C＃提高SIMD Sum的性能

如何解决C＃提高SIMD Sum的性能

我正在编写SIMD库，并尝试压缩所有性能。
我已经将> df[[752]] %>% ungroup() %>% data.frame() %>% class() [1] "data.frame" > df[[752]] %>% ungroup() %>% data.frame() %>% colnames() [1] "leg.id" "arb_identifier" "SecurityID" "date" "UnderlyingClose" "UnderlyingOpen" [7] "TotalReturn" "ReferenceExchange" "OptionID" "Expiration" "CallPut" "Strike" [13] "Volume" "OpenInterest" "ImpliedVolatility" "Delta" "Gamma" "Vega" [19] "Theta" "AdjustmentFactor" "BestBid" "BestOffer" "Last" "LastTradeDate" [25] "T" "stale" "old" "roll" "n_opt_shares" "delta.hedge" [31] "OrigBid" "OrigOffer" "PXRecov" "acquisition_date" "tranche_id" > df[[752]] %>% ungroup() %>% data.frame() %>% NROW() [1] 1 投射到array中，而不是创建新对象。
目标数组很大（超过1000个元素）。
有没有更有效的数组求和方式？
欢迎提出想法。

Span<Vector<int>>

解决方法

您的代码是好的。只能提高4％，方法如下：

// Test result: only 4% win on my PC.
[MethodImpl( MethodImplOptions.AggressiveInlining )]
static int sumUnsafeAvx2( int[] array )
{
    unsafe
    {
        fixed( int* sourcePointer = array )
        {
            int* pointerEnd = sourcePointer + array.Length;
            int* pointerEndAligned = sourcePointer + ( array.Length - array.Length % 16 );
            Vector256<int> sumLow = Vector256<int>.Zero;
            Vector256<int> sumHigh = sumLow;
            int* pointer;
            for( pointer = sourcePointer; pointer < pointerEndAligned; pointer += 16 )
            {
                var a = Avx.LoadVector256( pointer );
                var b = Avx.LoadVector256( pointer + 8 );
                sumLow = Avx2.Add( sumLow,a );
                sumHigh = Avx2.Add( sumHigh,b );
            }
            sumLow = Avx2.Add( sumLow,sumHigh );
            Vector128<int> res4 = Sse2.Add( sumLow.GetLower(),sumLow.GetUpper() );
            res4 = Sse2.Add( res4,Sse2.Shuffle( res4,0x4E ) );
            res4 = Sse2.Add( res4,1 ) );
            int scalar = res4.ToScalar();
            for( ; pointer < pointerEnd; pointer++ )
                scalar += *pointer;
            return scalar;
        }
    }
}

Here's a complete test.

为清楚起见，我不建议您做上面写的事情。不适用于4％的改善。不安全的代码是不安全的。您的版本可以在没有AVX2的情况下运行，并且可以从AVX512获得好处，如果没有AVX2，我的优化版本将会崩溃，即使硬件支持，也不会使用AVX512。