Jump to content
HWBOT Community Forums

phobosq

Members
  • Posts

    547
  • Joined

  • Last visited

Posts posted by phobosq

  1. Cough - please avoid 3:5 divider on socket 775 because it is bugged - I have to delete result because of this, you can use any other divider to make validation but not 3:5, which is in reality 2:3 - good mems nontheless :)

     

    I was so proud of it and that's how you dealt with it:

    tumblr_mnzw034wmV1s8njeuo1_400.gif

     

    Anyway, yes, 3:5 is bugged on 775, and safe dividers for DDR2 are 1:1, 1:2 and 2:3. I don't know about DDR3 on 775 because most of the time I use 1:2 :D

  2. So basically what you want is to make OC like chess or bridge - there are live streams of chess/bridge tournaments, but without commentary it's not that entertaining, even to enthusiasts.

    I think I understand your points - Massman asks "can we make it more popular", JunkDogg makes comparison to athletics and der8auer replies from his experience of doing 4-way LN2. But maybe that's the point - if the community started breaking the records live, then it would move OC closer to sport and gain additional attention? :)

    I remember there's was a bunch of guys in Georgia a year or so ago, that took Cedar Mill celly to university and did 8G+ on LN2. I think that's the spirit of live OC we're looking for.

  3. ...also in all my current testing on PSC vs Sammy they are totally different volts required in "imc,ioa,iod"....

     

    Care to share? I know it depends on a lot of factors and I am pretty OK with my air settings for PSC of +0.1/+0.1/+0.15 (+0.2/+0.2/+0.25 on phase), but I'm completely in the dark with Sammies, at least with G3258.

  4. Of course I can understand your frustration. But if you have problems, ask at forum or with PM for solution - the XTU not showing frequency bug is long solved, newest XTU and win7 for example work easily including showing frequency for me, which was last years problem. There was shown nothing at your XTU, try a new clean win7 sp1 install and use newest XTU version and test again, this should work :) I have removed results that show 45x100 4500 MHz because XTU showed spike of 4,61ghz as max, there is no way I can tolerate above limit at a low clock, for noone

    On the GPUPI, you cannot hide bench settings, even if you work for them. And others can´t as well, this is simply not possible because you need settings to check plausibility. If we would not understand that you worked hard for your bench results and LN2 and hardware is costly the GPUPIs would have been removed instantly.

     

    I saw that happen after I updated BIOS on Z97 OCF to 1.30, on 1.10 the clock reading in CPU-Z was stable, on 1.30 it fluctuated -5/+5 MHz from setting in BIOS. Anyway I don't argue with the rules, but it was also a point of concern for me.

  5. I have about 10 sticks of Ballistix 667C3, on DFI P35/Asus X38 they do up to 400-420 CL3 with 2.2 easy, but I fear that 2.1 on ASRock VSTA is like <2.0 on one of those above and 400 CL3 is difficult.

     

    @trodas - I see that you're quoting my offer of Team 800 C3 :P I still have them, although I only tested them on Gigabyte EP45-UD3R, which has this unique ability of running mems like no other board. Besides, those Teams are binned for 800 CL3-3-3-8 @2.3V.

  6. I used DI for RAM and CPU on Z97 OCF, I have kneaded eraser around socket covered with paper towel, armaflex and some more towels. For RAM I cover unused slots with electric tape and put paper towels everywhere around EK modules and pot, I never got moisture in 24-pin with that, but I believe that for LN2 it might not be enough.

  7. What do you consider as good cold scaling for Haswell? I had a few G3258 and a 4690K, and I saw 200-400 MHz increase in max frequency with the same voltage when going from air to SS. I tried DI, but it didn't give anything probably due to TIM.

     

    Questions:

    1. 400 increase from air to SS - good or not enough?

    2. How does it look with i5/i7?

    3. What more can I gain doing LN2, is it proportional to temp drop?

    4. How does delidding impact cold scaling?

     

    To be more precise - I measure "increase" in a following way: first do SuperPi 32M at maximum stable settings on air (for example 4800 1.4 vcore), then change to phase, go up with frequency without changing volts and keep running 32M until it fails. Then it's just subtracting one number from another. :)

×
×
  • Create New...