Maximum PC

Fan-based cooling systems for rigs have have temperature controls that speed up the fans if the pertinent chip set gets hot enough, but how do liquid cooled rigs control theirs? Do the pumps simply work harder to move all the liquid through the radiator faster? Or how does that work exactly?

You can regulate the pump in the same way, if you like, but generally you want to design your loop to be able to absorb as many calories as your system can produce.

I tend to agree here. Liquid cooling is fairly quiet if you invest in quality parts, and so it is generally best to turn off and "smart fan" controls in the BIOS if your system relies on the motherboard fan headers. The loop should run at a constant speed to dissipate as much heat as possible, especially if the GPU's or chipset are involved in the process.

Thanks for the replies, they're much appreciated! The reason I was asking is because I'm still debating on getting a liquid cooled Digital Storm machine and they have this ESA (Enthusiast System Architecture) that claims to monitor just about every component on a system that is sensitive to temperature. I really like the idea, but not sure yet if it applies to water cooled components. I'm looking on their forums to see if I can find anything about it, but no love so far. I'll probably have to register on the site and ask about it. Will make another post if I find out anything further on this subject.

ESA is an industry standard for system monitoring created by a consortium of companies headed by nvidia in 2007...

http://en.wikipedia.org/wiki/Enthusiast ... chitecture

http://www.nvidia.com/docs/IO/46624/Tech_Brief_ESA.pdf

Thanks, vig1lant3! The second link gave me the information I was looking for. Now, to find out if the different devices are ESA compliant, etc.

ESA is not a corporate wide standard...it's an nVidia ONLY standard that nVidia made. nVidia doesn't do anything with Intel chipsets, so that's pretty much out the window unless you run AMD on an nVidia Chipset. Not a standard it seems if you need to run certain parts for certain situations. ESA is an nVidia chipset only program by any other terms and far from being a standard adopted by any official consortium.

ESA died at birth. Jip has it right in that using the BIOS to control the fan headers on your, or any, motherboard is the best method. If you don't like hearing the fans ramp up on CPU load, then just turn it off and let them be constant. Temps will rise, but who really cares? It's not hurting performance for the CPU to go from 25C to 45C...you've got room to run in.

So basically, I shouldn't even worry about the ESA stuff? It's just that it seemed a great idea. It was claimed to be open source so that all companies could participate. So most don't want to use it just because it was developed by Nvidia? That's pretty sad, all that corporate politicking and bickering. Makes me wonder how they have any time left to actually produce anything, lol.

Ah, when I say standard I'm not saying universally accepted. In order for ESA to work a great number of companies had to be on board to even achieve the interface necessary to make ESA work. Admittedly it is more of a standard set by nvidia and partners, but it IS a standard. Once upon a time Motorola PowerPC chips were the standard in Macs, things change. I would not say that ESA died at birth. It's more likely that it never caught on because not too long afterward Intel introduced DMI and locked nvidia out of the chipset party by refusing them a license. What power user would not want an ESA capable machine?! Unfortunately, right now it's Intel's world and we just live in it.

Personally, I am of the opinion that BIOS controls are always favorable to software controls offered from within the OS. Overclocking, fan control, etc. We'll see how EUFI stacks up when it's universally adopted, but I always prefer to set available system settings in the BIOS.

As far as thermal controls go, a liquid cooling loop performs better when it runs at a constant speed. I'm not talking about the difference between 25C and 45C, I'm talking about the difference between 50C and 80C with a max overclock. This is the overclocking section, do not forget that. It takes a significant amount of time for a pump and radiator to catch up to a thermal limit when enabling power saving features.

ESA is open source. Any third party MFG can offer support...

http://www.anandtech.com/show/2370/3

Hence forth, all posts conforming to the Jipstyle standard must:

a) be honest, without reservation,
b) use appropriate and correct language,
c) teach a lesson or explain a concept, and
d) optionally, be humourous to at least one person other than the author.

There. I just created a standard. It is open in the sense that anyone is allowed to apply the standard and advertise that fact but it is closed in the sense that no one other than me is allowed to alter the standard itself.

