Using the range operator will result in significant performance loss

[HMBSbige]

Jit

using System;

public class X
{
    public void A(ref Span<byte> span, int i)
    {
        span = span[i..];
    }

    public void B(ref Span<byte> span, int i)
    {
        span = span.Slice(i);
    }
}

; Core CLR 8.0.23.53103 on x64

X..ctor()
    L0000: ret

X.A(System.Span`1<Byte> ByRef, Int32)
    L0000: sub rsp, 0x28
    L0004: mov rax, rdx
    L0007: mov eax, [rax+8]
    L000a: mov ecx, eax
    L000c: sub ecx, r8d
    L000f: mov r8d, r8d
    L0012: mov r10d, ecx
    L0015: add r10, r8
    L0018: cmp r10, rax
    L001b: ja short L002b
    L001d: add r8, [rdx]
    L0020: mov [rdx], r8
    L0023: mov [rdx+8], ecx
    L0026: add rsp, 0x28
    L002a: ret
    L002b: mov rax, 0x7ffb10a5ead8
    L0035: call qword ptr [rax]
    L0037: int3

X.B(System.Span`1<Byte> ByRef, Int32)
    L0000: sub rsp, 0x28
    L0004: mov eax, [rdx+8]
    L0007: cmp r8d, eax
    L000a: ja short L0020
    L000c: mov ecx, r8d
    L000f: add rcx, [rdx]
    L0012: sub eax, r8d
    L0015: mov [rdx], rcx
    L0018: mov [rdx+8], eax
    L001b: add rsp, 0x28
    L001f: ret
    L0020: mov rax, 0x7ffb10a5ead8
    L002a: call qword ptr [rax]
    L002c: int3

So just ignore IDE0057