Announcement

Collapse
No announcement yet.

Ryzen 9 3900X vs. Ryzen 9 3950X vs. Core i9 9900KS In Nearly 150 Benchmarks

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

  • #21
    Interesting that for code compilation there doesn't appear to be any reason to go for the more expensive 3950X vs 3900X: the best improvement is for the Linux kernel compilation and it's just 12%. The rest is in the 5-7% area. I have two theories as to why this might be the case.

    The first is that the limited memory bandwidth starts to bite. That would also explain why the Linux kernel sees the best scalability (which is generally seems to be the case): lots of smallish C-files that don't take a lot of RAM to compile (unlike, say, large C++ files). But then things don't appear to be scaling well when we go from 3950X to the 24-core Threadripper (which has twice the memory channels). For example, for GCC compilation, going from 12 cores to 16 we see 7% speedup while going from 16 cores to 24 (plus the two extra memory channels), we see 9% speedup (based on these results: https://openbenchmarking.org/result/...AS-COMPILING50). One thing that we are not taking into account in this second comparison is the RAM speed, though.

    The second theory is that for compiling real projects, single-core performance also matters a lot because of the serial linking steps. Again, the Linux kernel is the outlier here since there is only one linking step (I believe this benchmark does not build the modules, but even if it did, those are also quite parallelizable). Compare this to GCC which goes through quite a few linking "bottlenecks". And in this regard (single core turbo), the three processors are essentially the same.

    Thoughts?

    Comment


    • #22
      I know it won't change the final results, but try running the 9900KS with the performance governor instead of powersave?

      Comment


      • #23
        Originally posted by cbdougla View Post
        I have a Ryzen 1800X system and I've been thinking of plopping a 3900X into it - without upgrading the motherboard. I don't really need the PCIe 4.0 and the things I have read say that performance is pretty good on the x370 motherboards.

        My board (ASRock x370 Taichi) does have the 3900X on the compatibility list and articles like this just encourage me.
        I own a 1700 with a x470 mobo and I'm thinking to buy a 3900x only to see qtwebenbine's compile time drop under 1 hour

        Comment


        • #24
          Originally posted by MadCatX View Post
          Intel will have Ice Lakes and Tiger Lakes coming out this year so the Zen 3 better be good if AMD wants to stay strong in the race. Anyway, the ices have finally moved and both major manufacturers are rolling out products with actual improvements as opposed to Intel's refresh of a refresh of a refresh.
          I believe it when I see it. Right now I have no reason to even consider Intel from any perspective aside from harcore gaming for edge cases - aka very special games you play in for money in leagues. And that market is... erm... tiny?

          Comment


          • #25
            AMD will maintain the cadence. They are just starting to make a dent in server market share and do not want to lose momentum.

            You will see more OEM's with Epyc offerings in 3Q-2020.

            Also OEM's are looking at shifting their workstation market to emphasize AMD in 3Q. This market is IPC focused and want to time their products with the next AMD release where they are poised to overtake Intel.

            Intel won't have a new core until 4Q-2020 at the soonest and product until 2021.

            Should be an interesting year.

            Comment


            • #26
              As far as you all know i believe, gaming profile uses less watt than application profile at the same frequency. This happens because many applications use advanced multimedia extensions inside cpu that cost energy. For example a Ryzen 3600 cost 50w for games and 80w for blender at full Hz all cores use. Now Intel knows that so boosts frequency for games at near 5Ghz, Amd on the other hand allows only 200Mhz extra boost on 3000 series. If Amd with a free motherboard patch allow for a full 4.8Ghz boost when possible, then games will explode for Amd. Ryzen 3600 @4.8Ghz gaming profile will be less than 70w, blender can remain at 4.2Ghz. Amd should do that they promised: unlocked processors at least when thermally possible. It will be bad for people starting to think that Amd today tricks people like Intel and Nvidia do, to extort more money. I mean i paid my B450 motherboard with the promise that will have all the processors unlocked, what should i do now go to court?

              Comment


              • #27
                Ehh... None of this impresses me. Until I see a 60% single thread improvement from either intel or AMD I am not going to hoo and holler. There is literally nothing one of the newer CPU's would really offer most geeks. Games, graphic design, audio production and some video rendering my 4c8t i7 still packs a mighty mighty punch.

                Very very few people are going to need that kind of threaded power. The ones that do are running servers or are doing lots of lengthy video rendering, which kudos to them if they are driving a business and being successful enough to profit money from all those cores. Otherwise
                I see a lot of this hardware impractical.

                Looking at new hardware makes me want to buy new stuff but then I look at my 2017 build and notice how the same build is still slaying quite nicely.
                Last edited by creative; 18 January 2020, 02:16 PM.

                Comment


                • #28
                  Originally posted by cbdougla View Post
                  I have a Ryzen 1800X system and I've been thinking of plopping a 3900X into it - without upgrading the motherboard. I don't really need the PCIe 4.0 and the things I have read say that performance is pretty good on the x370 motherboards.

                  My board (ASRock x370 Taichi) does have the 3900X on the compatibility list and articles like this just encourage me.
                  I did this upgrade from a 1700 at 3.8ghz. The performance uplift was about double in highly multithreaded tasks, virtualization, or avx2 methods. If you need any of these things and are performance limited, it's a worthy upgrade.

                  I also waited for the 3950x to come out before buying. The price difference is substantial enough to prefer the 3900x (as these benchmarks show).

                  Comment


                  • #29
                    Originally posted by yeeeeman View Post
                    You could've at least include a similarly priced to 3950X CPU from intel in this comparison...For sure 9900k is beaten by a 16 core cpu.
                    Also, I don't see the point in using 9900ks, since this product is a special one released for OC-ing.
                    The scaling is not amazing tbh. 27% more performance for 50% more cores with 3900x? 47% more performance for 100% more cores with 3950x? Sure, prices/performance make the difference but given AMD fights a 5 year old process and architecture here, this is not an amazing result.
                    Simply using the geometric mean for comparison is idiotic when it includes many limited thread applications. Look at blender or particle simulation results.

                    Comment


                    • #30
                      It would be nice to see the MKL benchmark on AMD hardware with the following workaround: https://www.extremetech.com/computin...eadripper-cpus

                      Comment

                      Working...
                      X