Announcement

Collapse
No announcement yet.

Intel Core i5 13600K + Core i9 13900K "Raptor Lake" Linux Preview

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

  • #41
    Originally posted by atomsymbol

    Just a note: The boiling point of [pure] water depends on [atmospheric] pressure. If the pressure inside a liquid cooler is higher than 1013.25 hPa (= atmospheric pressure at the sea level)​ then the liquid (if it happens to be pure water, which is unlikely) will not be boiling. Adding chemical compounds to pure water changes the boiling point of the liquid. Even if there was pure water in a liquid CPU cooler and it partially turned into steam: the steam would increase the pressure inside the cooler and thus the liquid would stop boiling - question is where the steam would be formed (if it formed near the cooler's pump it would decrease flow of the liquid in the cooler; movement of bubbles inside the cooler would produce extra noise).

    https://en.wikipedia.org/wiki/High-altitude_cooking
    Funfact: cooling devices that work by boiling pure water at sub-atmospheric pressure are very widely used. They're called heatpipes! 🙃

    Comment


    • #42
      qarium The efficacy of race-to-sleep is greatly exaggerated. Basically every CPU runs well above the range of frequencies that are so slow that running faster and then sleeping would save energy. Try running this on a mostly-idle system:

      Code:
      #!/usr/bin/env bash
      
      report_power () {
          turbostat \
              --quiet --Summary \
              --show Core,Busy%,Bzy_MHz,PkgWatt,CorWatt,GFXWatt,RAMWatt \
              sleep 10
      }
      
      get_max_freq () {
          cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq
      }
      
      set_frequency () {
          cpupower frequency-set -g performance -u "$1" >/dev/null
      }
      
      main () {
          for freq in 1GHz 2GHz 3GHz $(get_max_freq); do
              set_frequency $freq
              report_power
          done
      }
      
      if [ $UID -eq 0 ]; then
          main
      else
          sudo bash "$0"
      fi

      Comment


      • #43
        Originally posted by qarium View Post
        …but one thing is clear: FUCK YOU INTEL... no one should buy any intel cpu anyway.

        Show us on this doll where Intel hurt you.

        Comment


        • #44
          Originally posted by drakonas777 View Post
          So i5 now is 320USD. I see no complains from birdie. I guess it's ok when intel raises the price LOL
          Intel is offering 14 cores unlocked K and KF CPUs around $300 price point where AMD only offers 6 cores and you are complaining about Intel’s prices? How about lowering the 7700x that gets beaten by an i5?

          Comment


          • #45
            Perspective from someone who works in semiconductors, both fab, and back end high speed testing, for many many years:

            nm process is pretty much meaningless. Why, there is no standard on how to measure and where do you measure, oh in this one place we have a 5nm gap, therefore we have a 5nm process. Don't think manufacturers aren't pulling that, I know of companies doing that. Sounds good on the marketing sheets.

            TDP pretty much meaningless because they have an equation with fudge factors in there. Want to know the real TDP, look at actual measured power consumption, almost all of the power that goes into a CPU gets turned into heat.

            Benchmarks, manufactures tune to certain ones to post the big numbers. I skip over those. I go to the real world testes that apply to me. I look at the types of games I play and applications I use. I think look to see if it is a meaningful result. For example in games, I run at 3440x1440 @ 120hz, so I am GPU bound in almost all games and anything over what my monitor can display is not going to help me. The blender renders are taking 20% less time, but I don't use it daily so waiting another minute is better for me than spending the money.

            When comparing upgrades you want to look at what you have, where you are looking to upgrade to and the competitors at that level. Then go through and check:
            1. Power draw, actual reviewer numbers from several sources.
            2. performance in the real world applications
            3. Take a hard look to see if you would really notice it, figuring at least a 10% bump is where you can just perceive an increase. I personally don't upgrade unless I get around 2X or more the performance of my current system.
            4. Price, how much is that upgrade going to cost you. For example the new AMD systems need a new motherboard, and ram, the system might need a new power supply.
            5. Don't be a fan boy. They are all mega corporations that are in the business to make money. They have no love for you, your identity isn't based on what you purchase. If you relate who you are to the brands you purchase, I suggest you talk to a counselor or therapist. I have owned blue and red over the years based on the one that best met my criteria at the price point I wanted.
            Keep in mind it can be easy to get caught up in the "Look at that shiny new thing and I will be faster than all of my friends."

