Playground/1brc
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
593 Commits 1 Branch 0 Tags
711aa2395a5ba8433ce3fdbe77d07b4e8a23a114
Commit Graph

593 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Roy van Rijn
7ced63f460 Rewrote to always read 16 bytes, this has less instructions on perf. (#562) 
It doesn't make a lot of sense since quite some code can be written shorter, but this is what gives the best numbers.
 2024-01-23 20:37:09 +01:00
3j5a
a9a05599cd CalculateAverage_3j5a off-the-shelf Java components + ArraysSupport (#566) 
* off the shell Java components, curious about official runtime results. thnx

my laptop results are around 12 seconds, e.g:
87.66user 1.32system 0:12.11elapsed 734%CPU (0avgtext+0avgdata 13980924maxresident)k

Architecture:            x86_64
  CPU op-mode(s):        32-bit, 64-bit
  Address sizes:         39 bits physical, 48 bits virtual
  Byte Order:            Little Endian
CPU(s):                  8
  On-line CPU(s) list:   0-7
Vendor ID:               GenuineIntel
  Model name:            Intel(R) Core(TM) i5-8400H CPU @ 2.50GHz

* off-the-shelf Java components... curious about official runtime results. thnx

laptop results are around 11 seconds, e.g:
./calculate_average_3j5a.sh  81.46s user 1.36s system 758% cpu 10.917 total

Architecture:            x86_64
  CPU op-mode(s):        32-bit, 64-bit
  Address sizes:         39 bits physical, 48 bits virtual
  Byte Order:            Little Endian
CPU(s):                  8
  On-line CPU(s) list:   0-7
Vendor ID:               GenuineIntel
  Model name:            Intel(R) Core(TM) i5-8400H CPU @ 2.50GHz

* off-the-shelf Java components + ArraysSupport..

laptop results are around 10.2 seconds, e.g:
./calculate_average_3j5a.sh  75.02s user 1.31s system 750% cpu 10.175 total

Architecture:            x86_64
  CPU op-mode(s):        32-bit, 64-bit
  Address sizes:         39 bits physical, 48 bits virtual
  Byte Order:            Little Endian
CPU(s):                  8
  On-line CPU(s) list:   0-7
Vendor ID:               GenuineIntel
  Model name:            Intel(R) Core(TM) i5-8400H CPU @ 2.50GHz

* method handle...

* full buffer read attempt

* MH

* MH cleanup
 2024-01-23 20:31:45 +01:00
Mathias Bjerke
337642d1ec 1brc contribution from mattiz (first attempt) (#567) 
* Contribution from mattiz

* Formatted code
 2024-01-23 20:28:58 +01:00
karthikeyan97
292edc629f fine tuning performance further (#526) 
* final comit

changing using mappedbytebuffer

changes before using unsafe address

using unsafe

* using graalvm,correct unsafe mem implementation

---------

Co-authored-by: Karthikeyans <karthikeyan.sn@zohocorp.com>
 2024-01-23 20:21:52 +01:00
Roman Musin
6c0949969a Native image + a few smaller optimisations (#564) 
* Inline parsing name and station to avoid constantly updating the offset field (-100ms)

* Remove Worker class, inline the logic into lambda

* Accumulate results in an int matrix instead of using result row (-50ms)

* Use native image
 2024-01-23 20:19:07 +01:00
yourwass
ba793e88cd Add Yourwass take on the challenge (#532) 
* Uses vector api for city name parsing and for hash index collision resolution
* Uses lookup tables for temperature parsing
 2024-01-23 20:04:55 +01:00
Yann Moisan
b3420d9348 improvements (#521) 
- inline computeIfAbsent
- replace arraycopy by copyOfRange

Co-authored-by: Yann Moisan <yann@zen.ly>
 2024-01-23 19:58:10 +01:00
Parker Timmins
c886aaba34 Deploy v2 for parkertimmins (#524) 
* Deploy v2 for parkertimmins

Main changes:
- fix hash which masked incorrectly
- do station equality check in simd
- make station array length multiple of 32
- search for newline rather than semicolon

* Fix bug - entries were being skipped between batches

At the boundary between two batches, the first batch would stop after
crossing a limit with a padding of 200 characters applied. The next
batch should then start looking for the first full entry after the
padding. This padding logic had been removed when starting a batch. For
this reason, entries starting in the 200 character padding between
batches were skipped.
 2024-01-23 19:43:34 +01:00
Artsiom Korzun
d858959097 parse value before going to map (#548) 
parse value before going to map
 2024-01-23 19:38:32 +01:00
Gaurav Anantrao Deshmukh
46d2058bd4 First optimal solution attempt (#539) 
* First optimal attempt

