Comments by "Winnetou17" (@Winnetou17) on "" video.

  1. LoL @ Java because they assume the end user and the developer are both idiots - you, sir, made my day :))
    6
  2. Context FYI That's an interesting question. From what I know the GPU is really specialized on very specific instructions (usually math related), while CPU has more diverse instructions. Besides having to rewrite all the software I think it would simply not be efficient. But in theory I think it could work. Going from 8x4GHz to 1000x1.5GHz cores. But you'll then use half of the GPU for CPU-related tasks.
    1
  3. I saw some videos when the i9 throttle thing was new. Even when throttled, it performed better than the i7 counterpart. So at least they had that. But I think the i7 was throttled as well.
    1
  4. I can't wait for the book to describe the 10nm crysis. I really hope that in about 5 years an Intel employee will releasee this.
    1
  5. To add up to what those above said, if I'm not mistaken, the efficiency of the second thread per core matters a lot on how the software is implemented and if the application happens to wait on stuff. If a program is poorly optimized or it frequently has to wait for disk or whatever, that 1 core 2 thread can actually get to 200%, same as 2 core 2 thread.
    1
  6. But isn't decompressing/compressing software (like 7z) both multithreaded and heavily optimised ? I remember that it's usually used as a CPU benchmark, and having faster or more cores always helps. And the difference is quite visible (almost linear in some cases)
    1
  7. #AskGN What actually happens when something like the CPU doesn't get enough voltage ? (Besides the crash/restart) Does having a better efficiency PSU help with safe undervolting ?
    1