kksnowbear

I said nothing that excludes considering other factors besides which GPU. In fact I said quite the opposite. And unless you claim to know more than MSI does about their own product, your statement about "any 4090" is inaccurate. I just cited an example where they specifically recommend a 1000w PSU. Not sure what you mean by "the recommended 850w", but that is not what some might recommend. Maybe you're OK with recommending a PSU that barely meets (or fails to meet at all) the manufacturers recommendations... My comment about 220 v 120 is solid, factual, and on point. It was also concerning why you and EightyDeuce were getting different measurements, not necessarily about choosing a PSU. Among other things, some people don't have 220v line to feed their PC. Yet it is still among the factors that matter. It is one factor of many, just as i said. Every statement I've made is 100% relevent to any discussion about PSUs. Can't help if you don't understand these things as well others might. Starts to look like you're just touchy that you're wrong about power supplies. Unfortunate.

What model/size backup did you buy, if I can ask? (I assume you mean "APC" as opposed to "APS") Will there be loads besides the one PC?

Yes, and my comments were taking EightyDeuce's grossly different numbers into account. Had they just been taken at face initially, it would be misleading. Software is still software, a meter is still...not (necessarily) software... My point was that doing this sort of thing, especially when making recommendations to those less knowledgeable, means it's important to account for differences. Otherwise it leads to misunderstandings (which can get out of control in a hurry). (Incidentally, to the question of whether it has to be 'super accurate'...well, all I can say is that if it's oversized by 1000 watts, it'll run just fine even if it's absurdly expensive, where if it's undersized - even by 50 watts - it *will* change the overall picture. It'll cause problems, and that's a fact.) Circling back a moment to the question of what power supply is appropriate to a 7800x/4090 system... Well...it depends. You'd want to know which 4090, as even different cards from the same manufacturer can be rated differently. For example, MSI recommends a 1000W power supply for their Suprim Liquid model, stating the minimum required is 850W, while the recommended PSU for a Gaming Trio 4090 is 850W. So that's 150W difference in the recommended PSU from the manufacturer, both cards being 4090s. Another (fairly important) thing to consider is that all switched-mode PSUs like those in our PCs operate most efficiently in the range of 50% load. So, in other words, you could *double* the calculated PSU output and it's not "too much" (except maybe the cost, that is). I am aware that people often throw money at (grossly) over-rated power supplies - it's actually one of my pet peeves, TBH. But the absolute fact is that sizing a PSU's output capacity on even twice the calculated load is not really absurd (just pricey!) Now, there are a lot of things to consider where trying to determine load - not the least of which being that the machine will seldom run at a maximum load (even when actually gaming), which affects determining what 50% is, of course...but it still applies that (especially if you're concerned with efficiency) you have to factor in how the equipment actually works.

You guys are measuring different things (at least somewhat) and it's affecting your results. Apples and oranges, if you will. For example - and this is just one example, mind you - it appears as if some1, being in Poland (unless I'm mistaken) is using 220v AC line (see HWinfo "UPS Input Voltage"). Meanwhile, EightyDeuce <sp>, I assume, is in the USA and (probably) therefore using 120v AC line. PC switched-mode PSUs are usually much more efficient at 220v than 120. That's one thing. Also, one of you is reading UPS load via software that we probably cannot fully account for its calculations (in any event this is not "wall power", and although it might be a realistic representation of what the PC PSU draws, it's still software-driven). The other appears to be using an inline meter device, which is (possibly...probably) measuring differently. Another factor is how these devices (your measuring devices) actually determine load; i.e. is it just VxA (which is actually watts, but not necessarily true AC power), or is it something more akin to RMS calculation, with power factor considered? Finally, although you both might have *similar* systems, you're not necessarily running the exact same tests on the exact same hardware. This could matter a fair amount as well. You both may well have (relatively) efficient power supplies...but the odds that they're exactly equal in efficiency, at the same moments in the same tests...well, the odds are not very high. Different measuring devices, different types of load, different conditions, different efficiency...could easily explain the delta in your measurements**. FWIW. Still a good idea to measure, don't get me wrong...just a lot of factors to consider even when you're using instruments. **(EDIT: I think we can all agree that there obviously must be some explanation for the two grossly different numbers EightyDeuce got when using different devices - which should itself be telling.)