* Removing debug lines

* Using default string equals method

---------

Co-authored-by: Gaurav Deshmukh <deshmgau@amazon.com>
 2024-01-23 19:32:03 +01:00
Gerd Aschemann
c232346e87 #540 Cache SDKman (#554) 
IMPORTANT: Only use SDKman provided Java - System JDK no longer installed!
 2024-01-23 19:13:13 +01:00
Jeevjyot Singh Chhabda
415b3eb5c5 b1rc challenge by @jeevjyot (#551) 
* b1rc challenge

* fixed a rounding error

* added the file back

* fixed file

* removed a file

---------

Co-authored-by: Jeevjyot Singh Chhabda <jeevjyotsinghchhabda@Jeevjyots-MBP.hsd1.ca.comcast.net>
 2024-01-23 18:43:58 +01:00
Jaromir Hamala
2c432abb96 jerrinot's improvement - fast-path for short keys (#545) 
* fast-path for keys<16 bytes

* fix off by one error

the mask is wrong for he 2nd word when len == 16

* less chunks per thread

seems like compact code wins. on my test box anyway.
 2024-01-23 18:29:22 +01:00
Marek Kohn
2be84e09ee Add 1brc solution by @makohn (#544) 2024-01-23 18:26:53 +01:00
Rene Schwietzke
6cf6da0c17 Added environment docs (#555) 
* Create ENVIRONMENT.md

* More intel added
 2024-01-23 18:21:36 +01:00
Gunnar Morling
95c0fb06b0 Leaderboard update 2024-01-23 18:16:12 +01:00
Van Phu DO
98a8279669 use thomaswue trick, use parallelism, slightly faster (#560) 2024-01-23 16:41:25 +01:00
Dr Ian Preston
8bae1b8781 Use simd for name comparison (#568) 
Co-authored-by: Ian Preston <ianopolous@protonmail.com>
 2024-01-23 16:37:33 +01:00
Antonio Muñoz
f7febea2f6 tonivade implementation (try 2) (#541) 
* tonivade implementation

* synchronized block performs better than ReentrantLock

* remove ConcurrentHashMap

* refactor

* use HashMap.newHashMap

* change double to int

* minor refactor

* fix
 2024-01-23 16:35:05 +01:00
Elliot Barlas
df891354d3 Inline and optimize value parsing code for each of the four semicolon position processing branches. This provides a small but noticeable speed-up. It also expands and obfuscates the code, unfortunately. (#563) 2024-01-23 16:31:04 +01:00
Sarkie
e16ad7660d Update README.md to add the _baseline (#552) 2024-01-22 11:00:01 +01:00
Jason Nochlin
b374643a1f evaluate.sh: Add note for "using Unsafe" (#547) 
* fix typo

* automatically label entries using Unsafe

* fix for entry in src/main/java-22/

* backfill leaderboard

---------

Co-authored-by: Jason Nochlin <hundredwatt@users.noreply.github.com>
 2024-01-22 09:20:26 +01:00
Li Lin
d2639b7ce1 Add linl33's implementation (#503) 
* Add linl33's implementation

* Update evaluate.sh

---------

Co-authored-by: Gunnar Morling <gunnar.morling@googlemail.com>
 2024-01-21 21:14:05 +01:00
Gunnar Morling
81963f963c Leaderboard update 2024-01-21 20:39:20 +01:00
Artsiom Korzun
ac4805ee45 subprocess spawner (#542) 2024-01-21 20:23:48 +01:00
Roy van Rijn
d8b071c878 Reverting ByteBuffer idea, using Thomas's trick instead. (#538) 2024-01-21 20:15:07 +01:00
Thomas Wuerthinger
d0a28599c2 Tuning and subprocess spawn for thomaswue (#533) 
* Some clean up, small-scale tuning, and reduce complexity when handling longer names.

* Do actual work in worker subprocess. Main process returns immediately
and OS clean up of the mmap continues in the subprocess.

* Update minor Graal version after CPU release.

* Turn GC back to epsilon GC (although it does not seem to make a
difference).

