Gadgetory

I'm testing it out of the box fry out of the box tangles so this is an Intel CPU versus an AMD cpu and the AMD CPU is measuring that at a twenty one point six which is a delta T over ambient of about zero and the Intel CPU is measuring at twenty one point six before getting to that this content is brought to you by the thermaltake flow RGB closed-loop liquid cooler which is a three hundred sixty millimeter radiator plus three 120 fans that are RGB illuminated if then we'll take it ringing fans at that this is a 4.5 done a stack pump which is one of the faster pumps you can learn more at the link in the description below so as you can see when we get rid of the heatsink that user last week was actually correct heat sinks are irrelevant when you get rid of the heatsink and you just measure them out of the box like a user would use them just like this they are actually completely identical in performance and I think if we were to take a CPU cooler just let that kind of sink the heat for a second okay that should be enough yep still 21.6 so they're completely irrelevant I apologize for last week I said something about how CPU coolers were actually very important to measuring temperature for CPUs and how you can't just disregard them but obviously this test proves if it's 21.6 now delta T over ambience of zero C and I take the cooler off still 21.6 well I guess there's no reason for the cooler industry to exist anymore so sorry sorry I knocked I didn't mean to put you out of business so that was our out of the box thermals test that was requested by popular demand last week you've now gotten it you're welcome welcome back to ask GN this is episode 61 I think we are going to be taking questions from the comment section and from our patreon discord you can leave quite in the comments below and I will dig through all of your dank memes as well as actual questions it gets harder every week to find the ladder and you can also post questions in our patreon discord if you go to patreon.com/scishow is so let's start with a question on streaming and gaming or just streaming in general this one was posted by otter wise in the patreon discord and otherwise asks you guys have a lot of content on gaming and streaming but what kind of hardware would be ideal for a dedicated streaming machine this is a really good question so most of our actually all of our benchmarks at this point have been just on a host system that games and streams to use your twitch simultaneously or records and clearly for those types of workloads really it's just more chords is more better for the host system because as you burden it with gaming and streaming tasks if your encoding and doing encoding tasks on the CPU it's just gonna get weighed down with something like a fork or for example so we know what to expect for the host system if you're doing a standalone streaming PC in some instances it might come out to be potentially a little bit cheaper the memory prices right now will really screw everything up but basically what I would do is just some r5 CPU that can handle the encode no problem at all you can buy the cheapest six core r5 you could spend five minutes overclocking it's a three point something gigahertz maybe three point eight or nine I think with 1600 work well for that and then really you could if you if you don't go too hard with the overclock you could just use the stock cooler to keep cost down further or throw some $25 cooler on there like the thermaltake contact 12 or the I don't really like the higher four to twelve the thermaltake contact twelve would be a better choice than the hyper 212 for that one but you throw something like that together that's a really cheap system get the cheapest memory you can get that's within reason and throw a light overclock on that as well a couple hundred megahertz or something slightly increase the mole flyer and you'd be good to go then you just need a capture card so I capture machine like that it doesn't need to be that powerful and you can either take parts you have lying around or just buy cheap r5 cheap board cheap memory and let it do its work the only thing to be careful of is if you go with be 350 gets some kind of cooling on the prm's if you do overclock it but otherwise that's what I would do and if you do that route for gaming and streaming in theory the secondary capture machine should resolve the lower frame time performance where you get spikes and frame time consistency during streaming from the host system that should be fixed by having an external machine and also depending on what you're doing it might be a little bit cheaper or around the same cost so it just comes down to do you have the space to do it and this is also something where you can salvage cases power supplies all basically everything because it's not like you're gonna be gaming with it it's just gonna sit there probably in a corner and then you can grab something like either synergy which they have they run ads with us right now you've seen the ads most of you you can run something like synergy so that you can just use one keyboard and mouse to interact with the host system that does the gaming and a streaming system or you could use that new Logitech keyboard and mouse that we saw at PAX which has Wireless and Bluetooth options and you set it up for Bluetooth on one machine and Wireless on the other and you just toggle between them that would be a software or a hardware solution to the problem so then you can only have one set of peripherals as well which would be pretty nice next question is from waffle who says I want to ask a question that pertains to new motherboards like the a C's crosshair hero six which come with a IO headers but don't know how to ask it it might go something like this should we buy should we be using