Re: [PATCH 2/3] kernel-shark: Change the mechanism of the multi-threaded search

[Yordan Karadzhov <y.karadz@gmail.com>]

On 2.05.20 г. 4:03 ч., Steven Rostedt wrote:
> On Thu, 30 Apr 2020 22:56:37 -0400
> Steven Rostedt <rostedt@goodmis.org> wrote:
>
>> On Mon, 27 Apr 2020 15:18:02 -0400
>> Steven Rostedt <rostedt@goodmis.org> wrote:
>>
>>>> I would say we can apply both. What do you think?
>>> I'll try it out and let you know.
>> Want to resend this patch with this change?
>>
> BTW, I had to test this on my laptop (which doesn't have 24 logical
> CPUs), but was able to reproduce it with the following patch. Perhaps
> we should try various numbers to make sure it works for other strange
> combinations.
>
> -- Steve
>
> diff --git a/kernel-shark/src/KsTraceViewer.cpp b/kernel-shark/src/KsTraceViewer.cpp
> index 12371ad7..6e2b5fc6 100644
> --- a/kernel-shark/src/KsTraceViewer.cpp
> +++ b/kernel-shark/src/KsTraceViewer.cpp
> @@ -676,7 +676,7 @@ void KsTraceViewer::_setSearchIterator(int row)
>   
>   void KsTraceViewer::_searchItemsMT()
>   {
> -	int nThreads = std::thread::hardware_concurrency();
> +	int nThreads = 24; //std::thread::hardware_concurrency();
Yes, I did the same.
Y.


>   	int startFrom, nRows(_proxyModel.rowCount({}));
>   	std::vector<QPair<int, int>> ranges(nThreads);
>   	std::vector<std::future<QList<int>>> maps;