Jump to content

Achtas

Member
  • Posts

    12
  • Joined

  • Last visited

Reputation Activity

  1. Informative
    Achtas reacted to Jurrunio in ASUS Crosshair VII and RAM instability   
    try disable GDM (gear down mode)
  2. Like
    Achtas got a reaction from Jurrunio in ASUS Crosshair VII and RAM instability   
    Alright, took a few days to try and see if it resolved it but sadly to no avail. Whilst no outright BSOD, applications still occasionally just close or crash (AutoCAD, Warframe, World of Tanks).
     
    As far as I can tell, yes. At least according to the manual, 2/4 is the recommended combination.
  3. Agree
    Achtas reacted to 19_blackie_73 in The CHEAPEST Heatsink on the Market   
    that's the thing I don't understand in recent LTT videos:
    It is ok to test it with a higher watt cpu for the lol factor, but if it is only designed to cool a 65 W TDP cpu, WHY DON'T THEY TEST IT with a 65W TDP chip to confirm at least if it holds up to what it promises to do?
    A really informative video would do the test with the 65W TDP chip, comparing intel stock (that actually comes with the 65W chips intel makes and not slapping one generic version of some other whatever TDP chip on the 7700k) and the 5$ chip. And to have a real comparison, you should run the tests with the stock thermal paste that comes with the coolers and one aftermarket on both so you can see if the heatsink is more limited by the thermal paste that comes with it or the heatsink itself has a bad design. And after that, you could slap it on the 7700k and see what happens.
    This video is like having a opel corsa with 69 PS, going to a drag event with tuned american muscle cars and then be suprised that you have the slowest car.
    I started watching LTT videos back in 2015 when choosing parts for my first pc. Back then the videos were informative and entertaining at the same time, but since they moved to the new studio, they went more and more crazy and less and less informative. Why can't you @LinusTech start again combining both and not going holy shit and over the top on everything? If the trend continues, I'm probably stop watching the videos, as this shifted form of entertaining is not my taste anymore, and I'm probably not the only one. Clickbaity titles is one thing, but keeping a level of information provided in videos should be possible. Because anybody who will buy that heatsink, probably won't slap it on a 7700k, but probably on a pentium/i3 and will now know not more than before watching that video.
  4. Like
    Achtas reacted to aisle9 in The CHEAPEST Heatsink on the Market   
    @LinusTech logic:
     
    "Hey guys, let's take a heatsink clearly labeled for a 65W TDP, throw it on a 91W freaking i7-7700K, then call it crap because it throttles!"
  5. Informative
    Achtas reacted to Rolling Potatoe in Liquid Ultra on r9 295x2 (CLU)   
    Lately Ive been messing with my pc a bit so I though my GPU might need a repaste (1-2 years old).
    My r9 295x2 runs cool enough with 2 fans (usually 68-70 °C tops with my fan curve) and stock clock (although memory is oced a bit and its undervolted by 40 mV),
    but review units seem to run several °C cooler usually. Running aida64 for 15-20 minutes on fixed fan speed gave me a max of 73°C (22.5°C ambient) on gpu2 (it gets the hot water from gpu  1) and a tiny bit of throttling from 1018 mhz down to 1002 mhz at times or so.
     
    Ive never used any of the supposedly amazing metal thermal pastes, so i though I might as well give it a try and use the nice stuff.
    Also I replaced the vrm thermal pads with some 1.5mm 7W/mK phobia ones.
     
    From the reviews I expected a 2-10°C temp drop, Id say 4 °C would be worth the bother.
    Lets see if 9 euro for 0,15ml of thermal paste is money well spend.
     
    To be honest the thermal paste job from sapphire is not bad, the paste is still malleable and not dry. The thermal paste layer is also not overly thick.

     
    The vrm thermal pad is only 1mm thick I think and very soft, almost pasty. I couldnt even get it off in one piece.

     
    The new pads on the heatsink, shiny white and new :D.

     
    After cleaning I put about 15 ul(0.015ml) CLU on the left and 10 ul on the right (very big, 438 mm^2) gpu die. (I thought some would get absorbed by the brush)
    I cut the hairs of one of the uncluded brushes short, which made the application very easy. It was pretty much like painting the gpu.

     
    All shiny and chrome, you can maybe see there is a bit less CLU on the right gpu
     

     
    Still quite alot left in the tube and very little stuff on the brush.

     
    Time to measure some temps, launching Afterburner I was immediately pleased to so the temp on gpu 1 at 28 °C. normally it would be 10°C hotter on a cold start.
    Launching aida 64 showed a problem though:

     
    My second gpu is a whole 12 °C hotter than gpu1, pretty much like with the stock paste in terms of thermals.
    I did put a bit less on the second gpu, so I opened the card up again to reapply a bit.
    Looking at the cooler surface it doesnt even seem like the blacks made proper contact, even though temps were superb on gpu 1.
     

     

     
    I was to lazy to clean the blocks, so I just spread it the CLU that was on there into the shape of the die and painted an additional 5-10 ul on each of the gpus.
     
    Time to test again, after 15-20 minutes of aida64 with fixed fanspeed and 22-23 °C ambient
     

     
    Some spikes to 63°C on the second gpu later on. 10 °C shaved off flat, Im happy :D. There was also 0 throttling.
    The vrm thermal pads dont seem to have made quite such a big difference, I can lower the vrm fan speed by 5% now without the gpu throttling though,
    so thats something.
     
    After ending  aida 64 the gpu temps only decreased rapidly for a few °C, seems like the water is not that much cooler than the gpus now.
    For any further cooling optimisation id probably have to got some more radiator space in the closed loop somehow.
     

     
     
     
    For now it seems Coollaboratory Liquid Ultra is pretty awesome, at least if you put it directly on the cpu or gpu die.
    If the temps deteriorate over time Ill let you know, for now I think I can give a recommendation.
     
    Results: 10°C lower temps
     
    Advice:
    1. Cut the hairs on your brush short, its easier to apply and wont suck up your expensive thermal paste this way.
    2. If you want to use a metal paste, Liquid Ultra is very easy to apply. Ive heard Liquid Pro can be a bit hard to apply.
    3. At least for a profiled copper block id say use 5ul (0,005ml, 0,5 on the syringe) per 100 mm^2 of your cpu or gpu die.
    Everything should be covered with a shiny, wet looking metal layer, if it looks dull like a scatched open product code you probably need a little more).
    This seems to work well and also has the advantage of making the syringe last for 10-15 gpus or cpus, I still have 120ul left.
    4. NEVER make a mess with Liquid Ultra or any other kind of liquid metal thermal paste. Its main ingredience is gallium (alloyed with a bunch of other metals with a graphite matrix), so it will eat through aluminium like you wouldnt believe and will conduct electricity extremely well.
     
    5. I know people like the dot method and stuff, dont do that with this stuff. You have to spread it thinly with the brush, if you do a dot it will form liquid metal beads that could slip anywhere and kill your pc. Its a gallium based thermal paste, not a silicone based one like the regular pastes.
×