Announcement

Collapse
No announcement yet.

Framework Laptop 16 Delivers Great Linux Support & Performance, Excellent Customizability

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

  • #21
    Originally posted by deusexmachina View Post
    Any news on the fans? All that benchmarking - how was the difference in noise?

    Note to those with multi-NVMe plans on this thing; it only has room for 1 2280 M.2, the 2nd NVMe would have to be 2242 or 2230 sized.
    I'm pretty sure the 2nd is actually 2230 only, not 2242.

    Lots of reviews coming up out there, see https://community.frame.work/t/revie...opping/43959/4 for a fair number of reviews, and more in that thread as well.

    One BIG thing to remember for ALL of the reviews, is that these are all late stage validation systems. AFAIK none of them are even early mass production.

    As per their blog post https://frame.work/blog/framework-la...views-are-live

    This is the largest batch of review units we’ve ever shipped, and we expect more reviews to trickle in over the next week. For context on our press units, many companies ramp into production, manufacture a large quantity to fill retail channels, and cherry pick some golden units from that as review samples. For us, because we’re entirely direct to consumer, our ramp is extremely fast, and units go to customers’ hands immediately. That means to get press units out weeks ahead of time to provide a sufficiently long review period, we send out production-intent qualification units. The “intent” part of that is that these are usually functionally identical to what ships to consumers.
    So some of the reviews show drastically higher sustained temps than others (a known and fixed issue with the heatpipes & fans found during validation), and likely at least some of the fan noise as well.

    Comment


    • #22
      Originally posted by Drizzt321 View Post

      6 ports that you can reconfigure to meet whatever needs you have. Not saying that it's ideal for you, maybe you need more in total, but it's 2 USB4 and 4 USB 3.2 g2 ports if going all USB. Sure, one is for power in, but that's still 5. Or 1 is instead a HDMI or DP output, and 1 can be gigabit ethernet. What do you need that you need to plug in more than 3 USB devices? ​



      And what other general consumer laptops have ECC RAM? You generally need to go up to workstation replacement to find that, and you're looking to pay way more for a machine.
      All computers ought to have ECC RAM. In this laptop's case, several topics were opened on the manufacturer's forums, linking a couple of them here.

      Framework shouldn't continue to perpetuate this forced market segmentation and we shouldn't keep on accepting it by buying such gear. Even if it's so well made in almost all other aspects.



      The wording from AMD is pretty strange, frankly: the official pages for the APUs (7640U, 7840U) now say “ECC Support: No”, but they used to say “ECC Support: Yes (FP7r2 only; Requires platform support)” back in May and I seem to remember even in summer, though I don’t have a copy (here FP7v2 is the DDR5 SKU, the other two are LPDDR5 ones). I don’t know what happened. Either AMD screwed up the circuitry and only has enough units with working ECC for the PRO bin (I doubt those are actually diffe...

      Shutting this down folks as it’s just going down a deep, dark rabbit hole that is unnecessary. This is what happens when you live on the bleeding edge and are working on delivering products in tandem with developers/manufacturers while they are still deep into development. AMD hadn’t even announced their new line of APUs when we started working on integrating the new tech into our mainboards and they were still very much a work-in-progress. As you can tell by the conflicting information on the ...

      ECC error logging just told me that my overclock on my desktop PCs RAM is not stable any more. About 1 corrected error every hour. I configured that overclock about a year ago with extensive (days of memory load) and error free stability testing. So either my memory has degraded, my powersupply is less stable than before or a plethora of other possible reasons can be the cause. But this example shows how ECC is a useful addition even for non professionally used systems. These hourly errors woul...

      Wow they actually did! Presumably AMD decided that’s gonna be a nice artificial market segmentation for their Pro CPU line. Given AMD’s track record on the desktop I really thought they would not do the same shenanigans als Intel. https://web.archive.org/web/20230505150811/https://www.amd.com/en/product/13186#:~:text=LPDDR5x-7500-,ECC%20Support https://web.archive.org/web/20231017032134/https://www.amd.com/en/product/13186#:~:text=LPDDR5x-7500-,ECC%20Support Imagine if you relied on that sp...