Ahh, OK I wasn't sure if it mattered or not...I am also not necessarily into all the lights...but I do it for clients who are PS: Most of it can also be turned off, if you prefer

FWIW The Gigabyte board will have it's own ARGB control software, and so will the Corsair RGB RAM. They will almost certainly not work together - meaning if you want to control both, you will almost certainly wind up running two different software apps, one for each. Corsair RGB RAM sucks in this respect (and I definitely like Corsair products, with this one exception). I've done perhaps a dozen RGB machines over the past year or so, and in every case I've used other brands of RGB RAM - which has functioned perfectly *and* allowed me to set up RGB 'sync' using only the motherboard software, be it Asus, MSI or Gigabyte. The few times I tried using Corsiar (strictly because that's what the client asked for), it has **never** wanted to cooperate with anyone else's software. Look around online, you'll see. Now, if you intend to just let all the RGB run at default (meaning it cycles rainbow-style and doesn't sync between the various components), then it won't matter...but when I do them, I intend to have it controlled/sync'd with one utility...and the motherboard is common to all RGB lighting if done properly. You can control it by pushbutton, but if you do this, the RAM will not be controlled by the button (unless you use add-on heatspreaders with built-in RGB connections that are then wired to your controller...but in that case you wouldn't buy RGB RAM) Just some input, FWIW. Some examples of units I've done: null

Realistically speaking, high-quality power supplies can last well beyond 10 years. In fact, it's not uncommon these days to see new units with 10-year warranties from the manufacturer. That's not to say they absolutely cannot fail in less than 10 years (any more so than any other warranty, mind)...but it should show that at least the manufacturers have that level of confidence in their designs. I have at least one or two that are in that age range (of ~30 or so total in my shop), and still in service as 'bench testers' that I have absolute confidence in. I have an old Corsair TX750 that's gotta be going on 15 years now, maybe more, and it's a workhorse, routinely handling even multi-GPU loads I throw at it sometimes. If yours is holding up well, you could reasonably forestall the PSU expense for at least a year (or even several). As I described above, it's far more than enough for a 4070 (again, assuming it's a decent quality unit). Of course, it would mean that at some point you'd have to tear the whole thing down to replace the PSU...well, potentially, anyhow. It's at least worth considering, IMHO Incidentally, even though an 850 may seem overkill, I don't think it's a bad choice at all, should you decide to go that way. Having some 'head room' in a PSU is not a bad idea, and it's not unusual to 'over spec' to about twice what the total wattage of your system is (not at all necessary, but not a bad idea either). The idea is that switch-mode power supplies (like our PCs use) are most efficient at around 50% load. So, if you buy one that's rated to output twice what your system requires in watts, you'd be roughly running it at around 50%, at least the majority of the time while actually gaming. That said, your system with a 4070 would probably require around 500W. (That's an educated estimate, neither a specific quote nor total wild-a$$ guess). And it's not going to actually use that 500W most of the time. So if you have a high-quality PSU that's rated 850W, then you could reasonably expect it would be loaded in the range of 50% (~425w) most of the time. Also, it does leave considerable room for upgrades down the line. An 850 is more than most manufacturers recommend for systems using GPUs on the order of high-end 40-series cards (NOT counting the obscene requirements of the 4090s). Plus, we all know eventually there will be a 50-series, and the power requirements for the 4090s has been a sore spot for a lot of people...you could speculate that the high end 50-series might address the 'run-away' in terms of power required. (Anyone remember the 900 series Nvidia GPUs, in the wake of the 700s?) Best of luck to you.

There were two variants of the 2060; one had a Thermal Design Power (TDP) of 160 watts, the other was 185 watts. Although a 4070 has a higher TDP at 200W, it's not so much higher as to cause problems in your case. Technically, even one of the 6+2 connections is enough to power a GPU at 225w (considering that the PCIe slot itself provides 75w), which is more than the 4070 actually requires. Often, manufacturers use hardware that is slightly more capable than what's actually required, either because of being 'safe' with power requirements or because of manufacturing considerations - for example, I can tell you there's no way the 2060 you have now would ever use 300w, which is what one 6+2 and one 6 pin connector will provide (again, given the 75w from the slot itself). Unless you have a really bad quality PSU or it's really (really) old/weak, you'll be fine. The 4070 200w TDP is well within what a 2x 8pin adapter can provide (technically, that's 375w, which is actually beyond even the PCIe 3.0 specification, FWIW). If I'm following what you said, your PSU has three 6+2 connectors, and the 4070 adapter only requires two, so there are more than enough connectors. The GPU itself won't even know the difference, with the adapter. As a formally trained electronics tech with 40+ years experience, I'm not personally a fan of adapters TBH...but (as much as it kills my soul to say this lol) they exist for a reason. HTH - good luck

'Course, you could try to find someone who might be interested in making an offer on the 6900XT you'd be looking to get rid of, to take some of the sting outta that shiny new 4090. And, you know...keep yer first born. *ahem*...

