Will you be able to remove the ram without removing the hard drives ?
Printable View
Will you be able to remove the ram without removing the hard drives ?
That is a total pass through. A guy at the last Intel LAN I went to had a boutique rig (not sure which one but a skull is their logo) in which they had a dual card DD loop in which the flow went down from card one to card two by passing card one as in this loop but instead of exiting the second card like this loop it returned the flow back up to the first card via the other barb and then exited the card loop via the right top barb . . .it would operate at stock but if you oc'd it even a hair all hell would break loose . . .
Well first of all, thx for the props :up:
I have a BS (bull:banana::banana::banana::banana::banana:) in Computer Engineering. This is not what fostered my growth (or education, really) in computers. I've been building (and taking apart) things since I was old enough to remember. I've had one of nearly every major computer ever made, and kept up on most of the architecture changes since I could read. Unfortunately this is becoming nearly impossible now with the changes coming so fast ...
Honestly the most important skill you can have is "googling". College teaches you jack squat, the most important skill you get from College is the ability to research something. I had to find crap in an Encycolpedia back in the day, but now I have a VAST resource for absolutely ANYTHING I want to know. That is what you call a paradigm shift
The key is to know WHAT to search for and HOW to find it! :shrug:
The next important thing is the willingness to screw up, and screw up often. The time you spend thinking about how to do something is time that you could have spent actually trying it. The moral is, do 10 minutes of research on say, creating cables, then spend a few hours actually trying it!
There is no substitute for experience in any trade. When people are awed with my fuel-injected '57 Chevy and ask me how I did it, it always seems to surprise them that I "just did it"! I tried and tried and by about the 5th time, I finally achieved something useful. Failure is the greatest teacher, and you can't be afraid.
The difference between a layman and an expert is very simple: experience.
Give it a shot and every time you fail, you have the opportunity to learn. Share with others when you get the chance (which means "dont be lazy") and try to create a dialoge with those people who are better than you.
So first, be clear on what you want, then go out and find as much information as you can. Then try it and see what happens! That's the best advice I can give you.
Good luck and when you get something done, hit me again for some feedback. :)
I know this works because I've tried it :shrug:\
Go out to EK and see how the cards flow, and you will quickly see why this works. Frankly after I thought about it, I was amazed that more people didn't do this. However it's always easy to see something after you understand it, just like you always find things in the last place you look :)
Except it IS forced to pass through both cards. The resistance in each water block is identical (or as close to identical as the manufacturing process allows), and the water is going to be forced to split between both blocks. Just take a couple of minutes to think about how the internals of the blocks are laid out. If he only had one of the tubes connecting the card, then yes, you're right, it wouldn't work, but with both connections there, it WILL work just like a Y-splitter, except the flow would be more even than you'd be able to get using a Y-Splitter.
Sunayknits: QFT is short for "quoted for truth", basically a way of saying you agree with something someone said.
Eller
Hmm, that sounds a bit goofy ... no surprise it was unstable. I think mine works well because the loop is super simple, and this loop is dedicated. Honestly I'm very surprised most ppl don't do it this way ...
Beavis, your Cherry build is blowing my mind. I have dreams about that crap :p:
QFT... good to know.
My first thought was that QFT stood for, Quit Effing Talking...... hheheehe
I am liking how your build is turning out. I had the same case, water cooled, and absolutely hated it. But, I also didn't put as much effort into mine.... I used to just put all the parts in, and hoped it would all work.... hehe
later,
breezyjr
Yeah this case requires quite a bit of thought and modding to get everything nice and tidy. I've had to move things around quite a bit to get it like I want. For example, I quickly ran out of 5.25" bays so I have to make a custom bracket for my iRams and that's gonna be a huge pain in the butt. :shakes:
So if a radiator can handle the capacity you should be able to create a 'virtual' two loops to cool a cpu and two gpu's
using a y connector to split the outgoing cool water: one goes to the cpu then a radiator
the other goes to gpu, gpu then the radiator
wouldn't it work?
