如何缩短VB.NET中5000条数据的计算时间（目标≤1秒）

核心优化方向：减少重复遍历、降低内存开销、优化并行效率 #

• 预分组替代重复查询
  原代码每次并行循环都调用.stresses.FindAll遍历全集合，5000个标签会产生5000次全量遍历，时间复杂度为O(n*m)。先一次性按Tag0D分组，直接获取每组数据：

  With gOutputStresses_LC(0)
      ' 替换原UniqueTagsList，直接分组
      Dim groupedStresses = .stresses.GroupBy(Function(s) s.Tag0D).ToList()
      
      Dim options As New ParallelOptions()
      options.MaxDegreeOfParallelism = Environment.ProcessorCount ' 利用全部CPU核心
      
      Parallel.ForEach(groupedStresses, options, Sub(group)
           Dim tstresses = group.ToList()
           ' 后续操作直接使用tstresses，无需再执行FindAll
  

• 简化排序逻辑
  原代码先升序排序再反转，直接改用降序排序一步完成，减少一次遍历操作：

  ' 替换tstresses.Sort(...)和Reverse()
  tstresses = tstresses.OrderByDescending(Function(x) x.Sxx).ToList()
  

• 消除冗余的回写查找
  clsMSHStress是引用类型，修改tstresses中的元素会直接同步到原stresses集合。删除Else分支末尾的For p循环，彻底避免每次查找元素的全量遍历：

  ' 删除以下冗余代码块
  'For p As Integer = 0 To tstresses.Count - 1
  '    .stresses(.stresses.FindIndex(Function(s) s.Tag0D = tstresses(p).Tag0D And s.Tag2D = tstresses(p).Tag2D)) = tstresses(p)
  'Next
  

• 避免不必要的列表创建
  原代码创建delta列表查找最大差值索引，可直接遍历计算，省去列表内存分配和额外遍历：

  ' 替换原delta列表相关代码
  Dim maxDelta As Double = Double.MinValue
  Dim splitIdx As Integer = 0
  For k As Integer = 1 To tstresses.Count - 1
      Dim currentDelta = tstresses(k).Sxx - tstresses(k - 1).Sxx
      If currentDelta > maxDelta Then
          maxDelta = currentDelta
          splitIdx = k
      End If
  Next
  

• 合并遍历计算，减少重复遍历
  计算平均值时避免单独调用Sum，在遍历过程中直接累加总和，减少一次遍历：

  ' 替换error_ <10时的平均值计算
  Dim sum As Double = 0
  For Each s In tstresses
      sum += s.Sxx
  Next
  AverageVal = sum / tstresses.Count
  

• 优化变量作用域
  将AverageVal、error_等变量移到并行循环内部，避免线程安全问题，同时减少重复初始化开销：

  Parallel.ForEach(groupedStresses, options, Sub(group)
       Dim tstresses = group.ToList()
       Dim AverageVal As Double = 0
       Dim error_ As Double = 0
       ' 后续逻辑使用内部变量
  

修改后的完整示例代码 #

With gOutputStresses_LC(0)
    Dim groupedStresses = .stresses.GroupBy(Function(s) s.Tag0D).ToList()
    Dim options As New ParallelOptions()
    options.MaxDegreeOfParallelism = Environment.ProcessorCount

    Parallel.ForEach(groupedStresses, options, Sub(group)
         Dim tstresses = group.OrderByDescending(Function(x) x.Sxx).ToList()
         Dim AverageVal As Double = 0
         Dim error_ As Double = 0

         error_ = Math.Abs((tstresses(0).Sxx - tstresses.Last.Sxx) / (tstresses(0).Sxx + tstresses.Last.Sxx))

         If error_ < 10 Then
             Dim sum As Double = 0
             For Each s In tstresses
                 sum += s.Sxx
             Next
             AverageVal = sum / tstresses.Count
             For Each c In tstresses
                 c.Sxx = AverageVal
             Next
         Else
             Dim maxDelta As Double = Double.MinValue
             Dim splitIdx As Integer = 0
             For k As Integer = 1 To tstresses.Count - 1
                 Dim currentDelta = tstresses(k).Sxx - tstresses(k - 1).Sxx
                 If currentDelta > maxDelta Then
                     maxDelta = currentDelta
                     splitIdx = k
                 End If
             Next

             ' 计算前半部分平均值
             AverageVal = 0
             For g As Integer = 0 To splitIdx - 1
                 AverageVal += tstresses(g).Sxx
             Next
             AverageVal = AverageVal / splitIdx
             For g As Integer = 0 To splitIdx - 1
                 tstresses(g).Sxx = AverageVal
             Next

             ' 计算后半部分平均值
             AverageVal = 0
             For g As Integer = splitIdx To tstresses.Count - 1
                 AverageVal += tstresses(g).Sxx
             Next
             AverageVal = AverageVal / (tstresses.Count - splitIdx)
             For g As Integer = splitIdx To tstresses.Count - 1
                 tstresses(g).Sxx = AverageVal
             Next
         End If
      End Sub)
End With

内容的提问来源于stack exchange，提问作者Anas Ahmed