Announcement

Collapse
No announcement yet.

OpenJDK Merges Intel's x86-simd-sort For Speeding Up Data Sorting 7~15x

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • OpenJDK Merges Intel's x86-simd-sort For Speeding Up Data Sorting 7~15x

    Phoronix: OpenJDK Merges Intel's x86-simd-sort For Speeding Up Data Sorting 7~15x

    Earlier this year Intel posted x86-simd-sort as a blazing fast sorting library that makes use of AVX-512. When the popular Numpy began using it they found up to 10~17x faster sorts for 16-bit to 64-bit data types. Today Intel software engineers released x86-simd-sort 3.0 and it also comes minutes after OpenJDK merged a modified version of this speeding sorting code into that reference JDK codebase...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Considering the LTS version of Java was just released less than a month ago, this won't make it into LTS for 18 months. So yeah, great news...

    Comment


    • #3
      Only remotely related question: who is choosing Java in 2023 and for what type of software. Enterprise applications that were built on the Java stack come to mind, but what else and why Java?

      Comment


      • #4
        Originally posted by rhadlee View Post
        Only remotely related question: who is choosing Java in 2023 and for what type of software. Enterprise applications that were built on the Java stack come to mind, but what else and why Java?
        Java in 2023 is like eating ice-cream but not with the mouth.

        Comment


        • #5
          Originally posted by bug77 View Post
          Considering the LTS version of Java was just released less than a month ago, this won't make it into LTS for 18 months. So yeah, great news...
          Oh huh really? That's kinda sad. This gigantic of a win would be a great target for backporting, if they're willing to put time into it.

          Wonder if Amazon would be willing to put it into their version of the JDK...

          Comment


          • #6
            Originally posted by rhadlee View Post
            Only remotely related question: who is choosing Java in 2023 and for what type of software. Enterprise applications that were built on the Java stack come to mind, but what else and why Java?
            Enterprise things built on Java generally don't use a new-ish version. It's terrible. Although Java finally has a useful feature set many programmers still don't get to use it.

            To find Java using software look for things that have dependencies on that Java stuff on apache.org

            Comment


            • #7
              Originally posted by rhadlee View Post
              Only remotely related question: who is choosing Java in 2023 and for what type of software. Enterprise applications that were built on the Java stack come to mind, but what else and why Java?
              Java remains king of API services because the overheard of the language means little when your applications existence is to orchestrate calls between other services and applications. For example, read from some DB and write to some queue etc etc which is the vast majority of applications in the cloud world. Then you get the ecosystem benefits and frankly, ease of development, that comes with Java. This is coming from someone who is proficient in many different languages and have used the major ones professionally in the past.

              You wouldn't write an encoding server or anything that is very memory-intensive or CPU-intensive in Java, but writing microservices and similar makes sense because everyone knows Java and all the tools work well with it. The only other things that really makes sense are Go and C# but neither have the same ecosystem.

              Comment


              • #8
                Originally posted by rhadlee View Post
                Only remotely related question: who is choosing Java in 2023 and for what type of software. Enterprise applications that were built on the Java stack come to mind, but what else and why Java?
                As a Java programmer (who doesn't particularly like Java anymore), I think I can answer that: a lot of companies, because that's the talent they can find in the market. When you factor in that over half of that talent is sub-par (i.e. people get into programming because they heard it pays well, not because they have an inclination or because they like it), your choices are rather limited.
                Java is also kept afloat by its huge library collection. There very few things you won't find a Java library for. That answers the "who?" question.

                As for the "what else?" question, imho because of the above Java is being used in a lot of places where it is currently a subpar choice. For example in the cloud, where you want fast startup, low memory footprint. Java, with its huge JVM (I know about modules and jlink, nobody uses them yet) is the antithesis of a fit language there.

                On a more positive note, think of the myriad of Apache Java projects that do their job really well, even if they were better off written in something else: ActiveMQ, Cassandra, HBase, Spark/Pig...

                Comment


                • #9
                  I notice an acute lack of machine learning applications in Java, although it would arguably be more suitable for feeding them then Python due to the JIT and due to being able to use more than one CPU (core). The world is kinda upside down. The data collectors really like Python and that was that.

                  Comment


                  • #10
                    Originally posted by unwind-protect View Post
                    I notice an acute lack of machine learning applications in Java, although it would arguably be more suitable for feeding them then Python due to the JIT and due to being able to use more than one CPU (core). The world is kinda upside down. The data collectors really like Python and that was that.
                    That's because of development speed. With Python you just save a file and run it in a few seconds. No need for complicated builds, jar files and whatnot. At the end of the day, that translates to $$$ saved (fewer man-hours) on a project and shorter TTM. Even if the performance may initially suck. If your product happens to catch on, you can just rewrite the critical bits in something better.

                    Also, did you just reply to me on TPU about choosing a medium for long term storage?

                    Comment

                    Working...
                    X