vst1q_u32ADD TO COMPARE ADDED TO COMPARE
Arm 64-bit (64 bits)/ NEON
View official documentation
Purpose:
Stores elements from a 128-bit vector into contiguous memory.
Result:
void
Example:
#include <stdio.h>
#include <arm_neon.h>
int main() {
uint32_t input_arr[4] = {
1, 2, 3, 4
};
uint32x4_t vec = vld1q_u32(input_arr);
uint32_t res_arr[4];
vst1q_u32(res_arr, vec);
printf("Stored values:\n");
for (int i = 0; i < 4; i++) {
printf("%u\n", res_arr[i]);
}
return 0;
}Prototypes
Assembly Instruction:
ST1
Usage:
void result =
vst1q_u32(
uint32_t a, uint32x4_t b
)
LLVM-MCA Metrics:
| CPU |
Latency (Cycles) (lower is better) |
Throughput (maximum IPC) (higher is better) |
|---|---|---|
| cortex-a72 | 2 | 0.5 |
| cortex-a73 | 2 | 0.5 |
| cortex-a75 | 2 | 0.5 |
| cortex-a76 | 2 | 0.5 |
| cortex-a77 | 2 | 0.5 |
| cortex-a78 | 2 | 0.5 |
| cortex-x1 | 2 | 2.0 |
| cortex-x2 | 2 | 2.0 |
| cortex-x3 | 2 | 2.0 |
| cortex-a710 | 2 | 2.0 |
| cortex-a715 | 2 | 2.0 |
| neoverse-n1 | 2 | 2.0 |
| neoverse-n2 | 2 | 2.0 |
| neoverse-v1 | 2 | 2.0 |
| neoverse-v2 | 2 | 2.0 |
| ampere1 | 2 | 1.0 |
| ampere1a | 2 | 1.0 |
| ampere1b | 2 | 1.0 |
| cortex-a55 | 4 | 1.0 |
Performance Plot:
SIMD Intrinsics Summary
| SIMD Engines: | 6 |
| C Intrinsics: | 10444 |
| NEON: | 4353 |
| AVX2: | 405 |
| AVX512: | 4717 |
| SSE4.2: | 598 |
| VSX: | 192 |
| IBM-Z: | 179 |