rte_vect.h(3)

📦
version 25.11.3 · 9 versions
, in ubuntu resolute, ubuntu stonking

💖 rte_vect.h 💖

okay bestie, so this is the manual page for rte_vect.h from the Library Functions Manual — yes, the actual one, section 3, we're in the big leagues now. It's the DPDK header that gives you SIMD vector types so you can write generic C and let the compiler go absolutely feral with vector instructions. Sit down, hydrate, we're doing this. 🫶

✨ NAME ✨

rte_vect.h — no cap, that's the name. Moving on, gorgeous.

🌸 SYNOPSIS 🌸

#include <stdint.h>

🧚 Typedefs

  • typedef uint8_t rte_v64u8_t
  • typedef uint16_t rte_v64u16_t
  • typedef uint32_t rte_v64u32_t
  • typedef uint8_t rte_v128u8_t
  • typedef uint16_t rte_v128u16_t
  • typedef uint32_t rte_v128u32_t
  • typedef uint64_t rte_v128u64_t
  • typedef uint8_t rte_v256u8_t
  • typedef uint16_t rte_v256u16_t
  • typedef uint32_t rte_v256u32_t
  • typedef uint64_t rte_v256u64_t
  • typedef int8_t rte_v64s8_t
  • typedef int16_t rte_v64s16_t
  • typedef int32_t rte_v64s32_t
  • typedef int8_t rte_v128s8_t
  • typedef int16_t rte_v128s16_t
  • typedef int32_t rte_v128s32_t
  • typedef int64_t rte_v128s64_t
  • typedef int8_t rte_v256s8_t
  • typedef int16_t rte_v256s16_t
  • typedef int32_t rte_v256s32_t
  • typedef int64_t rte_v256s64_t

🎀 Enumerations

enum rte_vect_max_simd {
    RTE_VECT_SIMD_DISABLED = 64,
    RTE_VECT_SIMD_128 = 128,
    RTE_VECT_SIMD_256 = 256,
    RTE_VECT_SIMD_512 = 512,
    RTE_VECT_SIMD_MAX = INT16_MAX + 1
}

👑 Functions

  • uint16_t rte_vect_get_max_simd_bitwidth (void)
  • int rte_vect_set_max_simd_bitwidth (uint16_t bitwidth)

🦋 Detailed Description 🦋

SIMD vector types and control. 💅

This file defines types you can use to do vector instructions with generic C code, plus the APIs that switch that code on. Definition in file rte_vect.h.

💜 Typedef Documentation 💜

typedef uint8_t rte_v64u8_t

64 bits vector size to use with unsigned 8 bits elements. it's giving baby vector, eight tiny unsigned cells:

a = (rte_v64u8_t){ a0, a1, a2, a3, a4, a5, a6, a7 }

Definition at line 31 of file rte_vect.h.

typedef uint16_t rte_v64u16_t

64 bits vector size to use with unsigned 16 bits elements. four of them, cozy:

a = (rte_v64u16_t){ a0, a1, a2, a3 }

Definition at line 38 of file rte_vect.h.

typedef uint32_t rte_v64u32_t

64 bits vector size to use with unsigned 32 bits elements. literally just a duo:

a = (rte_v64u32_t){ a0, a1 }

Definition at line 45 of file rte_vect.h.

typedef uint8_t rte_v128u8_t

128 bits vector size to use with unsigned 8 bits elements. sixteen lil guys, a whole friend group:

a = (rte_v128u8_t){ a00, a01, a02, a03, a04, a05, a06, a07, a08, a09,
a10, a11, a12, a13, a14, a15 }

Definition at line 53 of file rte_vect.h.

typedef uint16_t rte_v128u16_t

128 bits vector size to use with unsigned 16 bits elements. eight slots, immaculate:

a = (rte_v128u16_t){ a0, a1, a2, a3, a4, a5, a6, a7 }

Definition at line 60 of file rte_vect.h.

typedef uint32_t rte_v128u32_t

128 bits vector size to use with unsigned 32 bits elements. a four-piece, iconic:

a = (rte_v128u32_t){ a0, a1, a2, a3 }

Definition at line 67 of file rte_vect.h.

typedef uint64_t rte_v128u64_t

128 bits vector size to use with unsigned 64 bits elements. the power couple:

a = (rte_v128u64_t){ a0, a1 }

Definition at line 74 of file rte_vect.h.

typedef uint8_t rte_v256u8_t

256 bits vector size to use with unsigned 8 bits elements. thirty-two of them bestie, we're SO back:

a = (rte_v256u8_t){ a00, a01, a02, a03, a04, a05, a06, a07, a08, a09,
a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20, a21, a22, a23,
a24, a25, a26, a27, a28, a29, a30, a31 }

Definition at line 84 of file rte_vect.h.

typedef uint16_t rte_v256u16_t

256 bits vector size to use with unsigned 16 bits elements. sixteen wide, main character energy:

a = (rte_v256u16_t){ a00, a01, a02, a03, a04, a05, a06, a07, a08, a09,
a10, a11, a12, a13, a14, a15 }

Definition at line 92 of file rte_vect.h.

