Announcement

Collapse
No announcement yet.

Running Ice Lake Out In The Cold - Intel Core i7-1065G7

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

  • #11
    Actually there is a clear non-conducting fluid that motherboard makers use to test thermal abilities. They can chill or heat the fluid to any ambient temperature, dip the board in fully connected and boot it up while under the fluid. Sensors inside the tank detect how many BTU's the fluid is absorbing and a FLIR sensor records the hot spots.

    One company that used to make server boards had a robot that took *every* board through the fluid test for QA and then the robot would put it in a huge vacuum chamber where it would dry the board off. It then went down the belt for either case install or boxing. It was amazing to watch.

    Comment


    • #12
      Originally posted by edwaleni View Post
      Actually there is a clear non-conducting fluid that motherboard makers use to test thermal abilities.
      Is it called air?

      Comment


      • #13
        Originally posted by syrjala View Post

        Is it called air?
        I believe it is called Novec. It was developed by 3M.

        Comment


        • #14
          Originally posted by syrjala View Post
          Is it called air?
          Talk to some lightning about the non-conducting properties of air.

          Comment


          • #15
            Originally posted by andyprough View Post
            Talk to some lightning about the non-conducting properties of air.
            Tens to hundreds of millions of Volts will tend to make even traditional nonconductors into sufficient plasma to create a circuit. That said, air can be an insulator or a conductor depending on the circumstances. E&M gave me a headache in college.

            Comment


            • #16
              Originally posted by schmidtbag View Post
              I had a feeling this wasn't thermal throttling all that bad. Obviously the cold garage made a difference, but for a laptop like this, it isn't a difference that's going to be noticeable to the average user's typical workload.
              Oh, when your fingers are freezing, I'm pretty sure any saved second is a noticeable difference :P

              Comment


              • #17
                These results suggest one of two possibilities. Either this chip generates such a low amount of heat that it does not throttle at room temperatures or it generates so much heat that even the cold of the garage is insufficient to stop it from throttling. Almost every test Michael ran shows the same average and max CPU clock but the temps are about 5 to 10 °C lower during the garage tests. This seems to suggest that the throttling threshold is set quite low on this Ice Lake chip.

                @Michael: Is there a chance that Ubuntu 19.10 has the thermald daemon running by default with some unnecessarily aggressive settings? This might explain why the CPU throttles by the same degree even when it's running considerably cooler.

                Comment


                • #18
                  Originally posted by MadCatX View Post
                  These results suggest one of two possibilities. Either this chip generates such a low amount of heat that it does not throttle at room temperatures or it generates so much heat that even the cold of the garage is insufficient to stop it from throttling. Almost every test Michael ran shows the same average and max CPU clock but the temps are about 5 to 10 °C lower during the garage tests. This seems to suggest that the throttling threshold is set quite low on this Ice Lake chip.

                  @Michael: Is there a chance that Ubuntu 19.10 has the thermald daemon running by default with some unnecessarily aggressive settings? This might explain why the CPU throttles by the same degree even when it's running considerably cooler.
                  Well the turbo frequency also depends on the TDP and the voltage, and there is a limit after all. Maybe it just hit the limit, and can't actually go any higher.

                  Comment


                  • #19
                    I wonder how well it will work in the hot Indian summer...............

                    Comment


                    • #20
                      Originally posted by MadCatX View Post
                      These results suggest one of two possibilities. Either this chip generates such a low amount of heat that it does not throttle at room temperatures or it generates so much heat that even the cold of the garage is insufficient to stop it from throttling.
                      It appears to be throttling which likely happens at around 80°C.

                      one thing that I’ve yet to address though is the heat sink / fan assembly. How effective that is at leveraging the thermal differential is yet to be seen.
                      Almost every test Michael ran shows the same average and max CPU clock but the temps are about 5 to 10 °C lower during the garage tests. This seems to suggest that the throttling threshold is set quite low on this Ice Lake chip.
                      or the throttling is more complex than just temperature.
                      @Michael: Is there a chance that Ubuntu 19.10 has the thermald daemon running by default with some unnecessarily aggressive settings? This might explain why the CPU throttles by the same degree even when it's running considerably cooler.
                      I’d have to look up Intel’s specs to see what the maximum junction temperature is but a new process could have that set lower to extend chip life. There are all sorts of possibilities including configuration mistakes, this just highlights that there could be a huge research project here to determine the causes. It would be more interesting to see this chip in a desktop chassis, with a desktop style fan on the chip, tested.

                      Comment

                      Working...
                      X