On Tuesday, 21 April 2020 at 20:17:37 UTC, Eduard Staniloiu wrote:
> On Monday, 20 April 2020 at 21:22:37 UTC, welkam wrote:
>> On Monday, 20 April 2020 at 17:30:11 UTC, Tim wrote:
>>> allocations on the D heap?
>> Whats that? You mean GC heap? I tried that tool on DMD and it found its custom allocator so it did all the things I needed it for. But yeah for better support you would need to modify the source and since its written in C++ I dont think I will touch it.
>
> I just want to add that valgrind works really well with D.
> You can use massif to profile the heap allocations [0].
>
> [0] - https://www.valgrind.org/docs/manual/ms-manual.html

Thanks!

On Monday, 20 April 2020 at 21:22:37 UTC, welkam wrote:
> On Monday, 20 April 2020 at 17:30:11 UTC, Tim wrote:
>> allocations on the D heap?
> Whats that? You mean GC heap? I tried that tool on DMD and it found its custom allocator so it did all the things I needed it for. But yeah for better support you would need to modify the source and since its written in C++ I dont think I will touch it.

I have modified heaptrack to also track allocations on the D GC heap: https://github.com/tim-dlang/heaptrack/tree/heaptrack-d

When using it, druntime has to be linked dynamically. It also only works, when the application is launched by heaptrack, and not, when heaptrack is attached later. The modified heaptrack will now only work with D applications.

On Monday, 20 April 2020 at 07:49:33 UTC, Jacob Carlborg wrote:
> On 2020-04-20 03:59, zoujiaqing wrote:
>> PHP use xhprof analysis performance
>> Golang use pprof analysis performance
>> Rust use pprof-rs analysis performance
>> D language also needs a tool like this.
>> 
>> [0] https://github.com/google/pprof
>> [1] https://github.com/facebook/xhprof
>> [2] https://docs.rs/pprof/0.3.16/pprof
>
> DMD has the `-profile` flag. D programs should work in profilers for C/C++ as well.

Hi Jacob :)

Dynamic code performance analysis is difficult to track.
For example, each HTTP request data is different, in low performance we can analyze a single request.

On Friday, 20 November 2020 at 09:56:10 UTC, zoujiaqing wrote:
>
> Hi Jacob :)
>
> Dynamic code performance analysis is difficult to track.
> For example, each HTTP request data is different, in low performance we can analyze a single request.

Hello,

With Intel System Studio you can get Intel Inspector and Amplifier for free.
When built in x86_64 wth the .pdb you can analyze A D program (use -b release-debug for profiling).

On Friday, 20 November 2020 at 14:29:30 UTC, Guillaume Piolat wrote:
> On Friday, 20 November 2020 at 09:56:10 UTC, zoujiaqing wrote:
>>
>> Hi Jacob :)
>>
>> Dynamic code performance analysis is difficult to track.
>> For example, each HTTP request data is different, in low performance we can analyze a single request.
>
> Hello,
>
> With Intel System Studio you can get Intel Inspector and Amplifier for free.
> When built in x86_64 wth the .pdb you can analyze A D program (use -b release-debug for profiling).

Thanks!

But, Everyone should first understand the benefits of pprof!

Compared with the normal - profile, these are two concepts.