The (mini) review for pdqsort will start tomorrow 6/21, and run through 6/30. I will manage the review; Orson Peters is the author. The proposed boost source is here: https://gist.github.com/orlp/24f27aada1ed724d1809b372c33eb92e Documentation is here: https://github.com/orlp/pdqsort#pdqsort And a draft paper is here: https://drive.google.com/open?id=0B1-vl-dPgKm_T0Fxeno1a0lGT0E If you're interested, please answer these questions: 1. Are there any specific concerns you have about pdqsort vs. std::sort? 2. Does it compile for you when you try to use it as a replacement for std::sort? 3. How does the performance compare on your system vs. std::sort? 4. Do you think we should include pdqsort in Boost.Sort? 5. Do you have any concerns about switching the default/fallback sort in Boost libraries (including Boost.Sort) from std::sort to pdqsort?
I'm trying to read your post Alexander, but it shows up as an empty email in gmail. Did I do something wrong? On Tue, Jun 27, 2017 at 10:20 PM, Alexander Zaitsev via Boost < boost@lists.boost.org> wrote:
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost
I don't believe it's just me that can't view your post Alexander, as the post is empty on the Boost archives as well: https://lists.boost.org/Archives/boost/2017/06/236626.php On Wed, Jun 28, 2017 at 1:20 AM, Alexander Zaitsev via Boost < boost@lists.boost.org> wrote:
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost
On 6/27/2017 6:37 PM, Orson Peters via Boost wrote:
I don't believe it's just me that can't view your post Alexander, as the post is empty on the Boost archives as well: https://lists.boost.org/Archives/boost/2017/06/236626.php
It is not only you. All I have is empty content on this thread from Alexander.
On Wed, Jun 28, 2017 at 1:20 AM, Alexander Zaitsev via Boost < boost@lists.boost.org> wrote:
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Hello. Try to send my thoughts from another email. 1. Are there any specific concerns you have about pdqsort vs. std::sort? 2. Does it compile for you when you try to use it as a replacement for std::sort? 3. How does the performance compare on your system vs. std::sort? 4. Do you think we should include pdqsort in Boost.Sort? 5. Do you have any concerns about switching the default/fallback sort in Boost libraries (including Boost.Sort) from std::sort to pdqsort? 1. Only that pdqsort is really better than std::sort's Introsort 2. Yes, it does. 3. Pdqsort is much faster than std::sort. 4. Yes, i do. I think that pdqsort is really good sort algorithm - it's really fast and pdqsort is already implemented sort algorithm in Rust's libcore as sort_unstable. 5. No, i have not. Best regards, Alexander Zaitsev zamazan4ik@gmail.com 2017-06-28 16:42 GMT+03:00 Robert via Boost <boost@lists.boost.org>:
On 6/27/2017 6:37 PM, Orson Peters via Boost wrote:
I don't believe it's just me that can't view your post Alexander, as the post is empty on the Boost archives as well: https://lists.boost.org/Archives/boost/2017/06/236626.php
It is not only you. All I have is empty content on this thread from Alexander.
On Wed, Jun 28, 2017 at 1:20 AM, Alexander Zaitsev via Boost <
boost@lists.boost.org> wrote:
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman /listinfo.cgi/boost
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman /listinfo.cgi/boost
Hello Alexander, While I do appreciate the feedback very much, I think three times is more than enough :) Greetings, Orson Peters On Thu, Jun 29, 2017 at 1:38 AM, Александр Зайцев via Boost < boost@lists.boost.org> wrote:
Hello. Try to send my thoughts from another email.
1. Are there any specific concerns you have about pdqsort vs. std::sort? 2. Does it compile for you when you try to use it as a replacement for std::sort? 3. How does the performance compare on your system vs. std::sort? 4. Do you think we should include pdqsort in Boost.Sort? 5. Do you have any concerns about switching the default/fallback sort in Boost libraries (including Boost.Sort) from std::sort to pdqsort?
1. Only that pdqsort is really better than std::sort's Introsort 2. Yes, it does. 3. Pdqsort is much faster than std::sort. 4. Yes, i do. I think that pdqsort is really good sort algorithm - it's really fast and pdqsort is already implemented sort algorithm in Rust's libcore as sort_unstable. 5. No, i have not.
Best regards, Alexander Zaitsev zamazan4ik@gmail.com
2017-06-28 16:42 GMT+03:00 Robert via Boost <boost@lists.boost.org>:
On 6/27/2017 6:37 PM, Orson Peters via Boost wrote:
I don't believe it's just me that can't view your post Alexander, as the post is empty on the Boost archives as well: https://lists.boost.org/Archives/boost/2017/06/236626.php
It is not only you. All I have is empty content on this thread from Alexander.