      Comment


      • #23
        Originally posted by Antennae5101 View Post

        All computers ought to have ECC RAM. In this laptop's case, several topics were opened on the manufacturer's forums, linking a couple of them here.
        You can make the case for ECC RAM on a server, which runs the same software constantly and you want an absolute minimal chance of crashing or data corruption. But on a system for user interaction (especially a laptop), you're FAR more likely to encounter a crash and/or lose data from buggy software or consumer-grade hardware than experience an incorrect bit on otherwise good memory. That's the reason why there hasn't been a bigger push for ECC memory on consumer devices. It doesn't bring enough benefit to justify the added expense.

        Comment


        • #24
          Originally posted by Antennae5101 View Post
          All computers ought to have ECC RAM. In this laptop's case, several topics were opened on the manufacturer's forums, linking a couple of them here.

          Framework shouldn't continue to perpetuate this forced market segmentation and we shouldn't keep on accepting it by buying such gear. Even if it's so well made in almost all other aspects.
          While I'm not disagreeing, it'd be nice if they all came with proper ECC, there is also the cost involved in parts and validation and so on that's involved. A quite small company like Framework can only do so many things, and to hit even this (not low) price point, if they were to add full ECC it'd drive up the mainboard costs & validation by quite a bit I imagine, to say nothing of the fact that there is nearly no ECC laptop memory out there in the market that's easily available to consumers to buy direct. And what little there might be commands a premium. So you'd have to convince SODIMM manufacturers to start making some SODIMMs with full ECC available.

          They have enough of an uphill battle as it is. We can hope over time they'll be able to correct that, and maybe with CAMM2 form factor we might be able to start making headway.

          Comment


          • #25
            Originally posted by Daktyl198 View Post

            I would love a tracpoint nub, but Lenovo would never license the patent to other companies. That's a feature that will never be on a laptop that isn't a ThinkPad because it locks in people who like/are used to that feature.
            I'm not sure why you think this is patented and off limits for other OEMs. HP and Dell have both offered them for many years. And while I enjoy having them, they are dying off because most people don't use them, not because only Lenovo is allowed to offer them. I have an HP ZBook 15 G5 with a trackpoint that I need to sell.

            Comment


            • #26
              Originally posted by Mitch View Post

              I haven't thoroughly rested all the same games on my FW13 but the rdna3 on it is a champ. So far there is nothing I can't play.

              Video acceleration is a dud right now, but I think that's it. I just leave that off till they fix it
              Wait, hardware video decode is broken on RDNA3 iGPUs?

              Comment


              • #27
                Originally posted by pWe00Iri3e7Z9lHOX2Qx View Post

                Wait, hardware video decode is broken on RDNA3 iGPUs?
                It is? That sucks. Feels more like a amdgpu driver issue tho, rather than a Framework issue.

                Comment


                • #28
                  Presumably all your tests were with the power adapter that shipped with the system, right?

                  If not; make sure you benchmark with that specifically because it will allow better performance as this is the first 180W USB C power adapter in existence (afaik).

                  I think it would also be worth pointing out that any future battery life and performance tests should be done with power profiles daemon 0.14 which will tune EPP policy and values in addition to the amd-pmf (ACPI platform profile). This has pretty dramatic implications for efficiency and performance.

                  Comment


                  • #29
                    Originally posted by andyprough View Post
                    That ethernet port is a big problem. Should have gone with a collapsable ehternet port like Clevo did, then you can fit it into a slimline case.
                    It's good they didn't. Those are notoriously breaking in no time. It would just be an unnecessary point of failure.

                    Comment


                    • #30
                      Michael: The final power consumption charts seems wrong. The Framework 16 has higher avg power in every individual test but comes out way lower than FW 13 in the final chart. Is 13 and 16 accidentally swapped there?

                      PS: In 3D load the Tuxedo seems to perform consistently better while in CPU stuff it's worse. Is this some configuration that may be changed in BIOS, because cooling and max TDP seems better on the FW 16?

                      Comment

                      Working...
                      X