*Official* NBR Desktop Overclocker's Lounge [laptop owners welcome, too]

Your GPUs are the SC and mine is an SC2. Mine also has iCX memory and VRM thermal sensors and if I remember correctly the SC does not. But, the second slider is irrelevant with water cooling. With air cooling, the G slider is for the GPU/vRAM fans and the P (power) slider is for the VRMs and MOSFETs fan. With our hybrid coolers, only the VRMs and MOSFETs have a fan and heat sink. The GPU and vRAM are liquid cooled, so the G slider doesn't change anything.

Go ahead and do that. It's not like Maxwell voltage and maxed out still isn't enough. It only allows a Mickey Mouse +100mV offset and really needs to be more like +250-300mV.

 

No. Not even for Clevo 1080. Maybe someday (hopefully) he will.

 

Cool! I want it though!

I've read your other posts about overclocking, and not being able to overclock it very far, that's a pity, what boost clocks are you actually hitting during benching & gaming at your max overclock? I'm less enthusiastic about the voltage slider than Mr Fox - I have read that it can actually decrease stability if you just max out the voltage slider, and for my GTX 1070 the sweetspot in terms of using the voltage slider is 50%, which is enough to gain an extra notch of stability (a whopping extra 13Mhz! ;-) ). You might have just been unlucky in the silicon lottery, and with all those extra transistors in the GTX 1080ti it makes it less likely to be able to be stable at high overclocks.

 

Nice! I used to feel that way at my old place, but when I bought my house last year I hit the jackpot with UPS/FedEx and USPS delivery times. They usually come by 900-1000am. . When I'm up chilling in the living room at 7am my wife knows I have a toy being delivered that day.

 

People said the same about it to me with the Clevo 1080 voltage slider and also said clocking the memory too high would lower benchmark scores. It never happened. What we don't know is why they said that and why it never happened. Maybe those that say those things were unlucky in the silicon lottery. But, it still needs to be tried. Poor overclocking can also be GeFarts cancer driver induced. I've seen more than 500MHz difference in GPU memory overclocking capacity with different driver versions, which is really retarded. Maxing out the voltage slider results in a minuscule amount of added voltage. The Green Goblin screwed us all pretty much with Pascal design. For the most part, Pascal is only a good product because it has stock clocks much higher than Maxwell. The GDDR5X is also a huge upgrade from Maxwell for 1080+ GPUs. The interference with overclocking and engineered lameness like losing core clocks at totally normal operating temperatures is pretty sad though. It's even more sad that it's still miles ahead of AMD GPU performance even though it is all screwed up. What a sinfully pathetic mess computerland has become.

 

I've tried to wrap my head around why they designed Pascal the way they did in reference to the clockspeed reductions at such low operating temperatures. The only thing I can think of is core instability at said clockspeed when the temps rise.

 

Yup, probably so. That's why I think water cooling is the best way to get the most out of Pascal. If you want to keep the core up at full boost clocks it has to stay abnormally cool.

 

Yeah, people definitely should test out the voltage slider to see if they can get a higher overclock, probably in 25% increments from 0-100%, or if lazy I'd suggest just trying 50% and 100% - (I can't quite recall if 25% even gave any additional voltage).

 

Ok, I've set a new record for my highest Firestrike score (I've got GTX 1070) with the latest NVidia driver that came out yesterday (GPU score of 21,359), it's the same score as stock Founders Edition GTX 1080 now which I'm pretty stoked about:
https://www.3dmark.com/fs/14810862

(standard stock scores for GTX 1080: 20,268 from notebookcheck.net review, 21,905 from guru3d review) ( https://www.notebookcheck.net/Nvidi...sktop-Review-Pascal-has-arrived.165500.0.html,
http://www.guru3d.com/articles_pages/nvidia_geforce_gtx_1080_review,28.html)

The newest driver says in the release notes that 3DMark performance issues have been eradicated - some systems (including mine) experienced lowered Firestrike GPU scores with the latest NVidia drivers over the last year, and this latest driver solves it.

This puts my system at #219 in the World for i7-6700K combined with GTX 1070 - that's not too shabby considering over 60,000 Firestrike runs have been done by people with that config, putting my system in the top 0.3% - ha, I'm trying to spin this positive! ;-) Will update my sig with the latest score. I had it on 100% manual fans and core frequency was 2100-2114Mhz at all times, max temp of 55 degC. This is a rock solid 24/7 overclock, not a one off Firestrike run, could get a few more notches on core & definitely on VRAM by running a borderline stable overclock just for benching purposes, but this is a rock solid overclock.

