Announcement

Collapse
No announcement yet.

Core i3 10100 vs. Core i5 10600K vs. Ryzen 3 3300X Linux Gaming Benchmarks

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

  • #21
    Originally posted by Michael View Post
    Unfortunately I am down to one working WattsUp Pro (they unfortunately have a tendency to die after a while, the company is no longer in business, and all the ones I have tried buying off eBay, etc, have been dead as well). I am not aware of any similar device to WattsUp Pro in a modest price point that also works under Linux. Thus due to concurrently benchmarking many systems, it's only possible to sometimes also include WattsUp Pro metrics.
    There seem to be many current replacements for the "Watts Up Pro Portable Power Meter". Others here might mention these replacements? YouTube publisher "bigclivedotcom" often uses these instruments.

    In your tests here, we are interested in the comparisons, on the hardware & environment, but with the one variable being altered.

    When the tests, however, there are other variables to accompany the CPU: the motherboard & the CPU cooler. With the CPU, the mother board changes & the cooling fan changes. Power usage will vary greatly if the cooling fan is used very much. Different choices of cooling fans can affect power usage.

    The other cost is that generally each Intel CPU needs a new motherboard. AMD CPU's seem to be less likely to need replacement motherboards. These "minor" differences can contribute to the much greater cost & inconveniences to the administrators of the hardware.

    Comment


    • #22
      20W for X570 sounds OK, but it implies some serious I/O (like RAID PCIe 4.0 SSD or something). Personally I own Gigabyte X570 Elite paired with simple SATA SSD - I've never seen PCH temps above ~65 and my PCH fan is always in passive mode. No way that would be the case if 20W is like a "nominal power draw".
      Last edited by drakonas777; 07-27-2020, 03:01 AM.

      Comment


      • #23
        Originally posted by milkylainen View Post

        ~20W extra on the chipset over the counterpart, like continuously?
        Seems way excessive, but if you say so...
        I think it's more like 10W, but yeah PCIEv4 requires some juice.

        Anyway, the other thing going on is that the chiplet based architecture isn't particularly energy efficient compared to a single chip with everything - that's one reason that Renoir didn't go that way. Moving data around between the io die and chiplets requires constant power use.

        That's a constant amount of extra power draw, and it's not very noticeable on the higher end desktop parts. But for the low-power parts it makes up a bigger portion of the total.

        Comment


        • #24
          Originally posted by schmidtbag View Post
          The chipset is X570, that's why the power draw is so high.
          The article says "CPU Power Consumption Monitor", not "System Power Consumption Monitor".

          Comment


          • #25
            Originally posted by atomsymbol View Post

            The article says "CPU Power Consumption Monitor", not "System Power Consumption Monitor".
            That's all fine, but article also says:
            "During the Linux gaming tests, the Core i3 10100 CPU was consuming much less power than the Ryzen 3 3300X, at least according to AMD's new amd_energy open-source driver."

            That clearly shows that software was used to monitor power consumption, and software can't be trusted for those things (in general). We are nota rguing here that R3 consumes less power, maybe it is the case that i3 actually consumes less, we are giving possible explanations why that might be the case.
            That being said, we were wrong, since the power consumption was monitored via software, unless someone does tests with power line measurement, we can only guess. But that is very complicated process, first you need to calculate how much power motherboard and components consume in given workloads, since nowadays it really is workload dependent, for example, few motherboard components might be stressed and consume as much power they can, while others were partly idle, while in other workloads, it may happen that every component consumes maximum (rarely the case), so, you will always get misleading information regardless of what you do, unless, you measure it from the line, and calculate for every single possible scenario/workload, and then, plug it in in CPU power consumption calculation.
            Such methodology, while most accurate, would be insane to do, and such a time waster for very little benefits, this way (with software monitoring) or via power line, you get a "general picture", sort of approximation, while most accurate approximation would be from power line (socket).

            Comment


            • #26
              Originally posted by SilverFox View Post

              Yeah, Intel has seriously lost the plot, By the time they release 7nm Ryzen will be everyone's goto.
              I wonder if you savvy chaps could opine on the respective mobo/platform/ecosystem situations when choosing amd vs intel for the majority at ~this pricepoint.

              I sense that if you monetised all the pros & cons, amd AM4 would have a handy advantage.

              Comment


              • #27
                I wonder if you savvy chaps could opine on the respective mobo/platform/ecosystem situations when choosing amd vs intel for the majority at ~this pricepoint.

                I sense that if you monetised all the pros & cons, amd AM4 would have a handy advantage.
                Well, i've not used the ryzens as im stuck on kaby lake. I say stuck, because with intel, If you buy a newer Processor, Your buying a new board aswell. On ryzen you can keep your board (for now anyway). AMD are well documented for better value, More cores/threads, Cheaper price. Also, AMD Boards now have PCIe4. So if you want speed that speaks volume.........................

                I've been very happy with my intel i7-7700K, But when i upgrade, It's going to cost me a new board aswell which can cost the price of a new ryzen.

                Comment


                • #28
                  Originally posted by Michael View Post

                  All default mitigations are in place as shown on the system table.
                  I think you misunderstood the message heh. It was a joke the word future mitigations is key.

                  Comment

                  Working...
                  X