Gadgetory

hey everyone welcome back to another ask G episode we just got done last night with the live stream where Paul jumped in so we did a live stream overclocking of the 2080 TI fe and the 2080 TI x ii ultra from EVGA ended up beating the goal of the goal of Saudi 400 points in times by extreme graphics we got 70 for 35 Paul jumped in at one point and we we started a beef a bit of a beef between he tubers some west side versus east side going on so he got 7400 before I did and then about 30 seconds later we got 70 for 12 and everyone spams Paul's Twitter so but Paul did end up going going up higher on his score I don't know if he beat 74 35 yet but I think we're gonna keep it going so keep an eye out on that but we haven't asked yet so as always leave your questions in the comment section below for next week's episode we have some good ones this week got an answer from Johnny guru himself who works at coarser these days not Johnny guru anymore and then also a pretty good discussion on how game explosions work before that this video is brought to you by EVGA SRT X 20 atti XE ultra video card we recently used this to beat our founders edition overclocking results with its additional power target headroom and cooling capabilities the XC Ultra uses a 2.7 extra thick heatsink for a quiet operation under low loads but also maintains higher clocks on average over the FE model learn more at the link in the description below quick update on the store as always we boil shirts are officially 100 percent gone they will never be back we had so I stated this in the stream but we had we ordered the amount to obviously meet the the orders for pre-orders and all that stuff earlier and we end up ordering a couple of extra just in case any got lost in the mail during that first round of shipping because I would feel terrible if someone's got lost like by USPS on the way to Australia or something and we didn't have any to fulfill it so we were two extras fortunately none got lost we released the extra inventory on stream last night and now they're gone forever so thank you for picking them up if you got one a lot of fun with that shirt we really liked how the design came out and other than I mean if you missed it you missed it that's it so anyway thanks ty when you picked one up you are one of the few let's get into the questions the first one is from Andre bolo who says I think this was on YouTube asked why do games usually stutter when you are close to an explosion or some other demanding effect I was playing Darksiders at war mastered and I noticed a short stutter when throwing cars into each other in order for them to explode or make some Souls the closer I was to the explosion the more likely the game was to have a slight drop in fps isn't rendering isn't it rendering the same thing this is a great question and so you have asked also separately about why the name gamers Nexus which is like hardware Nexus or something well we started entirely in games and used to do a lot of coverage on games and how games work so I have a good answer for this one if anyone before getting into it if anyone has questions about game graphics specifically please leave them below because we have some good contacts in the industry industry at engine makers at Nvidia AMD we have Andrew in our own team does 3d art so we can get good answers to game graphics questions so for this one not exactly sure how Darksiders does it but there are a few ways games do explosions just cause is a pretty good example of impressive explosions in this one the explosions aren't truly dynamic but they're faked really well anything with a true dynamic volumetric effect would be difficult to run especially in real time but like most graphics it can be faked in clever ways true volumetric effects and explosions would have smoked Angell's for example cleaning - a plane as you fly through a smoke cloud like in Just Cause 2 but that's not so easy to do in real time you could also see real dynamic effects by flames so you have an expression of flame being displaced from the planes propeller for instance this becomes a computational fluid dynamic problem or CFD and it's not really feasible in real time on a grand scale but it can be done in isolated areas one way to fake this kind of effect though is to divide bounding boxes into voxels which are then used for a calculation of the results of the explosion at is fluid dynamics the bounding boxes use an algorithm that is typically built per game or per engine and that algorithm runs on a per voxel or even intra box'll level to determine how the explosion behaves with regard to the graphics output a lot of this is GPU intensive math and runs floats on the FP use or the CUDA cores if you prefer or stream processors but the resulting physics calculations tend to go into a single thread on the CPU if the effect begins to exceed the confines of the bounding box that means that the fluid is clipped so if you've ever seen smoke for example clip into the side of a 3d object like a car or a building and counter-strike or something although a lot of those are just sprites actually you'll see it cut off and the bounding box size could also dynamically change or scale to resolve this problem but that creates a huge performance hit like the one you're talking about so the bounding box size dynamically scales suddenly more computations are required and memory cost increases greatly as we noted in our old Nvidia vxa o coverage you could also expand the voxels but this would result in decreased resolution and fidelity which is very noticeable by the player so the ways of faking it not really perfect but that would avoid the clipping issues where fluid terminates abruptly and the visual quality of the facts would be greatly downgraded with the reduced resolution bounding boxes are also imperfect four diagonal effects or thin effects like playing contrails where a lot of additional voxel space is wasted and system resources are burned in the calculations a lot of things can cause slowdowns with game explosions one of them is exceeding the bounding box size to improve fidelity of the fluid