Gadgetory

Vega 64 may consume more power than a GTX 1080 but how much does that impact room-temperature that's what we wanted to know and we eventually expanded that concept to include how much a 940 watt mining machine increases room temperature and how much a 600 watt machine and so on we were able to effectively replace any need of a heater for the past week as well and right when it started get colder so in this test we're looking at the room ambient impact of various PC builds and wattages and whether or not you can replace your heater with your computer before we get into that this content is brought to you by the thermal take 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 rain 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 although this test is a little bit harder to control than our usual tests to quote friend of the site VSG it's the right balance of science flaw and madness to make for an entertaining educational video and that's really what we were targeting something that's kind of real-world very interesting and can be somewhat controlled if not perfectly this test started off just trying to look at actually 1070 and Vega 56 but there wasn't enough of a power load there to even really begin the test so we moved it up to Vega 64 and a 1080 and then we started ramping into the higher wattage units and ultimately this type of test is inherently very flawed but we'll lay out everything we did for testing and then hopefully establish how it can apply to you or where it applies the most because obviously when you're testing room scale temperatures you're dealing with things like insulation age of the house or building the external temperature what you're doing with AC or HVAC during the tests and if you're going to do things like block the door with a wet towel or anything like that it's preventing heat from escaping so there's a lot of places where there are decision points and we chose to stick it to mostly realism so we're controlling to a point and then we're trying to implement some more realistic scenarios for the testing in and for these tests we spoke with a thermal engineer in the industry and with vsg of thermal bench and that helps us plan but we did have to make a couple of decisions on our own there are a lot of variables here we're not going to attempt to control for all of them but we discussed the possibility of for example testing in a thermal chamber but the Chamber's we have access to and have used in the past are small just big enough for me to stand inside of it so it would have heated up fast and wouldn't really relay real-world room temperature we ultimately decided to test in an actual bedroom in an actual house and this means we have four walls insulation a door a small gap under the door a ceiling fan and all of the other things you'd have in a room for room setup we configured the ceiling fan to spin out of medium speed and left the 60 watt ceiling light on for the entire test as this will generate a minor amount of additional heat if insignificant the idea is that a user who's trying to combat rising room temperature would probably have the fan on to some degree and you'll probably have a light on while you're in there we considered blocking the underside of the door with a wet towel to help isolate the room and the testing but ultimately decided against it there's no real point in controlling for that type of variable when there are inherently so many others with this type of test and in a real world scenario no user is going to do that so we left that one out we also did control for as much as we could the external weather factors though we ran all tests on different days giving the room time to cool down between each test so we waited for similar or identical as far as they can be weather conditions outside to run the tests and that meant that the tests took a couple of weeks to run because we're waiting for the right conditions and also allowing the room to cool down rerunning tests as we found flaws with the methods and improving them and things like that so there's a lot of data that was used to improve testing for the final data set we also had to decide what to do with AC so we considered disabling it entirely but ultimately decided to configure AC to regulate a temperature of 70 degrees Fahrenheit or 21 degrees Celsius and here's the next challenge the thermostat is located in the hallway outside of the room iemon now that you're watching this on the builds were on this table and the hallway is not too far away it's a couple feet away so the thermostats there which means that the temperature sensor is also there and AC won't trip until that sensor reaches whatever the value is 70 F in this case so the temperature in this room is somewhat isolated but as more heat sort of slowly seeps out under the underside of the door and gets out there it will trigger AC now this was by design and were pointed it out for a few reasons we placed a thermocouple down the hall in one of the colder rooms as the coldest room in the house there's another one down the hall placed a thermocouple there that's our control room and then we placed two in here one was one foot away from the computer where you would be sitting the other one was in the center of the room about six feet away from the computer this room is sized that 15 foot 4 by 9 foot 9 we're measuring temperature every second or so over a period of several hours to allow steady states who be achieved if possible some tests went as long as seven hours and we're still not at steady-state at the end but had to call it there so