I think it's generally accepted that a used 3090 is a better proposition than a new 4070Ti. (Actually I'm pretty sure there was a thread recently about this same choice). The price would be in the same range for either, the performance would be comparable...but the 3090 has 24G VRAM where the 4070Ti only has 12G. Of course, you'd want to consider carefully where you get used hardware from. Ask specifically about warranty, trade-in, and things like the history of the card (mining, and environment like smoking or pets). If the seller can't (or won't) discuss these factors, I'd look elsewhere. Also, check the seller's other clients...I wouldn't suggest buying from anyone without an established background selling components/systems. And yes, you can use a 4070Ti on a Z390 board, though it won't run at PCIe 4.0 x16, as the Z390 board will only support PCIe 3.0 - but it'll still get 16 lanes, depending on your setup, and a 4070Ti wouldn't saturate 16 PCIe 3.0 lanes anyhow (means it won't alter performance at all). Yes, you'll likely need to change power supplies, but if the GPU (either one) fits your chassis physically, then it should work. (This does assume an otherwise properly set up and configured Z390 system with only one GPU and no other add-in PCIe cards except *possibly* small x1 boards like sound cards.) Best advice I can give is to consult with a trained and experienced professional with demonstrated, verifiable credentials. Internet forums are full of 'experts' who give inaccurate and misleading advice in a lot of cases. Truth is there are a lot of details that would really benefit from further discussion, and I'd encourage that. BTW you mentioned where you live, but didn't say where. If you don't mind saying, that may help. Good luck.

Yeah, that's something I've looked at/used myself - but I often hesitate to give blanket recommendations for sites like that, for various reasons. One big reason is that some people just blindly accept what the site says without actually thinking about other factors that are part of the question. I've seen instances where these sites return results that are absolutely wrong, and someone who just wants a "TL/DR" answer isn't going to know that it's wrong, or put any effort into understanding why. Specifically, that site has you choose the type workload (which is a good thing) but only gives you a very limited set of choices: CPU heavy, GPU heavy, or "General". While it's good to have some range of choice, it's not going to take details into account which matter, such as the exact game (it varies, a lot) and VR (it accounts for higher resolution but not for other differences related to VR). It also appears the list of resolutions is limited. As with all these online tools, it requires a huge grain of salt approach. I know that, and I use that approach - but while anyone can follow a link, a three-field automated 'tool' (like that bottleneck website) just cannot replace experience, training, and good old-fashioned study 'legwork'. Another example is the numerous power supply calculators that are online. They work (to an extent) but they're pretty much doing some basic math to arrive at a conclusion, which doesn't consider a number of other factors that a trained and experienced professional will take into account. They can be downright wrong, and in many cases are driven by marketing (i.e. what a company wants to sell) instead of prudent expertise and advice. People buy 1200W power supplies without having any idea what they're doing; they spend way too much for something that isn't really necessary and/or fail to take into account some things that *are* a good idea. Anyhow, just be careful - and remember the best thing is *always* to consult an actual trained and experienced professional with demonstrated, verifiable credentials.

Ah...well that does make a difference then... With (very) limited knowledge of VR and zero experience lol I do understand it is effectively a huge jump in resolution. So that's definitely a factor. The more your resolution goes up, the more is demanded of the GPU, and after a certain point the GPU becomes a limiting factor. This may well be the case with VR and a 1080Ti, but (as with most things computer) 'it depends'. Different VR headsets have different resolutions, and (I think) settings with VR become a lot more crucial. A 1080Ti will do OK up to even 4k, but VR is a different ball 'o beans, as I'm sure you already know. I'm thinking this accounts for the "GPU limited" observation with the 1080Ti. So I suppose one could argue that a 4070Ti isn't a bad idea for that kind of demand. But that GPU is still overmatched (in a general sense) to the 8700k CPU, IMHO. Meaning: Add the 4070Ti to keep up with VR requirements, but I'd be concerned that even with the MT patch, the 8700 might leave a bit to be desired. In your case, the best advice would seem to be get a GPU as good as you can afford (since you'll need it anyhow, based on your expressed commitment to VR). If it's mismatched with the 8700k, so be it - again, given your commitment to VR, it could be that you'd need a better CPU anyway, in which case you can (and likely will) upgrade that later. I don't think it will be "unplayable" in the meantime, though that's often subjective itself. The 8700k isn't slouchy by any means (my son uses an 8086k, very similar)...just that, as circumstances demand more, well...everything's got to increase in performance to keep up. And, of course, when you add a much more powerful component at one point, then the 'weakest link' moves/becomes apparent somewhere else. Best of luck