When Chumly points out that ESA is an nVidia standard, he is not saying that it isn't open .. it is. It is simply backed by nVidia. Why they would do so is up to them .. but chances are quite good that their standard leverages their hardware in a manner that helps them.

Another company may not implement the standard simply to avoid strengthening their competitor's brand .. even if the standard makes sense and improves performance. Hell, they may even release a competing standard that performs exactly the same function but ever so slightly differently.

And before anyone cries 'omg corporations and capitalism is horrible', please review KDE v. Gnome. Even FOSS has stupid shit like this.

Cute, but irrelevant. Yes, a standard is easily created, and easily ignored as UEFI has been thus far. That will be an upgrade that is shoved down our throats based on the fact that it supports newer hardware that BIOS does not., Sometimes that's what it takes. Even cutting edge tech guys can get comfortable and resist change...me too!

Crying corporate foul doesn't fly, or maybe I misunderstood your post. Compatibility is always the most important word in computer hardware. It's the only way to guarantee user adaptation, and it has nothing to do with strengthening or weakening a competitors brand. It has to do with building your own brand within an existing framework. ESA was a logical offering while nvidia was still in the chipset game. It was a way to market themselves to mobo MFG's. "Hey look at what we can do" sort of thing?

Now that Intel has set DMI as a standard, and nvidia has all but exited the chipset market, ESA will largely be an afterthought but that doesn't mean it's not functional or useful...and it will always be an OPEN source standard. It's just one that wasn't widely adopted.

I was not putting down capitalism or corporations in general, hell no! I was simply putting down their petty bickering over things like who helps initiate standards and what not. It's the posturing and the egotistical politicking that I despise, not the making of money. Making money is a virtue because that is how the free market innovates and creates all the wonderful tools and implements that improve our lives. But all the sniping and snarking by corporations at each other is just like watching children on a playground. It just makes me roll my eyes in amused disgust. An obvious example is watching that guy who's the head of Oracle, what's his name again?

My reply was largely in response to our MOD's statements. Sorry should have quoted, but I do understand what you're saying. Intel and nvidia have been at each others throats for years and it definitely affects the industry as a whole, as well as consumers. Nvidia's NG-ION is a perfect example. When Asus released the 1215N the specs looked great, but the GPU sits on a PCIe x1 lane because Intel refused nvidia a DMI license. What's the point of having a discreet video option with dedicated VRAM if it can only access 1 lane of a 16bit wide bus because Intel is trying to kill GPU computing?

That was my point.



If you thought I was crying corporate foul, you did misunderstand my post.



It does, actually. That is why AMD, Intel and others are constantly fighting legal battles with each other. That's why Intel is always trying to lock people out of their chipsets unless they pay royalties on each chipset sold.

Point 1. Then we agree

Point 2. Sometimes in my dotage I do tend to misunderstand things... ah well, such is life...

Point 3. Couldn't disagree more...

Actually the majority of Intel's legal wranglings are anti-trust issues based on unethical business practices. For example, offering vendors huge discounts on parts in exchange for capping their usage of competitors products is illegal. Yet regardless of lawsuits, there's actually a great deal of cooperation between AMD and Intel. Consider the cross-licensing deals between both companies to share architecture and technology. AMD wouldn't even exist with out an x86 license from Intel. Intel and nvidia just agreed to a huge cross licensing deal in January as well (regardless of their well chronicled distaste for one another)...

http://www.eetimes.com/electronics-news ... sing-deal-

Legal battles are simply the corporate way of marking territory, and setting legal precedents for what can and cannot be allowed to occur in a competitive capitalist environment.

I understand what you're trying to say about the chipset point, but it doesn't do anything for your argument. It's more of an example of Intel's borderline business practices than it is a point of trying to strengthen or weaken a competitors brand. It's just another source of revenue as far as Intel sees it and it's largely pointless anyway. All of the major components of a traditional chipset are integrated on-die with the CPU cores and cache these days. That's the whole reason VIA abandoned the chipset market, and nvidia is doing the same.

To be completely honest, there might be more cooperation between tech companies than there is in any other market sector. Mergers, acquisitions, and licensing deals happen every day...