Announcement

Collapse
No announcement yet.

Intel Core i7 1165G7 "Tiger Lake" Linux Performance With The Dell XPS 13 9310

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

  • #31
    Originally posted by gregzeng View Post
    Sometimes I might insist a staff member prepare an executive summary of these benchtests. Readers comments seem not so useful this time. The AMD CPU is famed for not allowing external GPU. If screen brightness (nits), 4k or 8k display is wanted, then Intel CPU is needed.
    On lightweight notebooks or laptops, power used by the CPU & GPU is important. Power forces fan cooling, battery depletion and other discomforts. With so many confusing proprietary names, it can be assumed perhaps that AMD chips are better at power efficiency, but by how much?
    On the price differences per unit, these are not significant if other performance features are so very different. Is this the case here?
    What are you rambling about? How is screen brightness depending on the CPU? Why do you claim that you can't have an external GPU with AMD chips, when there are several products on the market offering exactly that? Why would a discrete GPU be relevant in an "ultrabook" level laptop? How are the "confusing" proprietary names even related to power consumption?

    Comment


    • #32
      The workload selected for most of these benchmarks is highly MultiThread friendly.
      Unsurprisingly, in these cases, more cores == more parallelism == more performance.

      The problem is the implied generalization : highly parallel workload work better with more cores, yes.
      But how many such workload does users actively use ?
      Is that representative of their daily activity ?

      Well, of course, it depends. What is the user ? What are his daily routine and his tools ?

      This is varied of course.
      But frankly, for the vast majority of users, I would be surprised if their primary workload was driven by highly parallelized MT software.

      Comment


      • #33
        Originally posted by evergreen View Post
        The workload selected for most of these benchmarks is highly MultiThread friendly.
        Unsurprisingly, in these cases, more cores == more parallelism == more performance.

        The problem is the implied generalization : highly parallel workload work better with more cores, yes.
        But how many such workload does users actively use ?
        Is that representative of their daily activity ?

        Well, of course, it depends. What is the user ? What are his daily routine and his tools ?

        This is varied of course.
        But frankly, for the vast majority of users, I would be surprised if their primary workload was driven by highly parallelized MT software.
        Michael did include a number of single-threaded benchmarks and called them out specifically - apparently Tiger Lake was slower than Ice Lake in many of them and Renoir ended up faster than Tiger Lake even in the single-threaded tests. The only exception IIRC was a couple of the Selenium benchmarks (2 or 3 out of 5).
        Test signature

        Comment


        • #34
          So, seems like something is wrong with Tigerlake performance, atleast on Linux. Dunno if it's heat, power, or scheduler woes.

          Zen 2 is still the better choice for CPU performance.

          Also, AMD still doesn't have basic things like power monitoring in place yet. Sure it's coming in a few months, but why isn't it already there? Crap like this is why I'd still rather buy Intel, because their stuff is supported and just works.

          Comment


          • #35
            Originally posted by evergreen View Post
            The workload selected for most of these benchmarks is highly MultiThread friendly.
            Unsurprisingly, in these cases, more cores == more parallelism == more performance.

            The problem is the implied generalization : highly parallel workload work better with more cores, yes.
            But how many such workload does users actively use ?
            Is that representative of their daily activity ?

            Well, of course, it depends. What is the user ? What are his daily routine and his tools ?

            This is varied of course.
            But frankly, for the vast majority of users, I would be surprised if their primary workload was driven by highly parallelized MT software.
            True, but it does help a lot with responsiveness, with multiple apps each potentially using multiple threads. And you can't deny how useful multiple cores/threads are for software development.

            ​​​​​

            Comment


            • #36
              Originally posted by HEL88 View Post

              Dell XPS is the way better laptop than these Lenovo. 250/300 vs 500 nits ?? Really? etc,
              Lately, and I say this as someone with a bit of fanboyism towards Dell, Dell is seriously rounding corners.
              In any case, usually the business oriented laptops are much better than the "premium" laptop (though Dell's budget lineups are usually similar between consumer and business laptops).

              I wonder how the Intel/AMD variants of the Lenovo ThinkPad L15 and L14 laptops compete against each other, since I am currently saving up for the AMD-equipped (Ryzen 5 Pro) variant of the L15 once it will be available in my country (it should be a few more months), and these are amongst the best (iGPU-only, unfortunately) mainstream laptops on the market (the T-series is slightly higher quality, however it's value (not MSRP) has dropped, due to soldered RAM even in the non-S models (one DIMM only there).

              According to Reddit users, the ThinkPad L14 is possibly the only current 14-inch laptop with nonsoldered RAM.

              Comment

              Working...
              X