a IO headers or not are there any performance benefits if not what should we be doing and should the consumers care if a motherboard has an AI ou have header what do you think I don't know if I answered this one before but those headers on motherboards are primarily there for things like the motors so the liquid cooler motors are generally something like 8 pull as opposed to 4-pole on most fans so depending on how they're set up and what motherboard it is I don't know what the crosshair does but they might have changed the math in there so that it reads the RPM accurately if you plug something like an eight pole engine twenty seven or an eight pole pump into some of the headers on some of the motherboards like we saw when we tested the engine 27 sometimes those headers will read it as a four pole motor and if it does what will happen is you'll get two times the RPM read out in BIOS as reality so it ends in 27 it's been at 2500 rpm but our BIOS was reporting 5000 which is enough to take off your hand so clearly there's there's a reason there just in terms of making sure you have an accurate reading of reality other than that it's there are some different options on some of the boards we've worked with where the i/o headers will have different profiles for their percent duty cycles for how they're the pump behaves under different conditions as opposed to a fan and I think that pretty much is it there might be one or two things I'm I'm not familiar with because I've only really used the i/o headers on the ACS boards lately so if you know of differences with the gigabyte board drama side boards let me know in the comments because I haven't really looked into that beyond what I've just said next question is from sir papa who says ask Gian thoughts on testing cases with water cooling alongside air cooling my thought is that cases get bigger as cases get bigger the air from the front fans loses its speed and therefore cooling performance do you think that an would be an equaliser between big and small cases therefore having purpose and testing or would the work would become through high the workload does become pretty high a couple of thins to this question it's a good one so we did publish the age 500 peat radiator placement guide after the review which basically does this it also tests the radiator in different mounting locations to see which one's best but first of all with regard to workload becoming too high that age 500 Peter 88 replacement video I'm almost definitely not going to make ROI on so you can see why we won't want to do that for every case review because we lose money on all time the next thing is in terms of thermal testing and cooling testing you run into challenges where some cases will support radiator mounts in different positions than others so it becomes a question of how do we standardize this do we want to always use it in the top position because there's pretty much always gonna be one or do we put it in the front position and then you have a lot more variables to deal with explaining to the viewers or readers when you're talking our performance because ideally you're not doing what we did for the Aged 500 P and testing the radiator in every slot because it's just not I mean it's just not sustainable as a business model but if you test only one position there's a chance of either the CPU is slightly worse than it could be or the GP is slightly worse than it could be because based on where the cooler is one of those components will cool better than the other normally so that's a consideration another big consideration is that the closed loop coolers are sort of big equalizers in that as you start using them to cool the CPU you lose some of the resolution on the differences between a really bad and a mid-range case so it starts you'd start making something like an Tek PA look somewhat acceptable just simply because you're using liquid to cool it and brute force so you do lose some test resolution on the differences between high-end or not even high ends but mid-range and low-end which is what we care about the most and then other than that I mean yeah it's basically it down to how do you standardize placement if a case doesn't support radiator mounting in one spot where the previous case did how do you reconcile that in your testing and your review and your notes because the performance is now going to be contingent more on the radiator location than on the case design itself and then you have concerns of obviously workload becoming too great to make money off of the content because you can't sustain an operation that requires money to to continue if you are spending more than you make and finally it's just the the fact that the CLC's do sometimes perform too well in to the point where you can't see as big of a problem with the truly bad cases as you would otherwise and a lot of the times the truly bad cases are ones that would be paired with an air cooler anyway so we wouldn't want to do just liquid coolers because you lose that resolution but doing both has its own challenges a great question though basically our solution to this is to try and do stuff like the aged 500 Peter 88 replacement guide on a per case basis with cases that people have clearly expressed a lot of interest in so not all of them but just stuff that we think we could actually get the traffic on like the H 500 P or something similar to that next question is from Jonas or Jonas Bjorklund who says follow-up question to the radiator positioning question you said that it is sub optimal to have the tubes going up and not down with a CLC I've heard this before but I really don't get why this is the case probably because of a lack of understanding of how the inlet and outlet to a radiator looks I watched your video on ail coolers but as I remember