@Mr. Fox , @Raiderman , @KY_BULLET & others, have you managed to get higher scores in Firestrike with the latest NVidia driver that came out yesterday? (driver 390.77 whql)

 

Nice job @Robbo99999 . I really hope Overclock.net gets their crap together soon I still cannot log into it >.>. If this continues I'm probably just going to order Primochill tubing or something and hope I buy the right one.

 

If you are going to use flexible tubing, I cannot imagine why it would make any difference which tubing you get as long as you get it from a reputable source. There is a good chance they all sell the same tubing with their own brand on the box. The coolant is probably more important. Maybe post a question for Jay on one of his liquid cooling YouTube videos (JayzTwoCents).

 

I havent tried the new driver yet, but I will. I have been a little leery of using a Ngreedia driver that is released close to a new architecture change. Ive heard they crap on the driver to make performance look bad just before releasing the next gen chip. I will try it and post results.

 

Good call on talking to Jay. He sounds like he knows his stuff when it comes to Liquid cooling. He just finished a custom loop a couple weeks ago on his personal build.

 

Ive been thinking about switching to hard tubing. It looks so nice and clean. Has anyone heard of any temp differences between the two?

 

800 points up on graphics

 

Yeah, so that's due to the 390.77 driver then, looks like it does improve performance in Firestrike for probably a lot of people. Timespy didn't seem affected though.

 

I cant run time spy, however I did order a cheapo SSD to install Windows eXcrement on. How long it stays on that drive is up to it.

 

Ha, well fair enough!

 

Below is a down and dirty Fire Strike in W7. No optimizations with killed services and what-not, just set the overclock and go.

Not as good as my two best scores, but still quite good. Looking at the link, you can see the evidence that W10 sucks the life out of a CPU, but Fire Strike imposes less of a penalty for CPU than all of the more respectable Futuremark benchmarks. W7 has about 7% higher Physics results in the same benchmark. The higher Graphics score in W7 is an anomaly, and K-Boost was enabled. K-Boost was not enabled on the slightly lower W10 score.



Since I am still waiting on the Tt Floe Riing cooler, I pulled off the side panel and put a fan blasting full speed on the CPU and GPU to address the inferior capacity of the air cooler. (I have around 400 CFM of cool room air getting crammed into the case with the AIO with all of the fans maxed out.) The air cooling seems to emit a LOT more heat into the case than the Floe Riing does. Of course, the three Vardar Furious fans on the AIO blowing like a hurricane on the GPU back plate with the Floe Riing was pretty useful as well.

 

Only a another day until you are under water again

 

Yeah, feels like it is taking forever for that RMA shipment to arrive. Can hardly wait. Thermaltake sending it to the wrong address and having to correct the delivery address really threw a monkey wrench into the works as far as delays. They should have sent a replacement package and asked UPS to cancel/return the one already sent and I would likely already be on water again.

 

I never thought of that! Maybe you should have told them to send a new shipment? On the bright side, I just tracked your power link, and it should arrive the same day as your cooler
Looking at your setup, at least you will be able to use yours.

 

Hindsight is always 20/20. If that ever happens again I will suggest it. Never had a mix-up like this before, but yes... one more day, and everything will be right again.

And, it has really taught me to appreciate the liquid cooling. Next project will be to incorporate a water chiller... not sure when I will have the money for it, but... Hailea HC-500A 110V (1/2HP=790Watt cooling capacity) Waterchiller

 

Thats a nice unit! Nice to see Koolance hasnt changed much. Still over priced as usual. $1600 for an 800w . Would be foolish for me to invest into anything like that with 1st gen Ryzen having very little overhead, with it topping out a little above 5ghz on LN2. Maybe the refresh will be different, I hope. Its just not as fun, when you have so little to work with. Still, from what I have read the 4.1ghz I am able to get is not too bad for this chip. No FSB overclocking either with Ryzen

 

The Koolance units look amazing, but seriously crazy overpriced for what you get. Not having much wiggle room for overclocking is the primary reason I did not consider a Ryzen 1800X. 1700X and 1800X are both very potent CPUs running stock clocks, but not being able to do any kind of crazy overclocking was just not something I could get excited about. Overclocking is my primary form of entertainment, so it's of utmost importance. If I were more excited about PC gaming than overclocking it would be a non-issue. TR is as much out of my budget as an i9 Extreme CPU, so that left me with 8700K as the only acceptable option. I love the 8700K, but the limited choices for overclocking enthusiasts sucks for the same reason that it sucks that the Clevo P870 is the only laptop that exists that I find to be an acceptable option. On one hand, CPU and GPU performance have never been so incredible. But, on the other hand, the severely limited options for those that value overclocking above all else are pretty sad.