just to give an idea thermocouples again positioned in two places in this room and the point of all of this is to highlight that this is an imperfect test but it's an academic experiment it's it gives us enough good data that we can still figure out how much a computer would heat up a room given these parameters that I've just spent the last few minutes laying out so the point of saying all that is so you can figure out if it relates to you and either way we've controlled for our environment as much as possible so it gives you a good idea now if you're sitting in a warehouse or something like that obviously your mileage will vary tremendously if you're in a warehouse but I don't have to really worry about this at all because it's just gonna spread out but that's where we're starting so your mileage will vary depending on your AC policy things like that but seemingly small things like keeping the door closed which we did can have a big impact on tests depending on how you do it at home so your results would it change based on that as well opening the door for instance would help a huge amount maybe for some reason you want to keep it closed maybe you have pets who jump on your test bench for example so those are our variables and our controls let's get into the data so to recap our AC power see quickly we had the AC on during the 900 watt mining machine test it came on occasionally and that's because if we allowed the room ambient to reach 40 50 C maybe not 50 but 40 C that's obviously becoming a fire hazard so we've got some limitations there the AC was basically off and all the other tests because the room didn't get hot enough to see peed out into the hallway and trigger AC so it was on for 900 watts and basically off for the other two we're starting with the most extreme experiment since that's the most fun one to look at this is from our 940 watt temporary mining machine which used a mixture of four GT X 1080 and 1080i GPUs we had three tea eyes and one 1080 all configured to 70% power limits with 100 to 150 megahertz core overclocks and a 250 megahertz under clock the fans were also blasting and we had a few fans positioned on the table to reduce heat in key locations and the idea was to avoid burning down the video set while all of these were going though we also had the CPU burning for crypto night for which we used a 1950 X under the lick tech three sixty millimeter cooler power sat at around 940 watts at the wall all said and done and for results here's what we got room ambient for this test started at around 20 degrees Celsius with our control room at around 20 point four degrees at one foot of distance the room temperature increases by five degrees within a half-hour and keeping the door closed this of course means that it's going to be a bit more accelerated than if you opened the door the temperatures at the 1 hour mark climbed to twenty-seven point five degrees for those using Fahrenheit that'd be a climb of about 69 Fahrenheit to about eighty one point five Fahrenheit in the span of an hour our AC kicks in at this point and regulates the control room and hallway down to the target temperature where we briefly saw temperature drops in this test room of about 1.5 degrees between AC cycles we managed to climb all the way up to about thirty degrees Celsius following three hours of testing and that's while AC was battling the temperature rise on occasion in Fahrenheit that puts us at 86 degrees effectively eliminating the need for heating in the room during winter actually it sort of starts necessitating AC middle of the room temperatures aren't quite as bad as being in the combat seat with a maximum temperature of about twenty eight point one degrees as for power consumption the charts about to get kind of messy this Green Line mapped to the right axis represents power consumption during the test this machine was running nice hash back before it was hacked and that's what loaded the rig fairly evenly for the testing period if you had a similar room setup two hours a bedroom attached to HVAC and with its door closed you're looking at high room temperatures from something like a mighty machine or a rendering machine or anything that generates a 900 watt load we went five degrees over control and just half an hour and reached nine degrees Celsius over in about three hours and just to establish a point here it doesn't matter how you generate the heat it's still heat it's still wattage so if you are pulling 900 watts through some insane gaming machine we're pulling 900 watts through a rendering machine or 900 watts through a mining machine it's all 900 watts doesn't matter how you create the power in this instance as long as it's created in a steady way it will heat the room up in the same fashion so it doesn't matter what you're using to create the temperature basically or create the heat load and the power load more appropriately stated this next test is closer to a high-end gaming or production machine and consumes about 600 watts of power looking at this fire strike chart from our gtx 1070 TI review we can see that a crossfire configuration was pulling about 460 watts for the system with our power modded vega 56 at 447 watts for the system a 600 watt machine would be similar to running to 1080 T is at Full Tilt's with a lighter load on the CPU or running a an overclocked I 9 CPU with a high end GPU this chart from our 79 80 X a review helps reinforce that it's not hard to make CPUs draw 500 watts alone if you're running in an h EDT class part these CPU power charts were measured at the rails so that's just the