Skip to content
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

Jump to bottom

Add first NEON SIMD opcode implementations to fast interp #3859

Merged
merged 15 commits into from
Nov 5, 2024
+1,128 −47
Merged

Add first NEON SIMD opcode implementations to fast interp #3859

Show file tree
Hide file tree
Changes from 12 commits
Commits
Show all changes
15 commits
Select commit Hold shift + click to select a range
78c9a5e
Add simde library and SIMD swizzle implementation
Oct 21, 2024
71d129f
Add some loads, stores, more eqs
Oct 21, 2024
566caeb
Fix SIMD128 conversion macro
Oct 22, 2024
d60d267
Loads, splats, comparisons
Oct 23, 2024
92ae702
SIMD 128 Comparisons
Oct 23, 2024
2776da6
Fill out most trivial opcodes
Oct 24, 2024
061f7cb
Address PR comments
Oct 24, 2024
db71c1d
shifts, all_true, bitselect
Oct 24, 2024
5d1c324
Implement shuffle manually/in C
Oct 25, 2024
6421605
Tidy load macros
Oct 30, 2024
060a0b1
Only build SIMDE on ARM for now
Oct 30, 2024
d35ed4d
Addressing PR comments
Oct 31, 2024
bcaa7c6
Addressing comments, fixing CI
Nov 4, 2024
2f651f1
Define WASM_ENABLE_SIMDE=0 for platforms with build issues
Nov 4, 2024
b20c300
Define WASM_ENABLE_SIMDE in config.h
Nov 5, 2024
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
6 changes: 6 additions & 0 deletions build-scripts/config_common.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -288,6 +288,9 @@ endif ()
if (WAMR_BUILD_LIB_RATS EQUAL 1)
message (" Lib rats enabled")
endif()
if ((WAMR_BUILD_SIMDE EQUAL 1) AND (WAMR_BUILD_FAST_INTERP EQUAL 1))
message (" Lib simde enabled")
endif()
if (WAMR_BUILD_MINI_LOADER EQUAL 1)
add_definitions (-DWASM_ENABLE_MINI_LOADER=1)
message (" WASM mini loader enabled")
Expand Down Expand Up @@ -321,6 +324,9 @@ if (WAMR_BUILD_SIMD EQUAL 1)
else ()
message (" SIMD disabled due to not supported on target RISCV64")
endif ()
if(WAMR_BUILD_SIMDE EQUAL 1 AND WAMR_BUILD_TARGET MATCHES "AARCH64.*" OR "ARM.*")
add_definitions (-DWASM_ENABLE_SIMDE=1)
endif()
endif ()
if (WAMR_BUILD_AOT_STACK_FRAME EQUAL 1)
add_definitions (-DWASM_ENABLE_AOT_STACK_FRAME=1)
Expand Down
4 changes: 4 additions & 0 deletions build-scripts/runtime_lib.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -142,6 +142,10 @@ if (WAMR_BUILD_LIB_RATS EQUAL 1)
include (${IWASM_DIR}/libraries/lib-rats/lib_rats.cmake)
endif ()

if ((WAMR_BUILD_SIMDE EQUAL 1) AND (WAMR_BUILD_FAST_INTERP EQUAL 1))
include (${IWASM_DIR}/libraries/simde/simde.cmake)
endif ()

if (WAMR_BUILD_WASM_CACHE EQUAL 1)
include (${WAMR_ROOT_DIR}/build-scripts/involve_boringssl.cmake)
endif ()
Expand Down
67 changes: 67 additions & 0 deletions core/iwasm/common/wasm_runtime_common.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -73,13 +73,20 @@ STORE_U8(void *addr, uint8_t value)
*(uint8 *)addr = value;
}

static inline void
STORE_V128(void *addr, V128 value)
{
*(V128 *)addr = value;
}

/* For LOAD opcodes */
#define LOAD_I64(addr) (*(int64 *)(addr))
#define LOAD_F64(addr) (*(float64 *)(addr))
#define LOAD_I32(addr) (*(int32 *)(addr))
#define LOAD_U32(addr) (*(uint32 *)(addr))
#define LOAD_I16(addr) (*(int16 *)(addr))
#define LOAD_U16(addr) (*(uint16 *)(addr))
#define LOAD_V128(addr) (*(V128 *)(addr))
wenyongh marked this conversation as resolved.
Show resolved Hide resolved

#define STORE_PTR(addr, ptr) \
do { \
Expand Down Expand Up @@ -264,7 +271,67 @@ STORE_U16(void *addr, uint16_t value)
((uint8_t *)(addr))[0] = u.u8[0];
((uint8_t *)(addr))[1] = u.u8[1];
}

lum1n0us marked this conversation as resolved.
Show resolved Hide resolved
static inline void
STORE_V128(void *addr, V128 value)
{
uintptr_t addr_ = (uintptr_t)(addr);
union {
V128 val;
uint64 u64[2];
uint32 u32[4];
uint16 u16[8];
uint8 u8[16];
} u;

if ((addr_ & (uintptr_t)15) == 0) {
*(V128 *)addr = value;
}
else {
u.val = value;
if ((addr_ & (uintptr_t)7) == 0) {
((uint64 *)(addr))[0] = u.u64[0];
((uint64 *)(addr))[1] = u.u64[1];
}
else {
bh_assert((addr_ & (uintptr_t)3) == 0);
((uint32 *)addr)[0] = u.u32[0];
((uint32 *)addr)[1] = u.u32[1];
((uint32 *)addr)[2] = u.u32[2];
((uint32 *)addr)[3] = u.u32[3];
}
}
}

/* For LOAD opcodes */
static inline V128
LOAD_V128(void *addr)
{
uintptr_t addr1 = (uintptr_t)addr;
union {
V128 val;
uint64 u64[2];
uint32 u32[4];
uint16 u16[8];
uint8 u8[16];
} u;
if ((addr1 & (uintptr_t)15) == 0)
return *(V128 *)addr;

if ((addr1 & (uintptr_t)7) == 0) {
u.u64[0] = ((uint64 *)addr)[0];
u.u64[1] = ((uint64 *)addr)[1];
}
else {
bh_assert((addr1 & (uintptr_t)3) == 0);
u.u32[0] = ((uint32 *)addr)[0];
u.u32[1] = ((uint32 *)addr)[1];
u.u32[2] = ((uint32 *)addr)[2];
u.u32[3] = ((uint32 *)addr)[3];
jammar1 marked this conversation as resolved.
Show resolved Hide resolved
}
return u.val;
}

static inline int64
LOAD_I64(void *addr)
{
Expand Down
Loading
Loading