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

Trimming add-across (Neoverse N1) #49

Merged
merged 3 commits into from
Jun 25, 2024
Merged

Trimming add-across (Neoverse N1) #49

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
1cce4a6
trying to reduce the cost of the 4-byte char
Jun 25, 2024
7bdac73
adding results
Jun 25, 2024
906201f
making some of the arm code prettier
lemire Jun 25, 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
33 changes: 33 additions & 0 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -178,6 +178,39 @@ boost as the Neoverse V1.
| Latin-Lipsum | 50 | 20 | 2.5 x |
| Russian-Lipsum | 4.0 | 1.2 | 3.3 x |


On a Neoverse N1 (Graviton 2), our validation function is 1.3 to over four times
faster than the standard library.

| data set | SimdUnicode speed (GB/s) | .NET speed (GB/s) | speed up |
|:----------------|:-----------|:--------------------------|:-------------------|
| Twitter.json | 12 | 8.7 | 1.4 x |
| Arabic-Lipsum | 3.4 | 2.0 | 1.7 x |
| Chinese-Lipsum | 3.4 | 2.6 | 1.3 x |
| Emoji-Lipsum | 3.4 | 0.8 | 4.3 x |
| Hebrew-Lipsum | 3.4 | 2.0 | 1.7 x |
| Hindi-Lipsum | 3.4 | 1.6 | 2.1 x |
| Japanese-Lipsum | 3.4 | 2.4  | 1.4 x |
| Korean-Lipsum | 3.4 | 1.3 | 2.6 x |
| Latin-Lipsum | 42 | 17 | 2.5 x |
| Russian-Lipsum | 3.3 | 0.95 | 3.5 x |

On a Neoverse N1 (Graviton 2), our validation function is up to three times
faster than the standard library.

| data set | SimdUnicode speed (GB/s) | .NET speed (GB/s) | speed up |
|:----------------|:-----------|:--------------------------|:-------------------|
| Twitter.json | 7.0 | 5.7 | 1.2 x |
| Arabic-Lipsum | 2.2 | 0.9 | 2.4 x |
| Chinese-Lipsum | 2.1 | 1.8 | 1.1 x |
| Emoji-Lipsum | 1.8 | 0.7 | 2.6 x |
| Hebrew-Lipsum | 2.0 | 0.9 | 2.2 x |
| Hindi-Lipsum | 2.0 | 1.0 | 2.0 x |
| Japanese-Lipsum | 2.1 | 1.7  | 1.2 x |
| Korean-Lipsum | 2.2 | 1.0 | 2.2 x |
| Latin-Lipsum | 24 | 13 | 1.8 x |
| Russian-Lipsum | 2.1 | 0.7 | 3.0 x |

One difficulty with ARM processors is that they have varied SIMD/NEON performance. For example, Neoverse N1 processors, not to be confused with the Neoverse V1 design used by AWS Graviton 3, have weak SIMD performance. Of course, one can pick and choose which approach is best and it is not necessary to apply SimdUnicode is all cases. We expect good performance on recent ARM-based Qualcomm processors.

## Building the library
Expand Down
25 changes: 13 additions & 12 deletions src/UTF8.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1352,6 +1352,8 @@ private unsafe static (int utfadjust, int scalaradjust) calculateErrorPathadjust
Vector128<byte> fourthByte = Vector128.Create((byte)(0b11110000u - 0x80));
Vector128<byte> v0f = Vector128.Create((byte)0x0F);
Vector128<byte> v80 = Vector128.Create((byte)0x80);
Vector128<byte> fourthByteMinusOne = Vector128.Create((byte)(0b11110000u - 1));
Vector128<sbyte> largestcont = Vector128.Create((sbyte)-65); // -65 => 0b10111111
// Performance note: we could process 64 bytes at a time for better speed in some cases.
int start_point = processedLength;

