Announcement

Collapse
No announcement yet.

AMD Ryzen 9 9950X & Ryzen 9 9900X Deliver Excellent Linux Performance

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

  • AMD Ryzen 9 9950X & Ryzen 9 9900X Deliver Excellent Linux Performance

    Phoronix: AMD Ryzen 9 9950X & Ryzen 9 9900X Deliver Excellent Linux Performance

    Last Wednesday was the review embargo for the Ryzen 5 9600X and Ryzen 7 9700X Zen 5 desktop processors that proved to be very exciting for Linux workloads from developers to creators to AVX-512 embracing AI and HPC workloads. Today the review embargo lifts on the Ryzen 9 9900X and Ryzen 9 9950X and as expected given the prior 6-core/8-core tests: these new chips are wild! The Ryzen 9 9900X and Ryzen 9 9950X are fabulous processors for those engaging in heavy real-world Linux workloads with excellent performance uplift and stunning power efficiency.

    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
    When taking the geometric mean of all the raw performance results, the Ryzen 9 9950X came out to being 19.6% faster than the Ryzen 9 7950X.
    126.34 / 107.25 = 1.178, i.e. 17.8%.

    Comment


    • #3
      Looks impressive. How massive will the difference between these benchmarks and the more Windows/gaming focused crowd be this week?

      Comment


      • #4
        Congratulations to AMD. This is the kind of innovation we are looking for.

        Also, thank you AMD for treating Linux as first-class in the recent years.
        Last edited by bezirg; 14 August 2024, 09:28 AM.

        Comment


        • #5
          It looks like Zen 5 has been optimized almost entirely for enterprise workloads.

          Desktop performance improvements are mediocre and unimpressive, if not downright poor.​

          In gaming 9950X (RTX 4090 at 1080p) is on average just 1% faster than 7950X.

          Enjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.


          And as if it's not enough AMD has done this:

          Ryzen 9000 power changes reportedly coming with AGESA 1.2.0.1.a PatchA Download more power.  There’s new information suggesting that the Zen5 desktop parts may soon have higher power limits by default. The upcoming AGESA 1.2.0.1A Patch A is said to increase the default power limits (TDP) from 65W to 105W. This change is expected to apply […]


          AMD needs to release Zen 5+ ASAP.

          Comment


          • #6
            How do I read the hardware table? I can't tell where one memory stick ends and the next begins for example.
            Last edited by ernstp; 15 August 2024, 08:01 AM.

            Comment


            • #7
              Anandtech summarizes a botched release:

              Based on what we know about the Zen 5 architecture, as well as the Granite Ridge chip overall, the reasonable guess here is that we’re seeing AMD’s uncore – the memory controllers and attached Infinity Fabric – stretched to its limit. Since AMD reused the original Ryzen 7000 IOD for Ryzen 9000, the only additional resources available to feed the CPU cores is the slightly higher bandwidth of DDR5-5600 memory. All other cache and interconnect bottlenecks remain.

              Consequently, when floating point performance in a single thread improves by a hefty 24%, that’s 24% more traffic through the caches and uncore/IOD to keep those CPU cores fed. And those resources aren’t quite there. To be sure, 9950X doesn’t hit a scaling wall here, as multi-threaded floating point performance is still ahead of the 7950X by 19% overall. But it’s just a bit worse – and enough so that we can measure it. I can only surmise that Zen 5 would have gladly taken more memory bandwidth and IF bandwidth if those were available. Thankfully for AMD, Rate results don’t rely on inter-thread communication, so there aren’t any hazards from threads from different CCDs talking to each other over the IF links.​[/quote

              Comment


              • #8
                Originally posted by avis View Post
                Anandtech summarizes a botched release:
                Anandtech tests the platforms at JEDEC 5600 CL46. A 6000CL30 kit easily balances the core improvements to the IF/Memory speeds

                Comment


                • #9
                  Originally posted by avis View Post
                  Anandtech summarizes a botched release:



                  https://www.anandtech.com/show/21524...9900x-review/5
                  How can you call it botched? They reused the IO die from the last generation. It was a tock release not a tick release. Intel.spent nearly 15 years releasing every other product generation in this same fashion.

                  EDIT: The thing that happened why you consider it botched is that they reused the last gens IO die... So does that mean you also think the last gen is botched too??? And do you consider every Intel release for a nearly 15 year period or so also botched???
                  Last edited by duby229; 14 August 2024, 09:56 AM.

                  Comment


                  • #10
                    Owner of 5950x who is dying to upgrade to 9950x, the power efficiency looks great.

                    I'm waiting for the X3D, hopefully as it'll really help database performance for a workstation.

                    Comment

                    Working...
                    X