 

There has been some talk about condensation here recently. Just curious since this water cooler would be significantly lowering temps in the lines, and with there being warm air inside the case, if that would cause condensation build up on the water lines?

Also, a crazy/funny thought just crossed my mind since I was a body man back in the day. Has anyone ever tried to run AC freon with a small ac pump through some lines to cool the cpu and gpu? But even if it was possible, I know that would build up condensation but damn would it be COOOOOLD! Probably to the point to where it would be too cold and possibly break something. Probably be the closest you could get to Liquid nitrogen without the mess lol!

 

No real issue with the chilled water because it does not get cold enough for that. These things were originally designed for fish aquariums. Maybe if you lived in a super hot and abnormally humid area, but not for most folks. I think they generally keep the temps at around 10C to 20C (50 to 68F, so not much different than air conditioning) during normal use and max load, with less variance between idle and sustained max load temps.

The latter is known as phase change cooling. It works great, but condensation is a huge lookout on all sub-zero cooling solutions. Linus has a nice video showing a phase change setup. I will post it if you have trouble finding it.

 

Hummmm....This is with my Aquarium water chiller un modified....
And a Phase Change for the cpu

 

^^^^^^^......and that my friends is the difference between Extreme, and enthusiast!

 

A water chiller will automatically move you out of enthusiast league.

 

http://hwbot.org/submission/3773925_
http://hwbot.org/submission/3773923_

 

@Raiderman Are you on the latest driver from Nigreedia? Just wondering how it's running if you are. Nice cinebench score by the way!

 

I am doing hard tubing not soft and there's like 5 different sizes of tubing and I had read about previously certain brands having issues with colored coolants staining the tubing and causing issues (which is what I am initially going to try to use since I bought into this scam and have 2 bottles of the coolant already) so I wanted to make sure I was getting one that had a good reputation for not getting stained or ruined and right sizes for the type of fittings I want to go with. Also read of UV light destroying certain types of tubing but I don't know how true that is VS other things that could have caused issues with peoples tubing. Definitely don't want to go with flexible tubing as possible. I think hard looks so much better and cleaner and more reliable if you do it right so I'm trying to jump right into that.

 

Yes it is installed, but I havent had a chance to fully test it.

 

I don't have any experience with it, so I cannot provide any direction. I agree the hard tubing looks nicer, but it seems like extra work, added cost, and more opportunities for leaks. The only advantage I am aware of is aesthetics. All I know is what JayzTwoCents has in their videos, and acrylic seems more failure prone/brittle than PETG. When I get around to custom loop I will likely just use flexible tubing to simplify matters and minimize costs since it does nothing to improve performance. I kind of like the idea of using braided sleeve over flexible hose or colored tubing better than clear tubing with dye. I do like the colored coolant look, but also like the no-nonsense look of plain old black tubing with nylon braid sleeve on the outside. If I were going to do hard tubing, I think I would go with copper.

https://www.youtube.com/results?search_query=acrylic+tubing+vs+petg

 

Opinions on this case?
https://www.evga.com/products/product.aspx?pn=100-E1-1014-K0

 

Looks nice and has lots of features (I watched the video). Good that it supports dual 420MM radiators and a couple of 120MM radiators on the back/left side. I love how huge it is. I like almost everything about it. The only thing I wonder about is flexibility and how cramped the workspace might be. I could not tell that from watching the video. All of the closed panels makes it hard to tell without hands-on. Knowing that it is made by EVGA, I suspect it is an excellent product, with superior build quality.

 

Its got some great features, and is huge. The way it is oriented is somewhat like looking at an old Tube television. Would be kind of cool to have the glass side being the front of the case.

 

I really wanted to do Glass or Copper but as this is my first time doing a custom loop and jumping straight to PETG instead of flexible, I'm nervous enough as it is but I would love to use a more solid option like glass or copper in the future going forward if I continue to do this in future builds or ever re-do my loop but we'll see what happens with this one first. Dealing with this makes me nervous as much as it is, I can't imagine doing it with something even more difficult to bend or build, or something like those phase changes. I don't think I would ever be able to do that extreme type of cooling that some people do.

