Announcement

Collapse
No announcement yet.

AMD Ryzen 5 7600X Linux Performance

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

  • #21
    Originally posted by oiaohm View Post

    This is absolutely not straight forwards.
    https://www.pcgamer.com/does-ram-spe...ing-amd-intel/

    .....real CPUs nightmare is that 5200 can be 1 to 2 percent slower than 4800 if it a miss match..
    LOL .

    Bullshit mate. Both RAM has the same latency in cycles, so 5200 have 8% access time too.

    And you link article about ZEN2 (!) and DDR4.

    ZEN4 get slower above 6000, so compare Intel with 4800 vs Ryzen with 8% faster 5200 memory it is dishonest and distorts the results particularly in games and some programs.

    Comment


    • #22
      Originally posted by [email protected] View Post
      After 2 agonizing years, I finally upgraded my system a couple months ago. Since I just do internet browsing and light gaming, judging by those results, I did right in not waiting for Zen 4. Got my R5 5600 for about 50% of the price of the R5 7600X (in my neck of the woods). Together with a cheap A520 mobo and memories, I got 80% of the gaming performance for half the price of the newer CPU combo. And my system runs cooler too, even with the box cooler.
      Yeah, nobody should feel bad about picking up an R5 5600 on the cheap.

      The 7600X being 35% faster than a 5600X sounds fantastic. But it took 26% more power on average and 58% more power peak to do so in productivity tests. Same story for gaming, 30% more power on average and 28% more peak. Takes a bit of the shine off.

      Comment


      • #23
        Given the huge costs for moving to DDR5 and AM5 boards, $199USD would have been more interesting but I feel $299USD is just too much for a 6core CPU these days...

        Comment


        • #24
          Originally posted by murlakatamenka View Post
          I remember times when AMD wasn't competitive, i5 was $200 and i7 for $300.

          Now AMD is back in the game (and there are M1/M2), but the price for AMD's i5 is $300. Competition benefits the customers, right? Right?!..

          Well, at the same time if you're not chasing the very newest things, you can get something very cost efficient like Ryzen 5600 for ~150$ from AliExpress. They cost even $135-140 there now. Basically 5600x, and what was its MSRP at launch? $300 it was...

          No wonder it's Groups top selling CPU on US Amazon rn:

          https://www.amazon.com/gp/bestsellers/pc/229189/

          ---

          ​​​​​​​New tech surely got pricier over recent years. Its performance is impressive though.
          I remember when an i7 on non-HEDT platform was 4c/8t...

          Comment


          • #25
            Originally posted by HEL88 View Post
            Bullshit mate. Both RAM has the same latency in cycles, so 5200 have 8% access time too.

            And you link article about ZEN2 (!) and DDR4.

            ZEN4 get slower above 6000, so compare Intel with 4800 vs Ryzen with 8% faster 5200 memory it is dishonest and distorts the results particularly in games and some programs.
            https://www.tomshardware.com/feature...5000-ram-guide

            This is with Zen 3. Its a repeating problem. Note what you wrote ZEN4 get slower above 6000. 5200 is on slow side for ZEN4. I have not seen good memory benchmarks for Zen4 on it scaling with ram to know how badly it hurt at 5200 ram.

            https://www.phoronix.com/review/ddr4-ddr5-alderlake
            Here is a ram benchmark with 12600k it starts showing diminishing returns at 3200-3600 ram speeds. The reality is giving 12600k 4800 vs 5200 memory you are looking at less than 1% performance difference not the 8% percent difference at best.

            The biggest gains for the 12600k is using ddr5 instead of ddr4 and that the memory controller is more suited to DDR5 than DDR4. Kind of shows why AMD just went bugger it we are releasing a DDR5 only chip.

            https://www.techspot.com/review/2354...ore-i5-12600k/
            Stock memory support includes DDR4-3200 or DDR5-4800, but we've been testing with even faster DDR5-6000 memory, which we evaluated the Core i9-12900K with. However we found that for the most part this high-speed memory offers very little extra performance when paired with an Alder Lake CPU. This lead us to conclude that most potential 12th-gen customers should ignore DDR5 and just go with DDR4.​
            I cannot find the party that did a full ram bench on the 12600k at the moment but its fairly much been found that DDR5-4800 is where the 12500K is basically maxed out and is it sweetspot for ram. You are looking at 3% performance gain from 4800 to 6000. Zen4 performance gain from 5200 to 6000 is a lot more.