* Minor tuning for another +1%.
 2024-01-21 20:13:48 +01:00
Artsiom Korzun
6b95ac6113 optimize branches (#534) 2024-01-21 19:54:43 +01:00
Elliot Barlas
47269cf30b Adjust rolling hash function to operate at int-scale rather than byte-scale. Ensure 8-byte alignment in key buffer for faster comparisons. (#523) 2024-01-21 19:38:32 +01:00
Gunnar Morling
d135bd9521 Leaderboard update 2024-01-21 18:01:52 +01:00
Roman Musin
7bfc7eaec6 Reduce allocations and heap size (#525) 
* Reduce allocations

* Shrink the heap size

* Calculate hash when reading name  (50-100ms difference)

* no need to reverse bytes

* bump heap size
 2024-01-21 18:01:23 +01:00
Gunnar Morling
3e1951379a #104 Running tests for PRs 2024-01-21 17:44:35 +01:00
Gunnar Morling
38c3e0f1ee Leaderboard update 2024-01-21 13:43:40 +01:00
Van Phu DO
410425c833 reorganize code, little bit faster (#509) 2024-01-21 13:25:18 +01:00
kumarsaurav123
8ba67cbc6d Use Array to store results instead of grouping by and custom class (#522) 2024-01-21 12:50:36 +01:00
adri
0a7726cc64 Improving first iteration by avoiding string creation as much as possible (#516) 
- It avoids creating unnecessary Strings objects and handles with the station names with its djb2 hashes instead
- Initializes hashmaps with capacity and load factor
- Adds -XX:+AlwaysPreTouch
 2024-01-20 21:27:34 +01:00
Gunnar Morling
36ffed1315 Leaderboard update 2024-01-20 21:02:06 +01:00
giovannicuccu
2c1264def9 Solution without unsafe (#507) 
Co-authored-by: Giovanni Cuccu <gcuccu@imolainformatica.it>
 2024-01-20 21:01:43 +01:00
Shivam Agarwal
f06de5faab Add 0xshivamagarwal Implementation (#508) 
* 0xshivamagarwal implementation

* .

---------

Co-authored-by: Shivam Agarwal <>
 2024-01-20 20:54:04 +01:00
karthikeyan97
f49a92019e using unsafe alone (#512) 
* final comit

changing using mappedbytebuffer

changes before using unsafe address

using unsafe

* using graalvm,correct unsafe mem implementation

---------

Co-authored-by: Karthikeyans <karthikeyan.sn@zohocorp.com>
 2024-01-20 20:49:54 +01:00
Yann Moisan
ac26c8b644 Improved version based on rafaelmerino (#511) 
* files created by create_fork.sh

* use indexOf

* improved implementation based on rafaelmerino

---------

Co-authored-by: Yann Moisan <yann@zen.ly>
 2024-01-20 20:33:14 +01:00
Roman Musin
9100ed6316 Epsilon GC + a number of other small tweaks (#513) 
* Version 3

* Use SWAR algorithm from netty for finding a symbol in a string

* Faster equals - store the remainder in a long field (- 0.5s)

* optimise parsing numbers - prep

* Keep tweaking parsing logic

* Rewrote number parsing

may be a tiby bit faster it at all

* Epsilon GC
 2024-01-20 20:30:25 +01:00
Dr Ian Preston
062f2bbecf Introducing the vector api. 1s faster on 4 core i7 (#506) 
Co-authored-by: Ian Preston <ianopolous@protonmail.com>
 2024-01-20 20:09:40 +01:00
Jaromir Hamala
114ba76d20 jerrinot's improvement (#514) 
* refactoring

* segregated heap for names

also a different hashing function. turns out hashing just first word is good enough
 2024-01-20 20:06:31 +01:00
Yonatan Graber
e1a0b79259 yonatang solution: a jdk8 friendly, no unsafe code, epsilon-gc friendly solution (#499) 
* 1bc challenge, but one that will run using jdk 8 without unsafe and still do reasonably well.

* Better hashtable

* the fastest GC is no GC

* cleanups

* increased hash size

* removed Playground.java

* collision-handling allocation free hashmap

* formatting
 2024-01-20 15:02:55 +01:00
Xylitol
8353a1cb3d Processing byte array backwards (#504) 2024-01-20 14:04:19 +01:00
Gunnar Morling
51f8ecfa43 Leaderboard update 2024-01-20 14:01:46 +01:00
Elliot Barlas
eaf87689f5 Use Arena MemorySegments rather than ByteBuffers. (#505) 2024-01-20 13:56:27 +01:00
Gunnar Morling
f435d64dff Leaderboard update 2024-01-19 22:17:54 +01:00
Roy van Rijn
7c983f3d66 Added dedicated reader (#493) 
Started running perf, perhaps this helps. No idea how to use it yet
 2024-01-19 22:15:49 +01:00