At the risk of being lazy, please review my comments pasted in from a different thread on this very sub-forum (below). I think your situation is very similar - but you don't mention what resolution you run. It matters, a lot. The 1080Ti is more than enough for 1080p and is reasonable (if slightly overmatched) to the 8700k. Obviously pairing a 4070Ti with that same CPU is going to create a more 'imbalanced' system than you already have, being that the 4070Ti is far more capable than a 1080Ti. Of course it will still work, and of course you will get better performance than with the 1080ti. However a 4070Ti is going to be overmatched to a 8700k, and will thus not perform at it's best in such a pairing. This is all the more true if you also run at 1080p, and is particularly true of a sim like DCS which seems to have issues with CPU utilization to begin (though the latter appears to be somewhat improved upon with the recent "MT Patch"). That said, I think even with the MT patch, a 4070Ti is going to be mismatched to a 8700k. It seems to me people are grossly overestimating exactly how much this MT patch is actually capable of. Essentially what you'd be doing is paying a lot more than the performance increase you're going to see will be worth. For some reason - as below, perhaps misguided - people seem to think that you can just continually add more and more powerful GPUs and the performance increase (usually described as "more FPS") will be in direct proportion to cost. This simply isn't true...the more mismatched the CPU and GPU are, the more one of the two is going to limit the other. Kindly keep in mind the foregoing, in its entirety, is strictly IMO (even though it's based on over four decades of professional experience, years of formal training, and first-hand day-to-day commissioned work with 15-20 gaming systems every year for many years now - including many for people here and on the other major flight sim forums, in addition to those done for friends and family).

Have to know more to be certain, but I'd be willing to bet your GPU is so underutilized because it's being 'starved' to death by the CPU. The 3060Ti, while not 'top-of-the-line', is a very capable card. In a (perhaps misguided) effort to get "more FPS" people often upgrade GPUs way beyond what a CPU is capable of keeping up with. It's not as simple as putting in a better GPU to increase FPS...the CPU/GPU act as a pair, and need to be 'matched'. I believe it's likely that a 7700 is fairly 'overmatched' by a 3060Ti. The CPU is probably working his rear end off, and still far from able to 'feed' data fast enough to the GPU...so the 3060Ti is sitting there bored. It sounds like what a lot of people refer to as CPU "bottlenecked" (though I really dislike that term because it implies other things that also aren't really accurate). I prefer 'mismatched' because I think it more accurately describes what's really taking place. Mind you, that's a guess and it really could be way off, as it depends a lot on what else is going on with the system as a whole...but, assuming everything else is in reasonable order, I'd say your CPU and GPU are pretty horribly mismatched On the original question: IMHO once you get up into the higher-end 30 series GPUs or beyond, the gain in performance at 1080p starts to diminish. Reviews at the time of the 30-series debut almost universally noted this...and concluded that newer GPUs really perform their best at higher resolutions. So that would all suggest you'd be better going 1440p.

I am simply responding to those who directed comments toward mine (and apparently, some still don't get it). Genuinely sorry for any imposition - but, to be fair, no one's forcing anyone to read any of this

Oh, and also (copied directly from the MSI article itself): TOP Mount The problem of air accumulation will be improved in the top mounting configuration. The air in the AIO system will accumulate on the top of the radiator, which basically does not affect the coolant flow in the system and does not reduce the thermal efficiency. This configuration is the most recommended in terms of thermal performance. However, some cases may not have enough space to install a 360mm or 280mm radiator on the top. Sounds a lot like...exactly what I've said all along. Top mount is the best performance and doesn't suffer air void problems, but the reason it can't always be used is because of mechanical constraints (a case that wasn't built for it). So get a better case...or accept something that even they acknowledge is not the "most recommended" (lol) Even though I don't agree with their approach/methods, they apparently agree with me

