Jump to content

KraZy

SETI Member
  • Content count

    251
  • Joined

  • Last visited

Posts posted by KraZy


  1. I have to say that the Swiftech kits are good, but they do have problems with the rads. The single rad does a *ity job of dissipating the heat from the water. If you plan to to any overclocking, you will be forced to get another rad of the same type, or replace it with a larger one like I did.

     

    The water blocks on the other hand, are impressive. Given that the rad could not dissipate the heat like I wanted, was proof to me that the water block was doing its job, and then some.

     

    Also, they are quiet. Verrry quiet. :) But this is common place with H2O systems, anyway.


  2. Given the near infinate combinations of hardware and environments, there is no one that can say that one method will work better than another. Yeah, if I had an extremely controled environment and some kick-butt cooling, I would be able to get some of those massive OC's too.

     

    (not picking on you Nuke..)

     

    Now take Nuclear for example. Rumor has it that he has quite a few CPU's and has had the time to test them. It seems that he has done his research into batch-numbers and the such. He knows what he is getting, and he seems to know his CPUs. I on the other hand have unmached trouble shooting skills and can tell you when and how a component is starting to fail. The point being, you have to know the limits of your hardware and how you can push it.

     

    I think you can usually start off with a resonable overclock from the start, and then work from there. My typical starting point for a system burn-in is usually around the +200mhz range. Call me crazy, but it works for me.


  3. why not set the client to use 100% and set its priority to idle. that way it'll make max use the processor when not in use and not get in the way when you are using ur computer.

    Actually, the folding is set to %100 usage, but the p4's HT physically knocks it back down to %50. (*that is, %100 of 'CPU 1' which seems not to interfere with any of my other apps.. )

     

    I played around with the priorities a little and can balance the load between the two 'cpus', but still couldn't get its usage any higher. Is there something that I am missing here? Turning off HT is not an option due to heat..


  4. Aight, I have been running my P4 3.2EE (WinXP Pro) against my AMD 2800 (Mandrake Linux) in folding for the last few days. Much to my amazement, my AMD is only behind by 3WU's. (p4 = 13, amd = 10) One would think that the EE would be making better time than that... Personally, I think Linux has alot to do with it.

     

    Both computers are using about %50 processor time for folding 24/7.


  5. Aight, if you look at my case (at the bottom picture), you will get a better idea of how I did my stealth...

     

    1. First, I dremel'ed off the edges of the cover plate.

    2. I filled the back of the plate with a layer of card board.

    3. I used -RUBBER CEMENT- (or equvilent!) to glue it on to the front of the cd-rom door.

     

    This took some good aiming, and not too much time. BTW, I stress the rubber cement because you will want the plate to be able to move back and forth a little. This helps when positioning the drive (during rebuilds) and allows the old button to be used.

     

    I have two ways to open the door, btw. 1. Software. 2. Push the lower right corner of the plate and listen for the satisfying 'click' of the button.

     

    (side note: whhooohoo! just noticed I got my Folding icon... finally....)


  6. Ditto on the crossover cable. Lets get into some cable basics while we are at it..

     

    Remember that each pin on an RJ45 connector does something ... in our case, the ones we need to look at, are the transmit (TX) and the recieve (RX). If you use a plain network cable to try to do networking, you are basically hooking up the TX to the TX and the RX to the RX on both computers. It aint gunna work. A crossover swaps the TX/RX wires on one side of the cable so TX goes to RX and vise versa. Simple, hu?


  7. just a pelt, although its -2 -5 idle full load 24/7 it eventually just gets to room temp with it

    My setup was the same way. I am sorry, but the idle/load differance S*CKED with my pelt. Got to go with a phase-change system.. if it can keep my cpu sub-zero under load, I am sold.

     

    I will have to say that my P4EE was one of the best investments that I have ever made. It overclocks great, and it runs cool. (heck of alot cooler than my 3.0C did, anyway.) I will see if I can break 4ghz on H20 tonight....


  8. (btw, any one got any good reviews on that --^ hsk? the inital buzz about the heat-pipe approach was kinda weak...)

     

    -sigh- Baaad reviews on the Jet in the OC world, anyway. Coolermaster needs to back out of the hsk market i think. (keep to the cases fellas.)

     

    Thermaltake had been one of the most reliable OC'ing hsks for a while now. Good choice.. just do a search in the forums for it and you will see all the reviews.

     

    Dunno about the Globalwin...

     

    My personal favorates are the Swiftech coolers, but I have a fondness for the Zalman quiet coolers as well.


  9. Go with the 'grain of rice' AS5 application, and let the hsk do the work. With the thousands of times I have taken my HSK off and put it back on again, I have found that method the most reliable. (this might sound like sacrilage here in these forums... but...) I look back at all the time that I spent getting that 'perfect' paper thin coat all over my die, and it was a waste of time and greese. Go rice.


  10. It took me a while to actually put down the $ for a good NB cooler, but now that I have it integrated into my water cooling, I can really tell the differance in my system stability. I don't think I will ever go back to stock cooling for my NB again. My advice is clear, get the best you can afford. Because something is always better than nothing.


  11. I rechecked all the connections carefully, but couldn't find anything amiss, so I ended up taking it down to the PC shop and they said I plugged the VGA power in backwards. So I did all this work and ended up paying $70!

    what the heck? Ok, please tell me it wasn't a molex. Granted, I have actually -seen- a molex plugged in backwards, this is next to impossible. (i thought the 25pin parallel cable was impossible to plug in backwards, until i saw that one day...)

     

    Unless you are using the smaller 4-pin power cable... ... ... I think that is one-directional too....


  12. Last time I tried to remove my heatsink it was suction attached to the chip. It would slide all around but would not come off.

    I had this problem when I used too much AS5 on my P4. I couldnt even take my water block off with out it pulling the cpu out of the socket. (yeah, the zif socket was locked too.)

     

    It just takes time man. If there is some of that ultra-nasty goop on it that is more like rubber cement (as seen on a ABIT northbrige), just be really carefull with it.

×