            4800 is where the intel 12600k performance gains slow down. Reality is intel 12600K with 4800 ram vs a Zen4 5200 ram is most likely still in the Intel favor for showing relative MAX cpu performance.

            5200 ram with the ZEN 4 chip is not going correctly show the ZEN 4 max performance because that not the sweet spot for Zen 4.
            4800 ram is on the sweet spot for Intel 12600k. Anything past the sweet spot is diminishing returns. Practical reality is there is only a max of 4 percent past the sweet spot in performance to get in the best case with CPUs. Yes this max of 4 percent gain applies to Zen4 with ram above 6000 like it applies to 12600K above 4800.

            Every modern CPU has a ram sweetspot once you are above that you don't see the calculable performance gains from ram clock speed any more.

            HEL88 the way I was reading you were thinking the Intel chip had been disadvantaged due to the lower ram speed. When in reality its the AMD ZEN 4 chip that being major-ally disadvantaged.

            The idea of put both chips on exactly the same ram speed use to work for compares before the ZEN line. Where the fast ram possible would max everything out. ZEN line brought the horrible of non uniform scaling. The newer Intel chips also don't have uniform scaling. Non uniform scaling means you can put faster ram in a system and in reality it goes slower because you getting bound up somewhere. AMD it getting bound up with the infinity fabric. Intel there is something in there that it get bound up on that Intel has not clearly publicly named.

            Correctly bench-marking CPUs is no longer easy.

            Comment


            • #26
              Originally posted by HEL88 View Post

              In this tests in many cases 7600x is slower than 12600k, but yes 7600x wins.

              But more important is that RAM was not equal for both platforms: Intel had 4800, but AMD had 5200! This is more than 8% difference.

              In test from TechpowerUp ram was equal for both platform: 6000.
              Is TechPowerUp the most or the least reliable benchmark provider? I don't know the answer to that question. But I do see strange things.

              For example, look at this result from TPU: https://tpucdn.com/review/amd-ryzen-...-1920-1080.png

              Then compare with this result: https://static.tweaktown.com/content...cpu-review.png

              We can conclude that interpreting the results is not super easy. Hopefully you also know that some review websites are Intel fanboys, and that some websites have been bribed by Intel anyway. There may also be review websites that are more fans of AMD. Intel is probably going to have the longest hand. It's very hard to trust reviews. A lot of websites are going to have a bias towards AMD or Intel.



              Comment


              • #27
                Its tough for me to follow some of ur points, but re:
                "if you have to have "new" all the time, AMD does make for a very compelling platform vs. Intel (and vs Nvidia)."

                The longevity of AM4 mobos is a side issue.

                Intel ("the toothpaste company to the chinese (they deliver in small dollops)") also conspired to offer very little "new" with the "new model" cpuS & their mandatory new mobo.

                AMD Zen a/ brought rapid change, well worth upgrading to, & separately, b/ as a bonus, democratised better perf by making it a very affordable retrofit on the same AM4 platform.

                Comment


                • #28
                  Originally posted by Anux View Post
                  5700G are notebook chips that are to inefficient for mobile use and get repurposed for desktop.
                  There will definitely be APUs in the future but if they have enough bad SKUs to put out desktop parts remains to be seen. The last one was 6000 series and it never made it to desktops. A 7700G3D would be a no brainer. Phoenix Point is the code name if you want to go on a research.
                  "5700G are notebook chips that are to inefficient for mobile use"

                  News to me: google "Is the Ryzen 7 5700U a good processor?
                  The AMD Ryzen 7 5700U is an upper-mid-range processor for mainstream laptop computers. It is a remarkable eight-core chip, since it provides higher-end performance while maintaining low power consumption of 15 Watts.24 Oct 2021​"