On the MSI article: Among other things, how hard were the GPU fans running during these tests? Factually, GPUs control their own temps (within limits) by altering fan RPMs and/or clock speeds/voltages. I can also create custom fan curves in software that will cause a GPU to maintain a certain temp across a *range* of internal temps, simply by running the fans differently. So, I could show the same GPU temps in a chassis that was warmer inside as compared to another. Point is since GPUs have independent thermostatic fan controls, GPU temp alone is not necessarily a reliable indicator of air temps inside the case, but that's all the "test" shows besides CPU temp. What about temps other than CPU or GPU? Motherboard VRMs? Chipset? Internal drives (especially M2, which are known to perform better when kept cooler)? Case fan data (How many? Size, RPMs, CFMs, pressure?). Controlled by what? The software used is named, but not settings, time duration etc. Therefore it's impossible to determine actual load conditions, especially as compared to "real world". Was the case sealed? Were any openings dampened? We don't know any of this because it's not disclosed. The fact that key data is missing means this "test" isn't proof of...anything, really. Well...I suppose it *does* prove that MSI has convinced at least one person that blowing heated air into a case means better cooling. After all, there's a reason the shell game still exists, and still works on some people. Incidentally, the case that they used for this test? It's around $400 far as I can see. Kinda doubtful their results translate precisely into other case types/configurations, especially depending on ventilation and fans. And the case's particulars factor into this sort of test significantly. And the CPU temps? Well, of course they are lower, you're blowing outside air directly at the radiator that's cooling the CPU. But as I already discussed, this just means that the air heated by the CPU is being re-introduced to the inside of the chassis, making it warmer than if it weren't heated. A front mounted radiator *might* be "better" for the CPU...but everything else in the chassis is warmer, without doubt. And that's just stuff I saw, at a glance, that indicates this is not an objective test. It's obviously designed to specifically show what they want it to show. I simply don't have time to waste on rigged "tests" that are clearly lacking the level of detail required to be any sort of meaningful. As it is, they're showing what they want you to see, while omitting crucial information - in other words, shell game.

A. (As regards all your comments about custom cooling loops): The OP talked about cooling his CPU with the EK unit, but here's an idea: Look at his pictures. He's got a 'hybrid' GPU and it appears the rad for that is in front (the EK is clearly visible up top). I'm not talking about custom loops, so again, you may well have been better off (and spent less) with an air cooler and leaving the front open for as much (unheated) fresh air intake as possible. And that would absolutely, undeniably mean cooler temps inside for things like the VRMs, chipset, M2 drives (which slow down if they get hot) etc...you're just not getting away from the reality that more cool air intake is better overall than just making a difference in one component. And an air cooler for the CPU - with an otherwise proper case and intake arrangement - could have cooled just as well as the liquid unit, without necessarily dumping heat on everything else. B. Yes, I am absolutely arguing that the point is (and should *always* be) to manage overall temps - because if done properly, this will almost invariably allow better individual component cooling (but without sacrificing other components as you acknowledge you are). Pump enough cool (ambient) air into a case to create positive pressure, you're not going to have throttling unless you've jacked something up too high. Fact is, if I displace heated air from inside the case with an adequate volume of cooler air from outside (which is how positive pressure works), the net is a reduction in temps - and that applies to *everything*. C. Far as the summer goes: As a scientist, you (should) know better than trying to introduce your "mood" into this. We're talking about system performance. So, mood aside, your system is going to run hotter in the summer, thus the situation with exhausting heated air from a radiator into your case *will* become worse. The coolant *cannot* remain at 32c and still remove more heat from your system - that's not possible. If there's more heat (summer) and the coolant temp stays at 32, it's not exchanging heat properly. Unless ambient stays the same, the intake temp will be higher, and your case will be warmer inside. This is what you seem to be overlooking: You're increasing the temp inside your case by dumping heated air into it from the radiator. No matter how small, it's a net increase - and that's more thermal load. No matter how slight, it's still (at least) 10c warmer than it would've been if you were jacking the front panel with cooler (unheated) ambient air, period. By the way, "It's too cold right now" is easily overcome - that is, if you're really concerned about properly testing something. D. My experience with maintenance of computerized systems means I've worked with this stuff 40 years plus. Yes, a lot in data centers, but at least half that is not. Over that time, I've built or overseen building and maintenance of hundreds (if not *thousands*) of PCs. I currently build around 15-20 every year, and most of these are gaming machines. This includes several builds and consults for members of this forum and others like it, who play one or more of the popular flight sims (DCS, IL2, MSFS, etc). At least a few of these people are on their second build that I'm doing. The systems I've built (and support) in just the past 4 years alone are in use by at least 30 gamers and around 6 or 8 streamers, maybe more. I'm sitting in a room right now with no less than 17 working gaming PCs, half a dozen with liquid cooling. More in a second room just across the hall, still others in a third room, and yup, even more in yet a fourth room. I really dislike having to break out the whole CV, as it were, but I *only* mention all this because you saw fit to question my "theory". I'm telling you I do this, in reality, every day - and have for several decades. You want to question my use of "theory", but I'm citing facts. And every bit of what I'm telling you is verifiable, in case you want to challenge that. This isn't just theory. And again, I'd be happy to compare respective, relevant experience any time. Incidentally, the background outlined above also includes the best formal electronics training in the US Navy, as well as first hand electronics repair and assembly work I still do every day, for local companies, one of which deals with (you'll love this) thermal gas systems. So while my experience and knowledge certainly includes gaming PCs, I would have to say it goes *well* beyond that of the typical casual gamer. I understand more than probably most here do about VRMs, heat in electronics, and so forth. I'm not some self-appointed internet "expert" who drives a bulldozer or sells insurance for a living by day. (Not that those aren't respectable professions, mind you, they're just not...*this*). I am also currently using a 5800X3d, so there's not much need to tell me what's involved in cooling one. E. The article from MSI: This warrants a separate post. However, I did say "almost" no empirical data, and I already saw the chart, so no need to throw that up. Most of the chart doesn't even apply to this discussion, and (as I will illustrate in my next post) the parts that do apply are simply more about marketing than any real testing - as I already explained, but will further discuss in a moment. In short, it's little more than marketing-based, carnival shell game BS...but let's actually qualify that description...