            Comment


            • #46
              Originally posted by OnetyOne View Post
              Intel is offering 14 cores unlocked K and KF CPUs around $300 price point where AMD only offers 6 cores and you are complaining about Intel’s prices? How about lowering the 7700x that gets beaten by an i5?
              Yes I am. Technological progress should result in more generational performance/features/resources for about the same amount of money (inflation adjusted). Not more performance/features/resources for more money. Companies try to brainwash users that the latter one is a norm now - but it should not be. "Thanks to" such outlets as LTT who regularly rationalize corporate side of story and help to propagate such a narrative.

              Comment


              • #47
                Originally posted by atomsymbol
                There is very little information on that webpage. Please stop referring to that webpage.
                From https://www.researchgate.net/publica...er_Consumption (2014):
                ".... This shows that performance oriented scaling algorithm finishes faster compared to energy-oriented algorithms thus providing the system with a faster run-to-idle performance. However, even executing the sequence at more than double the clock frequency (0.798GHz versus 1.862GHz), the performance gain are not linear and we only improved the execution performance by around 8%. We cannot expect that doubling the clock frequency will translate to double the execution speed as shown in the result.​"

                but they do not go from 0,79ghz to 1,86ghz... they go to 6ghz...
                also this claims that nothing else change but only the ghz number but in reality other part of the cpu also change like more cache and other stuff.

                also what you suppress is the point that some stuff can not be optimised but only go faster with higher clock speed.
                this means you either clock higher and then do race-to-idle or else you have slow cpu and bonus to this high power consumtion.
                Phantom circuit Sequence Reducer Dyslexia

                Comment


                • #48
                  Originally posted by atomsymbol
                  Just a note: The boiling point of [pure] water depends on [atmospheric] pressure. If the pressure inside a liquid cooler is higher than 1013.25 hPa (= atmospheric pressure at the sea level)​ then the liquid (if it happens to be pure water, which is unlikely) will not be boiling. Adding chemical compounds to pure water changes the boiling point of the liquid. Even if there was pure water in a liquid CPU cooler and it partially turned into steam: the steam would increase the pressure inside the cooler and thus the liquid would stop boiling - question is where the steam would be formed (if it formed near the cooler's pump it would decrease flow of the liquid in the cooler; movement of bubbles inside the cooler would produce extra noise).
                  https://en.wikipedia.org/wiki/High-altitude_cooking
                  do you know the 3M cooling liquid ? ( perfluorinated compounds (PFCs), stable fluorocarbon-based fluid,)


                  3M Liquid has much higher cooling energy density per kelvin 50% more than water and the boiling point is much lower than water.

                  "An example of one of the compounds 3M uses is FC-72 (perfluorohexane, C6F14). Perfluorohexane is used for low-temperature heat-transfer applications due to its 56 °C (133 °F) boiling point. Another example is FC-75, perfluoro(2-butyl-tetrahydrofurane). There are 3M fluids that can handle up to 215 °C (419 °F), such as FC-70 (perfluorotripentylamine).[2]​"
                  this 117C cpus is very interesting with these 3M liquids if you do not want it to boil you use this FC-70 with boiling point of 215C
                  but if you want to boil you use this FC-72 then it boils at 56C...

                  you cold build a closed very powerfull heatpipe like 3M liquid coolings system who can easily handle 1000watt cpus...

                  because the cpu goes to 117C and the boiling point is 56C this would mean you can push limitless power in the cpu without the temperature goes higher if the cpu is in a liquit like this.
                  Phantom circuit Sequence Reducer Dyslexia

                  Comment


                  • #49
                    Originally posted by yump View Post
                    qarium The efficacy of race-to-sleep is greatly exaggerated. Basically every CPU runs well above the range of frequencies that are so slow that running faster and then sleeping would save energy. Try running this on a mostly-idle system:
                    it looks like this race-to-idle is just a byproduct to the fact that many things in chip technology can only be made faster with higher clock speeds. this means you optimise cpus mainly for higher clock speed to speed up these stuff who can only be made run faster with higher clock speed and the race-to-idle effect is just a byproduct to that design philosophy.

                    all the highend cpus have high clock speed for this reason even apple M1/M2 have the highest clock speeds for all ARM cpus out there.

                    if a company like intel has a goal of 6ghz cpu the race to idle effect is a byproduct they get for free.
                    Phantom circuit Sequence Reducer Dyslexia

                    Comment


                    • #50
                      Originally posted by OnetyOne View Post
                      Show us on this doll where Intel hurt you.
                      like 25 years ago i had 1 time in my life bought an intel cpu it was a pentium 3 with 450mhz.

                      and this was at a time i was so young i did not unterstand what is a monopole

                      or what is evil politics like patent war and ISA war and so one and so one.

                      then i did grow older and i did unterstand all this and i never bought any intel system again.

                      Phantom circuit Sequence Reducer Dyslexia

                      Comment

                      Working...
                      X