I suppose it would probably work. The major benefit (to me anyhow) of 2 separate loops is to keep the gfx cards from heating up my proc/mobo. The cards generally run 15-20 degrees C hotter than the other parts ...
However the water temp doesn't change a whole lot through the loop, on the order of a few degrees at most, so I'm guessing 'virtual' loops might yield similar benefits even if they shared a reservoir.
Some more nice things about 2 loops: you can run different color coolant in each loop if you want, and you could potentially have them run the fans at different speeds, ending up with a quieter machine. Also, you will get better flow with separate loops. Just a few thoughts.
I mostly did it because this hobby is all about being eXtr3m3! :rolleyes:
in all tests shown gpu1 or 2 is seriously a lot higher.
The cores are no where near evened out.
Ive tried it as well, and i personally like my gpu's to be within 1-3C of each other when i sli or xfire.
That looping is another example of bling over performance. Sad to say. :\
Simular to how people throw in random 90's cuz they look cool.
The simplist option would be to paralell the cards if your that anal about tubing layout, and it would solve your uneven core issue.
oh one more thing, the best way to test it is to load up your GPU's and watch GPU 1 or 2 SKY rocket, while the other stays almost FLAT.
That is the result i and many others got while twiddling with this setup.
Was a great waste of 4 hours. :P
wow great build!
and i have a question for you (cause you have a mozart tx to, mine arived yesterday :P: ) i have read the manual but the pictures in there are soo bad and i can't find what i was wondering on the internet to..
so the question is: where exactly doesn the mylar tape supposed to go??
greatings, dope
a) I really don't understand your post
b) If you're talking about dual loops vs. single loops then it still doesn't make sense
c) When my machine is pulling 500 watts from the wall, with both GPUs going flat-out for many many hours, both cards are still within 1 degree
d) This IS a parallel layout
Forgive me if I misunderstood, but it's not clear what you're talking about.
Hey Dope, thanks man!
The mylar tape goes over the raised cutouts underneath the motherboard. I think there are 2. They could potentially touch the mobo and short it out. These mounts are designed for the BX standard IIRC, which is pretty scarce anyhow, so I just removed them with a cutoff wheel since I never plan to use them.
okey im sorry. Lemme see if i can simplify myself.
I am only talking about your gpu loop.
You said you had the first card setup as inlet outlet, and connected to the second card.
Basically the water is going in a U. The first card will heat up while the second card will stay still.
You said it was okey to config it like this. I am saying its not, because you'll have a very big delta between card 1 and card 2.
DB also warned you about this type of config, and you still said it was okey.
I am mearly double verifying what DB said. I have tried that config, and as i said card 1 and card 2 has too much of a temperature delta.
2. Can you show a pic of this. Cuz i never could get 1C within the cards. AS i said I noticed it being more then 10-15C varianace.
A program like ATI Tool or RTHDRIBL is great to load up your gpus
3. No a true paralell would have = flow to both cards. So inlet would be on both card 1 and 2 using a Y and exit via card 1 and 2 on outlet via Y.
Once again here is DB saying IT WONT WORK.
And i am saying ive tested this, and IT DOESNT WORK.
If you have pictures of it working. I would definitely like to see them. Im not calling shins, if you have a working setup, i would like to see how you pulled it off.
The closest i got to it was using a 3/8th id tubing instead of 1/2id tubing for the bridge on gpu1 and gpu2.
Adding restriction to the bottom cards will allow more water to flow to the top card. This might work, but i still wasnt happy with the temp deltas.
My last attempt although very painful, was to use some form of reducers to mirror the flow. Another failed attempted. :X
Sunay please dont get me wrong, i am really curious on how you pulled it off. I would love to tube like that, but my OCD wont allow my gpu's to be like that unless i can = the temps.
I like your Mozart TX Dual Loop
This my MoZart
http://i209.photobucket.com/albums/b...Build/b25r.jpg
http://i209.photobucket.com/albums/b...iator/R14r.jpg
http://i209.photobucket.com/albums/b...Build/b19r.jpg
http://i209.photobucket.com/albums/b.../Final/t4r.jpg
wait 4 upgrade 2 Dual Loop