That link from MSI shows an article, not test results. There's almost no empirical data, and zero mention of methodology or any other factor that is *crucial* to proper testing. There is no way at all to verify the findings of their "tests" without actual published data and procedures. The simple fact is MSI sells liquid coolers. They are motivated by sales. If they acknowledge that it's a bad idea to have a liquid radiator exhausting into a case, they're discouraging sales to people with cases that are ill-suited to radiators in the first place. I don't know how many cases in the world there are that can only house radiators in the front, but for every one of them, that's a potential lost sale unless the owner is convinced that it's not a problem. So I'm the marketing director over at MSI, I'm gonna tell my staff to get busy showing it's a good idea - without really publishing anything other than an article saying "Trust us...we tested it" This is a problem with cases people already have, and either don't want to or cannot replace, and that's all. It absolutely does not prove it's a good idea to exhaust a radiator into a chassis - it's biased marketing, plain and simple. And, as I've already explained: My points are not theory, they are practice, backed by formal training and over 40 years' experience in professionally-compensated computerized systems maintenance. I don't need to cite articles from others. If you want to compare factual data on our own actual first-hand real world experiences, I welcome the opportunity. It's pretty, but there's not enough space... well, anywhere, really. So you ended up putting the radiator up front because of your case. Fairly typical; case manufacturers consider quite a few factors when designing cases, but optimal liquid cooling isn't necessarily one of them. I'd have mounted the radiator on top, but there's no room up there. Good; we agree the mounting was driven by available space, as opposed to an arrangement for optimal cooling (which you acknowledge you'd have done if there were room). One problem with the idea that you're "not blowing hot air into the case" based on this example is that you're cooling the CPU. Different story if a GPU rad is installed as an intake (that's what the OP's doing I believe), and actually cooling a significant load. I can't speak about CPUZ as a stress test, I don't use that. Run Prime 95 for a few hours and then look at the CPU temp, not the coolant temp. If I fail to properly install a cooler (air or liquid) then the cooler/coolant itself won't get warm, but the device being cooled will be much warmer. It doesn't really matter what temp the coolant is, the goal is a lower temp at the CPU or GPU, not the liquid in the radiator. By the way, if your ambient air is 69F and the coolant is 32c...then you most assuredly *are* blowing heated air back into the case. I don't know about you, but 90F(32c) is pretty warm for most people. In our context, if the air being introduced isn't the same or near ambient, it's not really "cool". Result is that the inside of your case is warmer than if it had a direct intake of outside ambient air (69F in your example). The entire system will absolutely cool better if the intake air is cooler, and that's just physics. Putting a radiator in the front panel inherently means the typical primary path for introducing cool air is now occupied by something that's heating the air. Even in your case, there's a ~21F delta in ambient/intake vs. exhaust/internal, and it's strictly because there's a radiator on the front panel which is heating the main source of intake. You're going to be hard pressed to prove that heated air at the intake doesn't cause warmer temps inside. At that point, you're not arguing with me - you're arguing with the laws of physics and thermal dynamics. And yes, it's going to be warmer during the summer. Can't really count cooling in a room that's 69F ambient as a good test; I doubt it's typical for many situations especially during the summer. Best practice testing assumes worst case. When I 'burn in' an overclock, I intentionally make the room 80F (even using a space heater if need be) so that I'm testing what is a likely scenario (upstairs bedroom, game-crazed teenager with door closed), or at least something that is slightly beyond what can be reasonably expected - thus testing will exceed the need. Tests that don't consider likely extremes are sort of pointless. You also have to factor in a sort of thermal dynamics, whereby introducing warmer air means the components inside will in turn run warmer, which will of itself (and due to physics) mean the air inside the case is now warmer too, and so on...in a closed box, with air intake that's not enough volume and/or not low enough a temp, the (actual component) temps *will* eventually increase beyond what they'd have been with cooler air coming in - again, just physics. Your CPU might have actually been cooler with a good air cooler, leaving the front panel for good intake fans - and it might've cost less, too. (Also everything else, including the motherboard VRMs, would be cooler too - something a liquid cooler simply cannot do). Your situation is like many: Cooling system performance is impacted due to mechanical constraints. Nothing personal, it is what it is, but let's just be honest: What makes it acceptable to you is that it fits in your chassis - but that's at the expense of blowing heated air into your system. I get it, it's a compromise - but let's just call it what it is, and not act as if the laws of thermal dynamics don't apply.