that particular detail wasn't covered so we Vance I've answered this one a few times over the last few years but the very short of it is those a few things but one of them if you haven't seen our TL DR how liquid coolers work watch that but as for liquid coolers and orientation basically this comes back to me saying mount with the tubes oriented down not up so if you're mounting the front of the case see on the tubes down here even if it's gonna look a little funny as opposed up high and that's just gravity it's because the air in the tank will float to the top and if the tubes are at the top they're gonna be where those air pockets can form and so what happens then is you can get some gurgling and the pump every now then when air is sucked through the tubes because it's sitting at the top of the tank and the radiator so as it gets sucked through the tubes it gets pulled through the pump and you got some bubbling noises that people don't tend to like theoretically there could be some impact to longevity there as well but I don't have I I mean we haven't done any longevity tests like that but theoretically there could be the main concern that was just noise and then after that longevity is a potential concern that's the main reason you do it that way though a big absolute no tell with radiators is don't ever take something like a hybrid card and mount it with the radiator in the bottom of the case with the tubes up because then you're fighting gravity to a point where as permeation occurs naturally over time and you have less liquid in the loop you're gonna have bigger air pockets those will sit between the the barbs the flanges and the tank and the radiator and that'll cause issues with pulling air through the loop and causing damage in fact it's bad enough that EVGA is hybrid installation guides that come in the manual say not to install it that way next question Alex Coleman says Steve I've been watching the thread Ripper review since I want to get it for work 3d rendering my question is why do you keep calling it - blender rendering I know you guys use blender to test performance but that's not the point I'm trying to make there are so many render engines out there it might confuse newcomers to the industry that it's only good for blender wouldn't it be better when referring to 3d rendering performance or video encoding performance to call it like that instead of blender rendering or premiere rendering you don't call streaming performance OBS or xsplit performance so as I addressed in this was from discord and discord streaming performance we don't call OBS or xsplit because they're irrelevant to the equation it's just it's they're just rapping h.264 encoding and then you apply some settings to it so all you need to know is those settings it kind of matters like we do actually say we use OBS for this but I don't say OBS encoding I say h.264 encoding as for the actual question we call it blender rendering or premiere rendering because we don't test the other applications so I can't in good faith just say rendering and think that it will apply equally to all applications because it doesn't premiere is a great example premiere if you've ever worked with it on a scale like we do or greater you know how much fun it can be to work with and how reliable it is and premiere has very specific and vigilia behaviors with rendering so for us to say a cpu renders videos X percent better than another CPU would be disingenuous if our only data point is Adobe Premiere because Adobe Premiere has its own unique and complex behaviors that even we don't fully understand so and most other people for that matter so you can't really just say you can't take Adobe Premiere data for rendering videos and say CPU a renders videos 5% faster than CPU B so if you're using Sony Vegas or a DaVinci Resolve I'd look up the second one or something like that they're gonna have way different performance than what we do with Premiere and also it varies based on video file and codec and things like that so AVC HD 1080p 60 on one with certain parameters for rendering in terms of bitrate warps and color correction scales things like that it will render differently and perform differently in Premiere than software B so it is not accurate just to just call it video rendering as for blender rendering we use cycles the cycles renderer in blender and we use specific settings defined in blender on the per project basis for what's how it's being rendered how many samples what size things like that and same thing there we're not Steen 3ds max we're not testing Maya so I don't know how they perform they could very well perform differently if they're not using the same renderer then performance could be different and we're not gonna say 3d rendering if all we're testing is blender so that's that's why I say blender rendering and premiere rendering because I tested blender and premiere and there are a lot of differences potentially between software that we can't account for without testing it as well so that should pretty much answer that one last question for this week actually it's not really a question cookie King gaming says on our knock to a liquid versus air coolers video that we just posted gamers Nexus stopped using charts these results are not tested or proven so basically this is fake content if you call stupid charts content disliked okay anyway thanks for watching you can leave your comments below for the next episode hopefully their questions I don't want to encourage things like the last comment mostly because I'm gonna have an aneurysm if I try to understand what it means he could subscribe for more as always patreon.com slash gamers and access to helps out directly or join the patreon community discord where some of these questions came from I'll see you all next time