什么是“其他”？是否发布Nsight探查器显示的停转原因？ [英] What are "Other" Issue Stall Reasons displayed by the Nsight profiler?

问题描述

我的内核在CC 3.0（Kepler）上的性能较CC 2.0（Fermi）差。在Nsight探查器中，翘曲问题效率图表显示60％的时间中没有合格的翘曲和问题失速原因图表显示其中60％是由于其他引起的。

I have a kernel that is performing poorly on CC 3.0 (Kepler) as opposed to CC 2.0 (Fermi). In the Nsight profiler, the Warp Issue Efficiency chart is showing that 60% of the time, there were no eligible warps and the Issue Stall Reasons chart is showing that 60% of these are due to "Other".

我想知道其他问题停滞的原因是什么，我可能会

I'm wondering what the Other issue stall reasons are and what I might do to reduce them.

CUDA 5.0。 / Nsight 3.0。 RC / CC 3.0。

CUDA 5.0. / Nsight 3.0. RC / CC 3.0.

推荐答案

在Nsight Visual Studio Edition 3.0 CUDA Profiler中，问题效率显示翘曲停顿原因的饼图。停顿原因是指令获取，执行依赖性，数据请求，纹理，同步和其他。

In Nsight Visual Studio Edition 3.0 CUDA Profiler the Issue Efficiency displays a pie chart of the warp stall reasons. The stall reasons are Instruction Fetch, Execution Dependency, Data Requests, Texture, Synchronization, and Other.

对于计算能力3. *设备，其他类别是时间百分比活动扭曲由于以下原因而停顿：

For Compute Capability 3.* devices the Other category is the percentage of time that active warps are stalled due to the following reasons:

• 执行单元忙（减少使用低吞吐量整数运算）


• 注册银行冲突（有时由于大量使用向量数据类型而使编译器问题更严重）


• 每个调度程序的扭曲太少

对于计算能力5. *和6. *设备，其他类别是由于以下原因而使活动翘曲停止的时间百分比：

For Compute Capability 5.* and 6.* devices the Other category is the percentage of time that active warps are stalled due to the following reasons:

• 注册银行冲突（有时会由于大量使用矢量数据类型而使编译器问题变得更糟）


• 扭曲等待解析分支


• 优先级较低且当前不考虑进行调度的扭曲

对于5. *和6. *，尤其是gp100，如果内核达到了最后一个原因可能很高（〜75％）每个扭曲调度程序有32个扭曲。

For 5.* and 6.*, especially gp100, the last reason can be very high (~75%) if the kernel reaches 32 warps per warp scheduler.

这些停顿原因归为另一类，因为很难确定开发人员可以采取哪些措施来解决这些问题。

These stalls reasons are grouped into the other category as it is hard to identify actions that a developer can taken to resolve these issues.

这篇关于什么是“其他”？是否发布Nsight探查器显示的停转原因？的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！