On the front rad, one thing I wonder is, how it that different from having an internally exhausting sir cooled GPU? Good question, one that should be discussed...however, at least initially, consider that I don't believe I've said anything about internal air cooled GPUs. The discussion has been about front mounting a liquid radiator as opposed to top/rear mounting (as you can see, both assume liquid cooling). Broadly and generally, concerning the woes associated with mounting radiators, I'll say this: It's not smart (to me) to spend a ton of money on a component, and more money still on a cooling system, and then diminish the performance of said arrangement by shoe-horning it all into a case that's a poor fit, just because that's what you have. Ya spent all that money on GPUs, CPUs, and cooling, get a case that aids in the goal (cooling) instead of opposing it (introducing heated air). Unfortunate if you'd have to dismantle anything, but it's either that, or you're intentionally choosing an arrangement that even common sense tells you cannot be optimal. Not directed at any individual, but it seems apparent here that some people will sacrifice "optimal" to save a little work or cost - while introducing the very thing they're paying to get rid of. And that's fine, if it's your choice. But call it what it is. My only problem is when someone tries to toss the laws of thermal dynamics out the window by acting like there's no difference. Fact and physics say otherwise.

On the "This is fine", as I recall, Jay was referencing the "this is fine" meme, not saying that was the right way to mount things. Well, I would certainly stand corrected if that's the case - and admittedly, I didn't watch it, only part(s) of the first video. I appreciate the correction. To be fair and accurate, though, I did refer to the picture (here) and it appears wrong prima facie. I also said I wasn't sure why it said that. Some people just take the wrong thing from somewhere it wasn't intended. On that point, this reinforces my contention - here and in other threads *ahem* - that you really have to be careful about the impression you give others. *Some* might know better but others might not. Now if you'll allow, I'd be happy to address your other questions separately. EDIT: Sorry, I watched about 4 minutes of the JayZ video and just couldn't take any more. (Don't get me started). I will say that it's difficult to imagine how very *little* a person would have to know for some of these videos to actually be useful. Seriously. Also, I'll have to stand on my premise about the picture. It remains that if I saw it, others can. And of itself it depicts an inaccurate concept. (I can literally feel it being posted all over the internet by people claiming it's proof that "This is fine", even if maybe some of us know better).

