Announcement

Collapse
No announcement yet.

Intel vs. AMD Performance-Per-Watt On Ubuntu 14.04 Linux

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

  • #21
    Originally posted by ua=42 View Post
    Watts x time = joules (Smaller number would be better)
    time/watts (smaller number would be better)
    Yes to the first.

    No to the second. You have that wrong. Short time and low power are desirable. But you have that formula preferring short time (okay) and high power (not okay).

    Comment


    • #22
      I am curious how the Iris Pro favors against the AMD. Intel is obviously better w/ cpu. The Iris Pro is the best intel has to offer for gpu.

      Comment


      • #23
        I guess the correct formula to use depends on what performance means. If you only care about the power spent to finish some workload, then the formula would be workload/(total power used). Because the workload is the same, you can use 1/(total power used) (larger is better).

        If by performance you mean the rate at which the workload is completed, you can use workload/time. Again, because the workload is the same, you can use the inverse of the time (again, larger is better). This just for the "performance" part. Divided by total power used, you get (1/time)/(total power used), which is equivalent to (total power used)/time.

        So the charts are fine if they used it, as ua=42 says they do.
        Last edited by cataphract; 18 January 2014, 02:01 PM.

        Comment


        • #24
          Originally posted by cataphract View Post
          I guess the correct formula to use depends on what performance means. If you only care about the power spent to finish some workload, then the formula would be workload/(total power used). Because the workload is the same, you can use 1/(total power used) (larger is better).

          If by performance you mean the rate at which the workload is completed, you can use workload/time. Again, because the workload is the same, you can use the inverse of the time (again, larger is better). This just for the "performance" part. Divided by total power used, you get (1/time)/(total power used), which is equivalent to (total power used)/time.

          So the charts are fine if they used it, as ua=42 says they do.
          work/time/energy simplifies to 1/time/energy as you've said if we're comparing the same work but that is rearranged to 1/(time*energy).

          Comment


          • #25
            Originally posted by cataphract View Post
            Divided by total power used, you get (1/time)/(total power used), which is equivalent to (total power used)/time.
            No .

            (1/time)/power = 1/time/power…

            1/(time/power) = power/time. (which is not what we want.)

            Edit: beaten by vick.

            Comment


            • #26
              Originally posted by ua=42 View Post
              http://phoronix.com/forums/showthrea...d=1#post390102

              Yeah. Just posted a post in the test suite forum. Hopefully Micheal will see it or this forum thread.
              A lot of people already commented that his performace-per-watt articles are always wrong (this isn't the first) and yet, he keeps posting useless articles that are completely wrong. It probably generates more comments, and more clicks, this way :P

              Comment


              • #27
                Originally posted by ua=42 View Post
                The workload was the same.

                So you would just need to know the joules used.

                Watt x time = joules.
                Like this:
                cpu i3 2120 i5 3470 i7 3770k i3 4130 i5 4670 i7 4770k A10 7850k
                P_avg[W] 63,7 76,4 96,8 75,6 91,1 113,8 116,3
                time[s] 202,1 118,8 88,92 161,96 94,03 75,43 163,14
                Energy[J] 12874 9076 8607 12244 8566 8584 18973

                Comment


                • #28
                  Well done, tuke81 !

                  Comment


                  • #29
                    Originally posted by vick View Post
                    work/time/energy simplifies to 1/time/energy as you've said if we're comparing the same work but that is rearranged to 1/(time*energy).
                    Oops, you're absolutely right

                    Comment


                    • #30
                      Power vs. Energy

                      Originally posted by devius View Post
                      These results are all wrong again.
                      Let's take the case of the Timed Kernel Compilation with the i3-4130 and the A10-7850. They both take about the same time to finish the test, but the i3 uses 40W less power on average to do so! That's about 34% less used by the i3 to do the same job in the same amount of time, which means the i3 is more power efficient than the A10. However on the bogus "performance-per-watt" chart the A10 appears with a much better score. [...]
                      The confusion comes from the term "power consumption". Many people think "power consumption" to be equal to "energy consumption". In fact, the energy consumption is the integral of power consumption over time. Therefore your complaint is more about the fact that you would like to see a conclusion about "energy consumption". On the other hand, the problem with energy consumption comes with workloads that usually run "forever" or at least not with a fixed amount of time e.g., playing a computer game. For the latter, power consumption can be treated equivalent to energy consumption.

                      Anyhow, it can't be repeated often enough: consuming a little bit more power over a much shorter duration still holds a lower energy consumption.

                      Comment

                      Working...
                      X