffer, unsigned char* seed, int checkpointCount, int skipCheckpointCount, int hashingIterations, unsigned char* inRes, unsigned char* inCheckpoint, unsigned char* outCheckpoint, unsigned char* resetSalt, unsigned char* resetSeed, int maxThreadCount);

#if defined(__cplusplus)
}
#endif


#endif
12 changes: 0 additions & 12 deletions apps/arweave/c_src/vdf/vdf_exp_arm.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -990,18 +990,6 @@ void sha2_p2_32_32_norm_rev (unsigned char *output,
}
}


void long_add(unsigned char* saltBuffer, int checkpointIdx) {
unsigned int acc = checkpointIdx;
// big endian from erlang
for(int i=SALT_SIZE-1;i>=0;i--) {
unsigned int value = saltBuffer[i];
value += acc;
saltBuffer[i] = value & 0xFF;
acc = value >> 8;
if (acc == 0) break;
}
}
void _vdf_sha2_exp_arm(unsigned char* saltBuffer, unsigned char* seed, unsigned char* out, unsigned char* outCheckpoint, int checkpointCount, int skipCheckpointCount, int hashingIterations) {
unsigned char tempOut[VDF_SHA_HASH_SIZE];
// 2 different branches for different optimisation cases
Expand Down
Loading
Loading