CPU for those knowing that our next test is at 600 watts so it's a bit more achievable but still on the high side this machine progresses somewhat linearly as the temperature never reached a point of tripping the AC from the hallway sensor the 900 watt machine had just enough heat seeping into the hallway that a/c triggered more frequently and so this test once again shows its imperfections and limited control that said it's good to see how the room temperature behaves when left uncontested by a/c and is all ultimately experimental not comparative anyway with no a/c to help out the room temperature starts at about 20 degrees Celsius for all of our probes reaching only 21 degrees after half an hour so no big deal by the one hour point we're at 22 degrees this climbs again to 24 degrees after 2 hours so it's almost perfectly linear temperature still hasn't reached steady state and so we allowed the test to run for 7 4 hours at the end of all this we're at 25 point six degrees and climbing higher as it's still not perfectly steady state that said seven hours is a long time to run a machine at full tilt so keeping with our theme of realism if you're running a 600 watt machine for seven hours you're probably rendering or encoding something and that means you can probably exit the room if you wanted to our final test is the one that prompted all this we know that Vega is power-hungry more power-hungry than its direct competitors the 1070 or the 1080 for 56 and 64 so we wanted to compare 1080 to have a 64 card and see that given the higher power load created by Vega 64 is there also an appreciable change in room temperature with this particular room so that's the next and final test this test was a loopy and firestrike benchmark for a few hours and here's the power consumption chart the Vega 64 system consumed about 360 to 380 watts on average where the GTX 1080 system consumed 302 330 watts on average this 50 watt peak to peak amplitude manifests itself minimally in room thermals as you can see in this chart of temperature from the combat seat once again a 1 foot distance after a few hours the Vega 64 card heated our room by about an additional one to 1.25 degrees Celsius over the 1080 card and given the inherently difficult nature of controlling this test you know we can sort of write this off as functionally equal and start mattering as you stacked more and more cards but with single card gaming machines under gaming workloads with this type of power load it's just not showing up in rooms this is also without AC ever triggering to even turn on so there wasn't enough of a temperature change for it to fire if the room were smaller we had more machines running or if it were already hot outside this might be more noticeable our chart of temperatures from the middle of the room it shows largely the same performance though the temperatures are even closer as we get farther from the PC making it even less relevant and that pretty much wraps it up so this testing was a lot of fun it was different again it's an academic experiment the 900 watt to 1 kilowatt load whether that's a gaming machine overclocking machine or mining clearly creates a big impact and just to exit all the objective data and speak subjectively walking into this room at the end of the test at the end of the mining test was noticeable it's like walking into a wall of heat like if you transition straight into a desert basically so going from 20 degrees on the other side of the door to 30 in here is definitely noticeable it's enough to make you sweat a little bit depend on who you are and so that's an area where obviously you pay for what you're running in terms of how you feel personally in the room now opening the door again subject to leave this wasn't measured opening the door definitely helped a pretty good amount in terms of human feeling probably drop the temperature by about maybe 3 degrees after it was allowed time for air to kind of circulate out of the room again and so that would certainly help if you had that capability another option of course turn on a/c to run it all the time or open a window or any number of things you could do to combat this but clearly within a relatively controlled environment there is a a direct increase in room temperature versus your computer's power consumption so that's pretty cool and this is where things come into play of how much does power consumption matter how much does a heat load matter all that all that stuff but between two competing cards head-to-head even with a an 80 watt difference amplitude peak to peak is obviously not a huge change so we were looking at one degree 464 versus a 1080 after a couple hours probably not going to notice that much in this type of room maybe if you were gaming in a really small room with no AC and you got other machines running you might notice it at that point but doesn't look like it's gonna be nearly as noticeable as just going from something like single card to crossfire or SLI for example so that's all for this one as pretty funds has to do let us know what you thought of this idea post comments below if you have similar test ideas to this one that or maybe less easy to control but still fun from a scientific standpoint and as always you can go to patreon.com/scishow and exit stops out directly if you would prefer to not use patreon you can go to store that gamers nexus net the bridge is a shirt like this one subscribe for more and we'll see you all next time crossfire or a for example so that's it for this one kind of fun let us know what you think of this idea is tanked cut