R/ARMeilleure/Optimizations.cs
Wunk 17620d18db
ARMeilleure: Add initial support for AVX512 (EVEX encoding) (cont) (#4147)
* ARMeilleure: Add AVX512{F,VL,DQ,BW} detection

Add `UseAvx512Ortho` and `UseAvx512OrthoFloat` optimization flags as
short-hands for `F+VL` and `F+VL+DQ`.

* ARMeilleure: Add initial support for EVEX instruction encoding

Does not implement rounding, or exception controls.

* ARMeilleure: Add `X86Vpternlogd`

Accelerates the vector-`Not` instruction.

* ARMeilleure: Add check for `OSXSAVE` for AVX{2,512}

* ARMeilleure: Add check for `XCR0` flags

Add XCR0 register checks for AVX and AVX512F, following the guidelines
from section 14.3 and 15.2 from the Intel Architecture Software
Developer's Manual.

* ARMeilleure: Remove redundant `ReProtect` and `Dispose`, formatting

* ARMeilleure: Move XCR0 procedure to GetXcr0Eax

* ARMeilleure: Add `XCR0` to `FeatureInfo` structure

* ARMeilleure: Utilize `ReadOnlySpan` for Xcr0 assembly

Avoids an additional allocation

* ARMeilleure: Formatting fixes

* ARMeilleure: Fix EVEX encoding src2 register index

> Just like in VEX prefix, vvvv is provided in inverted form.

* ARMeilleure: Add `X86Vpternlogd` acceleration to `Vmvn_I`

Passes unit tests, verified instruction utilization

* ARMeilleure: Fix EVEX register operand designations

Operand 2 was being sourced improperly.

EVEX encoded instructions source their operands like so:
Operand 1: ModRM:reg
Operand 2: EVEX.vvvvv
Operand 3: ModRM:r/m
Operand 4: Imm

This fixes the improper register designations when emitting vpternlog.
Now "dest", "src1", "src2" arguments emit in the proper order in EVEX instructions.

* ARMeilleure: Add `X86Vpternlogd` acceleration to `Orn_V`

* ARMeilleure: PTC version bump

* ARMeilleure: Update EVEX encoding Debug.Assert to Debug.Fail

* ARMeilleure: Update EVEX encoding comment capitalization
2023-03-20 16:09:24 -03:00

68 lines
4.5 KiB
C#

using System.Runtime.Intrinsics.Arm;
namespace ARMeilleure
{
using Arm64HardwareCapabilities = ARMeilleure.CodeGen.Arm64.HardwareCapabilities;
using X86HardwareCapabilities = ARMeilleure.CodeGen.X86.HardwareCapabilities;
public static class Optimizations
{
public static bool FastFP { get; set; } = true;
public static bool AllowLcqInFunctionTable { get; set; } = true;
public static bool UseUnmanagedDispatchLoop { get; set; } = true;
public static bool UseAdvSimdIfAvailable { get; set; } = true;
public static bool UseArm64PmullIfAvailable { get; set; } = true;
public static bool UseSseIfAvailable { get; set; } = true;
public static bool UseSse2IfAvailable { get; set; } = true;
public static bool UseSse3IfAvailable { get; set; } = true;
public static bool UseSsse3IfAvailable { get; set; } = true;
public static bool UseSse41IfAvailable { get; set; } = true;
public static bool UseSse42IfAvailable { get; set; } = true;
public static bool UsePopCntIfAvailable { get; set; } = true;
public static bool UseAvxIfAvailable { get; set; } = true;
public static bool UseAvx512FIfAvailable { get; set; } = true;
public static bool UseAvx512VlIfAvailable { get; set; } = true;
public static bool UseAvx512BwIfAvailable { get; set; } = true;
public static bool UseAvx512DqIfAvailable { get; set; } = true;
public static bool UseF16cIfAvailable { get; set; } = true;
public static bool UseFmaIfAvailable { get; set; } = true;
public static bool UseAesniIfAvailable { get; set; } = true;
public static bool UsePclmulqdqIfAvailable { get; set; } = true;
public static bool UseShaIfAvailable { get; set; } = true;
public static bool UseGfniIfAvailable { get; set; } = true;
public static bool ForceLegacySse
{
get => X86HardwareCapabilities.ForceLegacySse;
set => X86HardwareCapabilities.ForceLegacySse = value;
}
internal static bool UseAdvSimd => UseAdvSimdIfAvailable && Arm64HardwareCapabilities.SupportsAdvSimd;
internal static bool UseArm64Pmull => UseArm64PmullIfAvailable && Arm64HardwareCapabilities.SupportsPmull;
internal static bool UseSse => UseSseIfAvailable && X86HardwareCapabilities.SupportsSse;
internal static bool UseSse2 => UseSse2IfAvailable && X86HardwareCapabilities.SupportsSse2;
internal static bool UseSse3 => UseSse3IfAvailable && X86HardwareCapabilities.SupportsSse3;
internal static bool UseSsse3 => UseSsse3IfAvailable && X86HardwareCapabilities.SupportsSsse3;
internal static bool UseSse41 => UseSse41IfAvailable && X86HardwareCapabilities.SupportsSse41;
internal static bool UseSse42 => UseSse42IfAvailable && X86HardwareCapabilities.SupportsSse42;
internal static bool UsePopCnt => UsePopCntIfAvailable && X86HardwareCapabilities.SupportsPopcnt;
internal static bool UseAvx => UseAvxIfAvailable && X86HardwareCapabilities.SupportsAvx && !ForceLegacySse;
internal static bool UseAvx512F => UseAvx512FIfAvailable && X86HardwareCapabilities.SupportsAvx512F && !ForceLegacySse;
internal static bool UseAvx512Vl => UseAvx512VlIfAvailable && X86HardwareCapabilities.SupportsAvx512Vl && !ForceLegacySse;
internal static bool UseAvx512Bw => UseAvx512BwIfAvailable && X86HardwareCapabilities.SupportsAvx512Bw && !ForceLegacySse;
internal static bool UseAvx512Dq => UseAvx512DqIfAvailable && X86HardwareCapabilities.SupportsAvx512Dq && !ForceLegacySse;
internal static bool UseF16c => UseF16cIfAvailable && X86HardwareCapabilities.SupportsF16c;
internal static bool UseFma => UseFmaIfAvailable && X86HardwareCapabilities.SupportsFma;
internal static bool UseAesni => UseAesniIfAvailable && X86HardwareCapabilities.SupportsAesni;
internal static bool UsePclmulqdq => UsePclmulqdqIfAvailable && X86HardwareCapabilities.SupportsPclmulqdq;
internal static bool UseSha => UseShaIfAvailable && X86HardwareCapabilities.SupportsSha;
internal static bool UseGfni => UseGfniIfAvailable && X86HardwareCapabilities.SupportsGfni;
internal static bool UseAvx512Ortho => UseAvx512F && UseAvx512Vl;
internal static bool UseAvx512OrthoFloat => UseAvx512Ortho && UseAvx512Dq;
}
}