 

Actually, I think using copper would be the easiest of all hard tubing options. They have excellent bending and cutting tools that produce a professional result with almost no effort. As noted by Jay in his video, the outside dimensions on copper tubing are the same as PETG, so the same compression fittings are used for all hard tubing. The inside diameter is only relevant for flexible tubing with barb fittings. The obvious pros for copper are ease of use and durability (almost indestructible). The cons to using copper are cost and it doesn't offer as much bling to those that care about that.

 

FINALLY! Man, that seemed like it took forever.

 

I've never done watercooling, but knowing a bit about the flow of liquids, I would have thought flexible tubing could provide a higher flow rate - the gentle curves of flexible tubing vs the more tight & angular curves of hard tubing.

 

Yes, I think you are correct about the flow of fluids. Eliminating sharp bends improves flow. But, I don't know that it would be enough to result in any measurable change in temperatures. Working on a system with flexible tubing has to be a little easier. You can move a CPU or GPU out of the way without having to drain fluids, take apart more things and replace o-rings, etc. I like how the rigid tubing looks, but that extra work seems like too much of a pain in the ass to have to put up with. I take things apart too frequently for that to sound like fun to me. But, I think building it the first time would be fun and rewarding. I'd even give myself a little bit of extra slack in the flexible tubing to make maintenance and part swapping extra easy.

 

I hadn't considered the ease of maintenance aspect that you mentioned either, so that's another plus point. Regarding temperature differences between flexible vs rigid tubing, it would be interesting to see some testing showing if there's a difference - praps pump speed could be lowered in the flexible tubing option while maintaining the same temperatures as the hard tubing, which might result in less noise.

 

Ahh crap, the cooler beat the power link...lol. Congrats man, I know the feeling of "taking forever"!
@Robbo99999 water cooling is fun, but its also a pain in the ass, more so with open loops. If you have to do any maintenance on the pump or reservoir, you have to drain it, making a mess most of the time. Thats why I am looking at some no spill quick connects now. Maybe switching to distilled water also.

@Mr. Fox I never asked you if they had an update on your flow ring that addressed the issues you had with it?

 

The inside diameter actually does make a difference for all hard tubing. I believe there are 2 or 3 metric sizes and 2 or 3 of our sizes (lol) and the compression fittings for hard tubing need to match the ID and OD of it.

 

Watch the video. Hard tubing uses compression fittings. That has an o-ring on the outside. So the inside diameter makes no difference about fit. Smaller inside diameter may decrease flow rate and increase velocity, but inside diameter (according to Jay) does not matter except with barb fittings because nothing gets inserted into the tubing. It all goes on the outside. Maybe Brother @Johnksss can chime in on that since he has been more involved with this than I have been. Maybe Jay is wrong, but I don't think so looking at how the products are made.

---

Sweet score, Brother @Trafficante

On another note... back in business. So glad to be back on water again. Thanks for the PowerLink Brother @Raiderman.

 

Compression fittings have to match ID and OD or else they won't fit or worse..They pop off because of in correct fit. They follow the same lines as normal fittings.
And the whole system can be made to use one size. This is the general consensus. The cpu and the gpu in a custom loop are always going to be the slowest point. And you want to get a high pressure pump. You can actually run two of those. One is a backup if the other one fails. Along with twice the inside pressure. My loop is a 5/8 ID - 3/4 OD

 

Oh man, that does tidy it up quite a bit, doesnt it? Or at least get the wires from hanging down. Thats kind of why I was asking opinions on that case, because it has a reservoir clamp, and I would be able to use my power link...lol. Silly to think about spending that kind of money on a case to be able to use it
Nice score btw!

 

Yes, it does look neater. I like it.

Funny thing on that score. After I put it back together I could not figure out why my Physics score dropped from 20K to 15K. I checked and rechecked settings and was scratching my head. wPrime 32M seemed normal, but 15K 3DMark 11 Physics score sucked. Well, one of the sticks of RAM had to be removed to get that air cooler off. That one was not fully seated, so the BIOS was ignoring it. I noticed during POST the boot screen said I only had 16GB instead of 32GB. I reseated that RAM stick and instantly got back the missing 5K points. It was running in single channel mode.