Actually it wasn't difficult or even expensive to do in my situation, and in fact I could *very* easily add a second radiator, both exhausting air (in fact I'm considering it). All while maintaining positive pressure, because that's also (in my book) a "near-imperative". And again, my current radiator both exhausts heated air and draws cooler, outside air. Just had to figure out how to set it up, way back in the planning stages of the build (which is unfortunately where the real failure is, most of the time). Part of the problem is people just accepting that it's too difficult or not worth it. They spend tons of money liquid cooling this and that, but then shrug at the notion of re-introducing the very heat they're spending time, money, and effort to deal with. They also fail to think through things, rush, and make mistakes. They heave money at the biggest, most expensive stuff available, even though that doesn't mean getting the best solution... How many people here actually have two radiators anyway? (And if they had that kind of resources, then they *really* should have no problem making sure it's done optimally). To be clear, "theory" in my case is practiced. I do 15-20 builds a year, in addition to friends and family, and have performed or overseen hundreds (maybe *thousands*) of PC builds over my career. This is not just theory. Far as testing goes, I don't know what data you're referring to, but I'd bet money there was a bit of a foregone conclusion in play, at least in some measure. The old "Well, we can't figure out how to solve it effectively, so we'll just decide it's not really that much of a problem". (See above re: expending money and effort, only to re-introduce the very thing you paid to overcome.) I'd prefer to see the testing data and results before concluding it's not that much of a problem, especially if it's not hard to overcome. We can agree to disagree about the location of the lines going to the radiator - again, I look at this as being declared acceptable due to lack of a better idea. No need to get upset at me just because I don't suffer that lack of a better idea. I've seen some references indicating the voids in mass-produced AIOs can go on 10%, and if you look at total volume of the pump, rad, and lines, I think you'd have to admit that it takes much less than 10% void to expose some part of the lines if they're mounted at the top of the system. So even if it's much less than 10% air, it's potentially enough that the lines aren't completely submerged internally. What's the basis for "extremely minimal risk"? (BTW how much is 'extremely minimal', anyway?) Your risk of being in a plane crash might be extremely minimal - but if it happens, your odds of walking off are "extremely bad". Everyone can do whatever they want, but let's not act as if there's no better way, simply because we couldn't think of it. As I said, I don't suggest it's worth it. People here spend thousands on CPUs and GPUs...but then we're going to entrust cooling to a system that's not optimal because some guy on Reddit said the effect is negligible? Hmm...yeah, that makes sense My advice is to find the better way. And, if it's approached properly, it's not necessarily hard or expensive. The part that amazes me is that people go bananas over tiny variation in frame rates (as one example) but then have no compunction whatsoever about dumping 100+F air back into their (sealed) case, while agonizing endlessly over 5c difference in temps. I said earlier I didn't want to get too far off the subject. Next thing you know, someone will be complaining about "derailing the thread" (which usually means someone doesn't like being proven wrong). What I'm saying pretty much speaks for itself - let the reader decide. The comments shown below are in the videos posted above, just so happens that the points illustrated show exactly what I'm saying. I don't always agree with GN, but the "proof of the pudding", as they say... you can clearly see the lines sucking air bubbles (and as the bottom pic shows, you have to factor in that liquid level in your sealed AIO could be even lower than these pics show)

Please see my edit, concerning top mount. The radiator *can* be mounted on top; I haven't said it shouldn't be. Sorry if this wasn't clear. EDIT: I just re-read the part where you said "on top the lines are above the pump"...kindly note I'm not saying the *lines* cannot be above the pump. I'm saying the point at which the lines connect to the rad should never be at the top of the (AIO) system. If you mount the rad at the top of the case (horizontally with respect to the orientation of a 'normal' desktop tower case), then the lines enter at the bottom of the rad - meaning there's more than enough liquid above that point to ensure the lines are always submerged. Any air void in the rad will naturally occupy the uppermost level of the rad itself, thus not affecting liquid pumping at all.

First of all, don't put a radiator on the front of the chassis. This avoids both the problem of where the lines go into the rad at, and heat being exhausted into the system. As for the bottom - I'm not sure if I understand correctly. You mean mounting the radiator at the bottom of the chassis (flat, horizontally) with intake through the bottom of the chassis itself? If so, it means that while you are getting cooler air directed at the rad, you're also still exhausting hot air into the system. Again - and this is my opinion but it's based on sound fact and physics, combined with decades of experience and formal training - radiators belong at the top or rear of a case. EDIT: I just re-read the part where you said "on top the lines are above the pump"...kindly note I'm not saying the *lines* cannot be above the pump. I'm saying the point at which the lines connect to the rad should never be at the top of the (AIO) system. If you mount the rad at the top of the case (horizontally with respect to the orientation of a 'normal' desktop tower case), then the lines enter at the bottom of the rad - meaning there's more than enough liquid above that point to ensure the lines are always submerged. Any air void in the rad will naturally occupy the uppermost level of the rad itself, thus not affecting liquid pumping at all.

Putting the radiator at the front will almost certainly mean exhausting hot air into the system, from the point which should be the primary intake for cool outside air. Not smart. Don't put it there, and preclude the problem. Can you provide a reference for how the radiators are "typically pretty well filled" (something authoritative, like a link from a manufacturer)? My comment has nothing to do with leakage. Anything that heats will expand. There has to be room for expansion, even in a sealed system. And yes, there is some area left empty because of manufacturing as well. As I said, the lines at the radiator *might* be submerged...I didn't say they are or are not, I said I wouldn't suggest that it's worth taking the chance. Radiators belong up top or in the back, exhausting heated air out of the system. Else you're defeating the purpose of (at least some part of) a cooling system. For the record, my own system has a 240mm radiator that exhausts at the top of the chassis, while also getting outside air (documented 20c cooler in load conditions) as it's intake. I'll leave it to the reader to figure that out, but that's how it is. The point is that just because manufacturers of cases and AIO coolers are constrained by a number of factors as to how they suggest their products are used, doesn't prevent the system builder from doing a better job with design and execution. One good step to start off with would be "Don't dump hot air back into the system." lol