Expand All @@ -1362,13 +1364,13 @@ private unsafe static (int utfadjust, int scalaradjust) calculateErrorPathadjust
{

Vector128<byte> currentBlock = AdvSimd.LoadVector128(pInputBuffer + processedLength);
if (AdvSimd.Arm64.MaxAcross(Vector128.AsUInt32(AdvSimd.And(currentBlock, v80))).ToScalar() == 0)
if ((currentBlock & v80) == Vector128<byte>.Zero)
// We could also use (AdvSimd.Arm64.MaxAcross(currentBlock).ToScalar() <= 127) but it is slower on some
// hardware.
{
// We have an ASCII block, no need to process it, but
// we need to check if the previous block was incomplete.
if (AdvSimd.Arm64.MaxAcross(prevIncomplete).ToScalar() != 0)
if (prevIncomplete != Vector128<byte>.Zero)
{
int off = processedLength >= 3 ? processedLength - 3 : processedLength;
byte* invalidBytePointer = SimdUnicode.UTF8.SimpleRewindAndValidateWithErrors(16 - 3, pInputBuffer + processedLength - 3, inputLength - processedLength + 3);
Expand Down Expand Up @@ -1402,7 +1404,7 @@ private unsafe static (int utfadjust, int scalaradjust) calculateErrorPathadjust
Vector128<byte> block4 = AdvSimd.LoadVector128(pInputBuffer + processedLength + localasciirun + 48);
Vector128<byte> or = AdvSimd.Or(AdvSimd.Or(block1, block2), AdvSimd.Or(block3, block4));

if (AdvSimd.Arm64.MaxAcross(Vector128.AsUInt32(AdvSimd.And(or, v80))).ToScalar() != 0)
if ((or & v80) != Vector128<byte>.Zero)
{
break;
}
Expand Down Expand Up @@ -1433,7 +1435,7 @@ private unsafe static (int utfadjust, int scalaradjust) calculateErrorPathadjust
// AdvSimd.Arm64.MaxAcross(error) works, but it might be slower
// than AdvSimd.Arm64.MaxAcross(Vector128.AsUInt32(error)) on some
// hardware:
if (AdvSimd.Arm64.MaxAcross(Vector128.AsUInt32(error)).ToScalar() != 0)
if (error != Vector128<byte>.Zero)
{
byte* invalidBytePointer;
if (processedLength == 0)
Expand All @@ -1457,18 +1459,17 @@ private unsafe static (int utfadjust, int scalaradjust) calculateErrorPathadjust
return invalidBytePointer;
}
prevIncomplete = AdvSimd.SubtractSaturate(currentBlock, maxValue);
Vector128<sbyte> largestcont = Vector128.Create((sbyte)-65); // -65 => 0b10111111
contbytes += -AdvSimd.Arm64.AddAcross(AdvSimd.CompareLessThanOrEqual(Vector128.AsSByte(currentBlock), largestcont)).ToScalar();

// computing n4 is more expensive than we would like:
Vector128<byte> fourthByteMinusOne = Vector128.Create((byte)(0b11110000u - 1));
Vector128<byte> largerthan0f = AdvSimd.CompareGreaterThan(currentBlock, fourthByteMinusOne);
byte n4add = (byte)AdvSimd.Arm64.AddAcross(largerthan0f).ToScalar();
int negn4add = (int)(byte)-n4add;
n4 += negn4add;
if (largerthan0f != Vector128<byte>.Zero)
{
byte n4add = (byte)AdvSimd.Arm64.AddAcross(largerthan0f).ToScalar();
int negn4add = (int)(byte)-n4add;
n4 += negn4add;
}
}
}
bool hasIncompete = AdvSimd.Arm64.MaxAcross(Vector128.AsUInt32(prevIncomplete)).ToScalar() != 0;
bool hasIncompete = (prevIncomplete != Vector128<byte>.Zero);
if (processedLength < inputLength || hasIncompete)
{
byte* invalidBytePointer;
Expand Down