release: add a few performance-related notes #9787

1. 
   theuni commented at 8:45 PM on February 17, 2017: member

   Tried to keep these simple and human-readable.

   I'm sure I've left out a few, please chime in with any others.

   Because these all depend on specific use-cases (number of cores, peers, bandwidth, etc), I've left out any specific performance figures. If preferred, we could instead list numbers plus reference system/scenario.

2. fanquake added the label Docs and Output on Feb 18, 2017
3. fanquake added this to the milestone 0.14.0 on Feb 19, 2017
4. in doc/release-notes.md:None in 5704b9e196 outdated

     38 | +
     39 | +Validation speed and network propogation performance have been greatly
     40 | +improved, leading to much shorter sync and initial block download times.
     41 | +
     42 | +- The script signature cache has been reimplemented as a high-performance,
     43 | +  low-overhead "cuckoo cache"
   

   ---

   ---

   

   TheBlueMatt commented at 10:25 PM on February 20, 2017:

   ...", allowing for more signatures to be cached and cache lookups to be performed much faster."

5. in doc/release-notes.md:None in 5704b9e196 outdated

     43 | +  low-overhead "cuckoo cache"
     44 | +- Assumed-valid blocks have been introduced which allows script validation to
     45 | +  be skipped for ancestors of known-good blocks, without changing the security
     46 | +  model. See below for more details.
     47 | +- In some cases, compact blocks are now relayed before being fully validated as
     48 | +  per BIP152.
   

   ---

   ---

   

   TheBlueMatt commented at 10:27 PM on February 20, 2017:

   ...", often allowing relay to happen fully without waiting for block validation."

   ---

   

   theuni commented at 6:28 AM on February 23, 2017:

   I'm not sure what you're getting at here, sounds the same to me :) Please suggest a complete alternative.

   ---

   

   TheBlueMatt commented at 6:53 PM on February 23, 2017:

   Mmm, nevermind, you're probably right.

6. 
   TheBlueMatt commented at 10:28 PM on February 20, 2017: member

   This should definitely make it in for release time, if we can.

7. 
   MarcoFalke commented at 10:40 PM on February 20, 2017: member

   Could mention #8610 as well?

8. in doc/release-notes.md:None in 5704b9e196 outdated

     46 | +  model. See below for more details.
     47 | +- In some cases, compact blocks are now relayed before being fully validated as
     48 | +  per BIP152.
     49 | +- P2P networking has been refactored with a focus on concurrency and
     50 | +  throughput. Network operations are no longer bottlenecked by validation. As a
     51 | +  result, block fetching is several times faster that previous releases in many
   

   ---

   ---

   

   jnewbery commented at 12:16 AM on February 21, 2017:

   ...several times faster than previous...

9. theuni force-pushed on Feb 23, 2017
10. 
    theuni commented at 6:29 AM on February 23, 2017: member

    Updated with the suggestions above, except for the one I'm unclear about.

11. theuni renamed this:
    [WIP] release: add a few performance-related notes
    release: add a few performance-related notes
    on Feb 23, 2017
12. in doc/release-notes.md:None in 1c194e0c3b outdated

      32 | @@ -33,6 +33,28 @@ frequently tested on them.
      33 |  Notable changes
      34 |  ===============
      35 |  
      36 | +Performance Improvements
      37 | +--------------
      38 | +
      39 | +Validation speed and network propogation performance have been greatly
    

    ---

    ---

    

    sdaftuar commented at 3:00 PM on February 23, 2017:

    "propagation"

13. in doc/release-notes.md:None in 1c194e0c3b outdated

      32 | @@ -33,6 +33,28 @@ frequently tested on them.
      33 |  Notable changes
      34 |  ===============
      35 |  
      36 | +Performance Improvements
      37 | +--------------
      38 | +
      39 | +Validation speed and network propogation performance have been greatly
      40 | +improved, leading to much shorter sync and initial block download times.
    

    ---

    ---

    

    sipa commented at 6:55 PM on February 23, 2017:

    Perhaps qualify this with "in some configurations"... people who are bandwidth starved or I/O bottlenecked won't need much difference.

14. jtimon approved
15. theuni force-pushed on Feb 23, 2017
16. release: add a few performance-related notes 95e68dfe35
17. laanwj merged this on Feb 23, 2017
18. laanwj closed this on Feb 23, 2017

    ---

19. laanwj referenced this in commit f00429666c on Feb 23, 2017
20. DrahtBot locked this on Sep 8, 2021

Contributors

theuniTheBlueMattMarcoFalkejnewberysdaftuarsipajtimon

Milestone
0.14.0