                  Comment


                  • #29
                    Originally posted by HEL88 View Post
                    ZEN4 get slower above 6000
                    No? https://www.igorslab.de/en/ryzen-700...mmendations/9/

                    Originally posted by msroadkill612 View Post

                    "5700G are notebook chips that are to inefficient for mobile use"

                    News to me: google "Is the Ryzen 7 5700U a good processor?
                    The AMD Ryzen 7 5700U is an upper-mid-range processor for mainstream laptop computers. It is a remarkable eight-core chip, since it provides higher-end performance while maintaining low power consumption of 15 Watts.24 Oct 2021​"
                    What's your question/point?​

                    Comment


                    • #30
                      Originally posted by oiaohm View Post

                      https://www.tomshardware.com/feature...5000-ram-guide

                      This is with Zen 3. Its a repeating problem. Note what you wrote ZEN4 get slower above 6000. 5200 is on slow side for ZEN4. I have not seen good memory benchmarks for Zen4 on it scaling with ram to know how badly it hurt at 5200 ram.

                      https://www.phoronix.com/review/ddr4-ddr5-alderlake
                      Here is a ram benchmark with 12600k it starts showing diminishing returns at 3200-3600 ram speeds. The reality is giving 12600k 4800 vs 5200 memory you are looking at less than 1% performance difference not the 8% percent difference at best.

                      The biggest gains for the 12600k is using ddr5 instead of ddr4 and that the memory controller is more suited to DDR5 than DDR4. Kind of shows why AMD just went bugger it we are releasing a DDR5 only chip.

                      https://www.techspot.com/review/2354...ore-i5-12600k/


                      I cannot find the party that did a full ram bench on the 12600k at the moment but its fairly much been found that DDR5-4800 is where the 12500K is basically maxed out and is it sweetspot for ram. You are looking at 3% performance gain from 4800 to 6000. Zen4 performance gain from 5200 to 6000 is a lot more.

                      4800 is where the intel 12600k performance gains slow down. Reality is intel 12600K with 4800 ram vs a Zen4 5200 ram is most likely still in the Intel favor for showing relative MAX cpu performance.

                      5200 ram with the ZEN 4 chip is not going correctly show the ZEN 4 max performance because that not the sweet spot for Zen 4.
                      4800 ram is on the sweet spot for Intel 12600k. Anything past the sweet spot is diminishing returns. Practical reality is there is only a max of 4 percent past the sweet spot in performance to get in the best case with CPUs. Yes this max of 4 percent gain applies to Zen4 with ram above 6000 like it applies to 12600K above 4800.

                      Every modern CPU has a ram sweetspot once you are above that you don't see the calculable performance gains from ram clock speed any more.

                      HEL88 the way I was reading you were thinking the Intel chip had been disadvantaged due to the lower ram speed. When in reality its the AMD ZEN 4 chip that being major-ally disadvantaged.

                      The idea of put both chips on exactly the same ram speed use to work for compares before the ZEN line. Where the fast ram possible would max everything out. ZEN line brought the horrible of non uniform scaling. The newer Intel chips also don't have uniform scaling. Non uniform scaling means you can put faster ram in a system and in reality it goes slower because you getting bound up somewhere. AMD it getting bound up with the infinity fabric. Intel there is something in there that it get bound up on that Intel has not clearly publicly named.

                      Correctly bench-marking CPUs is no longer easy.
                      your claims are false. Of course in some benchmarks are not RAM speed sensitive, but games are literally very strongly RAM sensitive. There is no such thing as RAM sweetspot, as most benchmarks are either bandwidth sensitive (like copying/reading big parts of memory) or latency sensitive (cache mismatches) and BOTH improve with faster RAM speed. Idea 4800 vs 5200 is fair is complete dishonest, any CPU in games is very RAM senstitive unless it is having a huge cache but only CPU with huge cache so far is 5800X3D and that one isn't sensitive.

                      Cyberpunk shows pretty much linear scalling for Alder lake up to 6200 MHz https://youtu.be/LU_w9fZvSso?t=629

                      Last edited by piotrj3; 08 October 2022, 08:22 AM.

                      Comment

                      Working...
                      X