Announcement

Collapse
No announcement yet.

300+ Benchmarks With AMD Threadripper 3960X vs. Intel Core i9 10980XE

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

  • 300+ Benchmarks With AMD Threadripper 3960X vs. Intel Core i9 10980XE

    Phoronix: 300+ Benchmarks With AMD Threadripper 3960X vs. Intel Core i9 10980XE

    Complementing our launch-day Intel Core i9 10980XE and AMD Ryzen Threadripper 3960X/3970X Linux benchmarks, here is much more data now that we've had the additional time for carrying out more tests... For your viewing pleasure this US holiday week are more than 330 benchmarks carried out on both the Core i9 10980XE and Threadripper 3960X in the same configuration while running Ubuntu Linux.

    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
    I know what's km my wish list this Christmas

    Comment


    • #3
      This is good and all, but why compare CPUs from different price brackets? I believe intel CPU is competitor vs Highest end Ryzen processors Ryzen9 [whatever], and not TR.

      Comment


      • #4
        Originally posted by Drago View Post
        This is good and all, but why compare CPUs from different price brackets? I believe intel CPU is competitor vs Highest end Ryzen processors Ryzen9 [whatever], and not TR.
        FTA: "For getting a more diverse idea of where the Core i9 10980XE Cascade Lake X and Ryzen Threadripper 3960X trade blows, I fired up a much broader set of benchmarks for comparison on these HEDT systems. Yes, the Ryzen 9 3950X is priced more comparatively to the i9-10980XE, but I was never sent a review sample of that processor so am using the 3960X for now -- if I get my hands on said processor, I'll certainly have a similar comparison on that front."

        And as shown with the results, in some workloads the i9-10980XE is comparable to the 3960X.
        Michael Larabel
        https://www.michaellarabel.com/

        Comment


        • #5
          Originally posted by Drago View Post
          This is good and all, but why compare CPUs from different price brackets? I believe intel CPU is competitor vs Highest end Ryzen processors Ryzen9 [whatever], and not TR.
          Actually, if you add the requirement for ECC memory, than the Intel price jumps to almost the same value, i.e. $1333 for Xeon W-2295, so these benchmarks are very relevant for whoever is considering which to buy between Xeon W-2295 and 3960X.

          I have used hundreds of computers and I have seen enough horror stories to convince me to never trust my precious data with computers without ECC memory, so the fact that Intel is willing to sell the CPU variants without ECC at 75% of the full price, does not matter at all for me.

          The price of the 3960X must be compared with the corresponding Xeon W, not with the crippled amateur part.


          This does not change the fact that the 3960X is somewhat overpriced, because a single computer with 3960X is more expensive than 2 computers with 3900X, even if for many applications with high parallelism and little communication, e.g. for program compilation, the double 3900X will be as fast or faster. The same is true when comparing a single computer with 3970X versus 2 computers with 3950X.

          If the Threadrippers would have been cheaper by at least $200 each, then after adding the cost of motherboards, cases, PSUs and coolers, a single Threadripper would have always been preferable vs. two Ryzen 9. Now that is not true.










          Comment


          • #6
            Originally posted by Drago View Post
            This is good and all, but why compare CPUs from different price brackets? I believe intel CPU is competitor vs Highest end Ryzen processors Ryzen9 [whatever], and not TR.
            There's always going to be an issue no matter what. If you compare CPUs by price, Intel fans will complain because the core count will be way off. But if you compare CPUs based on core count, the price will be way off. But even if you compare the 10980XE to the 3950X (which has fewer cores, is a lower performance tier, much lower wattage, and substantially cheaper), the 3950X has the performance to be a direct competitor.

            Comment


            • #7
              The 10980XE is holding up pretty well IMO. For having 33% more cores, the 3960X gets only 20% better performance and that is because of the AVX-512 advantage.

              Comment


              • #8
                Originally posted by yeeeeman View Post
                The 10980XE is holding up pretty well IMO. For having 33% more cores, the 3960X gets only 20% better performance and that is because of the AVX-512 advantage.
                I would rather say it is some kind of Amdahls law behaviour which means that corecount and speedup is not really linear scaling
                Last edited by CochainComplex; 27 November 2019, 04:39 PM. Reason: ...scaling was missung

                Comment


                • #9
                  To get a really fair comparison between AMD & Intel CPUs, one should use the same scaling driver and governor (i.e. ACPI-cpufreq performance).
                  And in this configuration, I have a feeling that the Intel CPUs would actually fare a whole lot better since Intel's P-state "powersave" is pretty bad (at least that is my experience from running a Linux kernel optimized for low-latency).

                  Comment


                  • #10
                    Originally posted by Linuxxx View Post
                    To get a really fair comparison between AMD & Intel CPUs, one should use the same scaling driver and governor (i.e. ACPI-cpufreq performance).
                    And in this configuration, I have a feeling that the Intel CPUs would actually fare a whole lot better since Intel's P-state "powersave" is pretty bad (at least that is my experience from running a Linux kernel optimized for low-latency).
                    The defaults were used. Though for the most part in these workloads, generally I see minimal difference with P-State powersave vs. performance, usually it's only with gaming tests where sometimes is such a difference.
                    Michael Larabel
                    https://www.michaellarabel.com/

                    Comment

                    Working...
                    X