No announcement yet.

OpenGL performance MoBo/Nvidia upgrade.

  • Filter
  • Time
  • Show
Clear All
new posts

  • #21
    Originally posted by Kano View Post

    So you would save impressive 27?. That must be a joke when you calculate that in % of your complete system. Why do you consider that expensive when you get MUCH more speed?
    10% is no joke IMHO.
    Again, I needed 2 things:
    1) Mobo through-output, and I got it.
    2) RAW bogomips, since i use gentoo and compiling takes time even with 12 GHZ CPU
    3) Nice bonus, 4 GHZ speed for single core applications - perfect for gaming, probably much better than what Intel offers at same price.

    Please refer to what was wrong in choice. What figures are wrong to you. Those 10% I save = enough to buy new cooler, which will help me squeeze more juice out of CPU.


    • #22

      Originally posted by Kano View Post
      Man, learn to read, they benchmark their GPUs with Intel systems!
      Intel is somehow magically superior to AMD?

      I got great Front side bus. I got better performing in most situations CPU compared to I5. Specifically at operation that i need the most from CPU - compiling my programs.
      I remember there was a time when Athlones owned CPU market. Intel couldn't beat them, so they decided to bully AMD with beating them unfairly in the market, AMD lost battle not because of inferior tech.

      Frankly I only see good things about AMD at the moment, with Vishera.
      Tests here from Phoronix are great proofs. Feel like doing some benchmark for comparison? Anyone got one of latest i5 here?
      We can simply do a test and publish it in different thread.

      I got no name GTX660 though and will upgrade my RAM very shortly.


      • #23
        Feel free benchmarking TF2:

        I used always the same settings, HDR on, bloom off, AA off, rest default:


        • #24
          Originally posted by Kano View Post
          My cpu is the S variant not the K but i would not worry about it. When you buy K to oc you need a better cooler anyway - if you don't oc then the cpu will run at that temp as expected but you will be able to hear the fan. I don't see a reason to monitor temp all the time and i really compile much. When it is withing spec all is ok - the cpu would at max throttle down but would never die. I think i pushed my cpu close to 100C while oc and it still lives.
          Intel boxes include different coolers, which depend on TDP. CPUs with smaller TDP include smaller coolers.
          82C was reached without overclocking. Desktop CPU can die soon if temperature exceeds 70C.
          How did you overclock 3770S?


          • #25
            The only ways to oc non K cpus are to raise the max turbo level and to increase the base clock (that means with any cpu below i5 only base clock increase is possible, in most cases around 5%). I could increase base clock to 105 and the max turbo is 43/42/41/40 - 4 steps higher than default. That means around 43*105 with 1 core at full load, 42*105 with 2 cores and so on. It is much harder to check this as you have to use a test tool that forces different loads. I think i increased the vcore by 0.2v. I overclocked the integrated gpu as well (around 200 mhz). Basically that way of overclocking is a bit stupid, the simplest way to get more speed out of this cpu is fixing all turbo levels to 39. That works without increasing vcore and is somehow the default with Asus boards as soon as you change memory settings (i have got no idea why). With this setting it was possible to use a Win 7 tool to adjust the baseclock and voltage. With full load the cpu easly reached 99C (using AC Freezer 7 Pro v2) but the speed increase was not really worth this effort. I prefer stock speed with TDP of 65W. I think 100C is then throttling begins - i dont think you will ever reach that with normal use. You can be 100% sure that Intel cpus do not die when they reach more than 70C.
            Last edited by Kano; 03-04-2013, 08:27 PM.


            • #26
              Originally posted by JS987 View Post
              Desktop CPU can die soon if temperature exceeds 70C.
              A bit too general isn't it? Any sources for this? If I look at (page 16) the max. temperature for all of them is at least 90?C. Why would they die 'soon' (what's soon? 1 year? 12 years?) if you're at 70% of the max. temperature?


              • #27
                Originally posted by droste View Post
                A bit too general isn't it? Any sources for this? If I look at (page 16) the max. temperature for all of them is at least 90?C. Why would they die 'soon' (what's soon? 1 year? 12 years?) if you're at 70% of the max. temperature?
                TCASE 69.1C
                To allow optimal system operation and long-term reliability, the processor must not exceed the maximum case temperature specifications as defined by the applicable thermal profile.


                • #28
                  Well I was talking about the core temperatures/TJUNCTIONS (and I'm pretty sure Kano too) and not TCASE.