typedef uint32_t rte_v256u32_t

256 bits vector size to use with unsigned 32 bits elements. eight of them, slay:

a = (rte_v256u32_t){ a0, a1, a2, a3, a4, a5, a6, a7 }

Definition at line 99 of file rte_vect.h.

typedef uint64_t rte_v256u64_t

256 bits vector size to use with unsigned 64 bits elements. four chonky ones:

a = (rte_v256u64_t){ a0, a1, a2, a3 }

Definition at line 106 of file rte_vect.h.

typedef int8_t rte_v64s8_t

64 bits vector size to use with 8 bits elements. signed now, but still cute:

a = (rte_v64s8_t){ a0, a1, a2, a3, a4, a5, a6, a7 }

Definition at line 116 of file rte_vect.h.

typedef int16_t rte_v64s16_t

64 bits vector size to use with 16 bits elements. signed, four cells:

a = (rte_v64s16_t){ a0, a1, a2, a3 }

Definition at line 123 of file rte_vect.h.

typedef int32_t rte_v64s32_t

64 bits vector size to use with 32 bits elements. signed duo, periodt:

a = (rte_v64s32_t){ a0, a1 }

Definition at line 130 of file rte_vect.h.

typedef int8_t rte_v128s8_t

128 bits vector size to use with 8 bits elements. signed sixteen, bless:

a = (rte_v128s8_t){ a00, a01, a02, a03, a04, a05, a06, a07, a08, a09,
a10, a11, a12, a13, a14, a15 }

Definition at line 138 of file rte_vect.h.

typedef int16_t rte_v128s16_t

128 bits vector size to use with 16 bits elements. signed eight, no notes:

a = (rte_v128s16_t){ a0, a1, a2, a3, a4, a5, a6, a7 }

Definition at line 145 of file rte_vect.h.

typedef int32_t rte_v128s32_t

128 bits vector size to use with 32 bits elements. signed four, obsessed:

a = (rte_v128s32_t){ a0, a1, a2, a3 }

Definition at line 152 of file rte_vect.h.

typedef int64_t rte_v128s64_t

128 bits vector size to use with 64 bits elements. and yes, the example really does start at a1:

a = (rte_v128s64_t){ a1, a2 }

Definition at line 159 of file rte_vect.h.

typedef int8_t rte_v256s8_t

256 bits vector size to use with 8 bits elements. signed thirty-two, absolutely unhinged (affectionate):

a = (rte_v256s8_t){ a00, a01, a02, a03, a04, a05, a06, a07, a08, a09,
a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20, a21, a22, a23,
a24, a25, a26, a27, a28, a29, a30, a31 }

Definition at line 169 of file rte_vect.h.

typedef int16_t rte_v256s16_t

256 bits vector size to use with 16 bits elements. signed sixteen, in formation:

a = (rte_v256s16_t){ a00, a01, a02, a03, a04, a05, a06, a07, a08, a09,
a10, a11, a12, a13, a14, a15 }

Definition at line 177 of file rte_vect.h.

typedef int32_t rte_v256s32_t

256 bits vector size to use with 32 bits elements. signed eight, chef's kiss:

a = (rte_v256s32_t){ a0, a1, a2, a3, a4, a5, a6, a7 }

Definition at line 184 of file rte_vect.h.

typedef int64_t rte_v256s64_t

256 bits vector size to use with 64 bits elements. signed quad, final boss:

a = (rte_v256s64_t){ a0, a1, a2, a3 }

Definition at line 191 of file rte_vect.h.

🌈 Enumeration Type Documentation 🌈

enum rte_vect_max_simd

The max SIMD bitwidth value to limit vector path selection. this is the vibe check for how wide your code is allowed to go 💅

Enumerator

  • RTE_VECT_SIMD_DISABLED — Limits path selection to scalar, disables all vector paths. yes, all of them. scalar only, sad girl hours. 😭
  • RTE_VECT_SIMD_128 — Limits path selection to SSE/NEON/Altivec or below.
  • RTE_VECT_SIMD_256 — Limits path selection to AVX2 or below.
  • RTE_VECT_SIMD_512 — Limits path selection to AVX512 or below.
  • RTE_VECT_SIMD_MAX — Disables limiting by max SIMD bitwidth, allows all suitable paths. This value is used as it is a large number and a power of 2.

Definition at line 198 of file rte_vect.h.

⭐ Function Documentation ⭐

uint16_t rte_vect_get_max_simd_bitwidth (void )

Get the supported SIMD bitwidth.

Returns

uint16_t bitwidth.

int rte_vect_set_max_simd_bitwidth (uint16_t bitwidth)

Set the supported SIMD bitwidth.

🚨 bestie, listen: this API should only be called once at initialization, before EAL init. do not be out here calling it whenever, that's not the move.

Parameters

  • bitwidth — uint16_t bitwidth.

Returns

  • 0 on success.
  • -EINVAL on invalid bitwidth parameter.
  • -EPERM if bitwidth is forced.

👑 Author 👑

Generated automatically by Doxygen for DPDK from the source code.

and that's the whole page, babe. this gorgeous little doc came to you courtesy of DPDK, Version 25.11.3 — generated automatically by Doxygen, so like, technically the robots wrote it and I just gave it the personality transplant it deserved. 🎀✨💖 mwah, go write some vectors, you absolute icon. 💅

📦 where this page comes from