On Wed, Jun 28, 2017 at 1:20 AM, Alexander Zaitsev via Boost <
boost@lists.boost.org> wrote:
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman /listinfo.cgi/boost
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman /listinfo.cgi/boost
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost
I had checked in deep pdqsort, and I think is a good algorithm, specially fast with small elements, which outperform the std::sort and the previous introsort of the library. With near sorted data ( sorted data and unsorted elements append at end, or inserted in the middle) the results are acceptable and better than std::sort the GCC, CLANG an VC++ compilers. I think must be included in the Boost Sort Library Francisco 2017-06-29 4:11 GMT+02:00 Orson Peters via Boost <boost@lists.boost.org>:
Hello Alexander,
While I do appreciate the feedback very much, I think three times is more than enough :)
Greetings,
Orson Peters
On Thu, Jun 29, 2017 at 1:38 AM, Александр Зайцев via Boost < boost@lists.boost.org> wrote:
Hello. Try to send my thoughts from another email.
1. Are there any specific concerns you have about pdqsort vs. std::sort? 2. Does it compile for you when you try to use it as a replacement for std::sort? 3. How does the performance compare on your system vs. std::sort? 4. Do you think we should include pdqsort in Boost.Sort? 5. Do you have any concerns about switching the default/fallback sort in Boost libraries (including Boost.Sort) from std::sort to pdqsort?
1. Only that pdqsort is really better than std::sort's Introsort 2. Yes, it does. 3. Pdqsort is much faster than std::sort. 4. Yes, i do. I think that pdqsort is really good sort algorithm - it's really fast and pdqsort is already implemented sort algorithm in Rust's libcore as sort_unstable. 5. No, i have not.
Best regards, Alexander Zaitsev zamazan4ik@gmail.com
2017-06-28 16:42 GMT+03:00 Robert via Boost <boost@lists.boost.org>:
On 6/27/2017 6:37 PM, Orson Peters via Boost wrote:
I don't believe it's just me that can't view your post Alexander, as the post is empty on the Boost archives as well: https://lists.boost.org/Archives/boost/2017/06/236626.php
It is not only you. All I have is empty content on this thread from Alexander.
On Wed, Jun 28, 2017 at 1:20 AM, Alexander Zaitsev via Boost <
boost@lists.boost.org> wrote:
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman /listinfo.cgi/boost
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman /listinfo.cgi/boost
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost
The review is concluded, and we will add pdqsort to Boost.Sort, and will switch the fallback sort for Spreadsort to pdqsort. I encourage library authors for any libraries in Boost where sorting is a non-trivial amount of the computation to consider doing the same. On Tue, Jul 4, 2017 at 2:44 PM Francisco José Tapia via Boost < boost@lists.boost.org> wrote:
I had checked in deep pdqsort, and I think is a good algorithm, specially fast with small elements, which outperform the std::sort and the previous introsort of the library.
With near sorted data ( sorted data and unsorted elements append at end, or inserted in the middle) the results are acceptable and better than std::sort the GCC, CLANG an VC++ compilers.
I think must be included in the Boost Sort Library
Francisco
2017-06-29 4:11 GMT+02:00 Orson Peters via Boost <boost@lists.boost.org>:
Hello Alexander,
While I do appreciate the feedback very much, I think three times is more than enough :)
Greetings,
Orson Peters
On Thu, Jun 29, 2017 at 1:38 AM, Александр Зайцев via Boost < boost@lists.boost.org> wrote:
Hello. Try to send my thoughts from another email.
1. Are there any specific concerns you have about pdqsort vs. std::sort? 2. Does it compile for you when you try to use it as a replacement for std::sort? 3. How does the performance compare on your system vs. std::sort? 4. Do you think we should include pdqsort in Boost.Sort? 5. Do you have any concerns about switching the default/fallback sort in Boost libraries (including Boost.Sort) from std::sort to pdqsort?
1. Only that pdqsort is really better than std::sort's Introsort 2. Yes, it does. 3. Pdqsort is much faster than std::sort. 4. Yes, i do. I think that pdqsort is really good sort algorithm - it's really fast and pdqsort is already implemented sort algorithm in Rust's libcore as sort_unstable. 5. No, i have not.
Best regards, Alexander Zaitsev zamazan4ik@gmail.com
2017-06-28 16:42 GMT+03:00 Robert via Boost <boost@lists.boost.org>:
On 6/27/2017 6:37 PM, Orson Peters via Boost wrote:
I don't believe it's just me that can't view your post Alexander, as the post is empty on the Boost archives as well: https://lists.boost.org/Archives/boost/2017/06/236626.php
It is not only you. All I have is empty content on this thread from Alexander.
On Wed, Jun 28, 2017 at 1:20 AM, Alexander Zaitsev via Boost <
boost@lists.boost.org> wrote:
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman /listinfo.cgi/boost
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman /listinfo.cgi/boost
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
participants (7)
-
Alexander Zaitsev -
Alexander Zaitsev -
Francisco José Tapia -
Orson Peters -
Robert -
Steven Ross -
Александр Зайцев