Jump to content
Search In
  • More options...
Find results that contain...
Find results in...


  • Content Count

  • Joined

  • Last visited

Everything posted by cj09beira

  1. thant the heavens for that "vulnerability" though as i forgot my password multiples times and that is how i go back in
  2. that excuse was BS from the start. this is the kind of accusation where getting hard evidence will be hard as not all agreements are done on paper and even if they are only top level guys would have access to them and good luck getting those, who knows if its true or not, one thing is for sure its something right up intel's and nvidia's alley
  3. it was stuck behind a interface that was too slow, and capacities that were too small, hopefully things improve once ccix comes around
  4. is it that dumb though, the same way you are arbitrarily ignoring any value in crypto he was making the same comment about the value of entertainment. there are many diverse uses for crypto that make it very much a useful thing, it can be used to avoid the use of your own country's coin if for some reason its not advantageous to do so like you needing to make a international transfer, as many times in poor countries they run out of dollars which leads to many problems
  5. although in that case its probably related to the issue of badly coded games
  6. because specially on arm, the socs are already pretty efficient, and it always much harder to get that little extra efficiency using just design (without using new nodes), on the software side it might be a bit easier but with screens using most of the energy in a phone how long a phone lasts is very much in the hands of consumers as to where they set their screen luminosity to. i would love the return of retractable baterries though as there used be custom double sized batteries with new back cover that were just great like this
  7. part of it is how they are really well differentiated part of it is how they have been doing it for long so software had time to adapt. still i really dont see how this is of benefit to desktop pcs at all, even laptops, as today many of the reasons why big little was useful are reduced by just how fast and precisely the cpus can clock up or down and disable inactive areas of the cpu
  8. i can wait for the intel fanboys to defend this as a 16 core, when they said bulldozer had only 4 cores, it will be glorious, i bet intel wont get in nearly as much trouble as amd did though
  9. if i remember right 10nm+ is when they expect to beat 14nm++++
  10. this is not actually a 16 core chip but rather a 8 + 8 where the first 8 are normal cores and the last 8 are much weaker atom cores, and those lack hyperthreading
  11. yes its worse than am4, sure not all am4 boards, but its worse then x570 that's for sure, as there you can have full bandwidth pcie 4 on your chipset m.2 as well, and lets be honest most of the time you aren't using all chipset devices at once, intel's io here is lacking not as much as before but....
  12. sooo you will be limited to a single pcie 4.0 m.2 not a good look guess they dont have the volume for it/yields
  13. you are actually quite wrong on that one, i personally have both versions of the card and specially if you were on just 8gb of ram on the system side having 8gb of vram vs 4gb made a world of difference, and in general it means less ram usage which is a big plus
  14. i am more interested in the samsung partnership that is now confirmed to be in the next samsung soc coming 2022
  15. and its over without anymore info, same as CES last year i guess, more general info than anything concrete
  16. for now thats all that's interesting the rest is "look how much of a good guy we are" and "look how much we help these random partners"
  17. now to wait for ultrawide versions with high refresh that dont sell for a kidney and a left nut
  18. so we now have patents/whitpapers from both companies on gpu chiplets, so we should see something cool come out soon enough
  19. i think this is why we saw amd increasing the on die caches as much as they did, they seem to be trying to reduce as much as possible the need for external talk only using gddr for massive textures/meshes etc, this should help keep the bandwidth needs of the chiplets lower remember zen 2 only merged 2 ccx's that were already on the same die, so it had zero effect on yields, it was a design improvement, they started with separate caches likely to reduce development cost as much as possible, not to increase yields.
  20. it is zen 3, not zen 2, and vega really isn't a problem at these small sizes, rdna cores are much larger and other structures are also larger (shader engines etc)
  21. from what i have seem of unions outside the US they aren't much better, the same tendencies happen just as often, and endup being a burden on the company and the employees
  22. in my house we can never have enough of them, as they get destroyed easily, and most members are still on microusb which we all know it isn't reliable