dynamics but it eats more vram and burns more cycles in calculating the floats another reason is CPU side where excessive in-game physics calculations will get forced onto one thread of the CPU potentially causing a pipeline stall if that thread can't keep up you could also have an issue of bouncing off of the frame buffer limit with resource consumption as more operations are pushed into memory via PCIe and where assets are pulled like real-time damaged assets for those vehicles draw calls things like that for an effect that is less volumetric and more geometric you run into an issue of excessive draw calls a draw call happens each time the CPU has to tell the GPU to draw another primitive like a triangle and it's primarily an issue in dx11 or older higher-level ap is also changing a significant amount of models or meshes simultaneously result in a sudden influx of draw calls that must be juggled by both the CPU and the GPU which could be happening in your instance so Nvidia tried to solve a lot of this with game works flow by only allocating voxels immediately affected by explosions and using Z buffers to change voxel allocation as the fluid expands or contracts but the technology never got wide adoption so we'll show some of that footage in the answer anyway you just won't see it in games these days and those aren't the only ways to deal with explosions in games of course sometimes you'll just see really simple sprites so battlefield 5 that demo that's supposed to be really impressive with all the rage racing and everything well if you look at it closely enough the fire that's in there not the raytrace fire but the normal just fire on vehicles it's a two-dimensional sprite with a lower frame rate than the rest of the game so it's it's really not a perfect way to do it but actual fire effects are really expensive to do in terms of resources so there are lots of ways to do explosions that's kind of the main one that we talked about a couple years ago on that article and bounding boxes are a good way to do it it's just you run into some resource cost issues you run into issues where the effect clips you could do dynamic explosions but those are extremely resource intensive and cause the issues you're talking about and aren't too common but hopefully that helps out next one is from this is another actually really great question with an answer from John gros aka Johnny guru he used to run the power supply review site so the question was from Merlin to one see who said what impact if any does an uninterruptible power supply have on the performance of a power supply in a PC in terms of loss of efficiency due to voltage ripples etc also would you recommend a UPS for hooking up valuable computing or gaming equipment in general to further aid in power surge protection and not only against brownouts so John's answer I'm just going to quote it cuz I can't do better his answer was a UPS doesn't have an impact on the PS use performance while it is still on main power ie not running off of battery but obviously your efficiency is going to be shot to hell not that the PS use efficiency is changed but since the UPS has to maintain a charge to its batteries and other UPS add-on features like LCDs that tell you the battery level power consumption etc it's consuming power in addition to whatever your computer is consuming as for is a you PS when running on battery good for your power supply that entirely depends on the ups and the PSU there are a lot of variables involved in the plethora of UPS is out there switch over time or how long it takes to switch to battery power when the mains drop and whether this period of time is greater than the holdup time of your PSU actual output capability UPS s are typically marketed with VA capability but not in watts ie a 1000 VA UPS may only support up to 600 watts and the shape of the sine wave coming from the UPS went on battery power considering the latter depending on how square the sine wave produced by the UPS is it could potentially damage a cheaper PS use bulk capacitors sometimes to achieve a 230 volt RMS output the wave that may exceed the voltage rating of the capacitor now a good quality power supply that sees this should shut off and not all UPS is will over voltage to obtain a proper RMS voltage it really comes down to doing the proper research prior to purchasing your PSU and your UPS bottom line is however a decent quality UPS and a decent quality PSU aren't going to have any issues whatsoever as for the use of an AVR it's simply better to just have a PSU that doesn't crap out on you turning a brownout it's not uncommon for a 230 volts only PSU to shut down if the mains drop below 180 volt or 190 volt depending on the build of the PSU but if you have a PSU that supports a full range of input voltage the PSU won't even blink pun intended if the voltage drops all the way down to 90 volts of course if you're somewhere where the power drops out completely for periods of greater than 16 milliseconds you should get a UPS as for surge protection never take for granted that your power supply has built-in surge protection or even if you know it does that is enough or won't result in a dead power supply even though the rest of your components should survive intact get a good power strip he says and by good he means that one that you know is dead when it gets hit with a surge it should either stop working completely or have a light that tells you it's still protecting I prefer the latter or one with a built-in breaker that trips during a surge and can be reset - these can be expensive like my favorite trip light isobar but are very cool so TLDR quality UPS quality PSU quality power strips no problems ever so that answer is from John or Johnny guru as many of you likely know I'm Johnny guru for the longest time and still has been one of the leading power spot review sites when John worked there definitely kind of created a market for power spot reviews so this is for sure an expert that can be trusted on the answer and he's these days if you're curious he's at Corsair but still provides great I mean you'll notice this is the difference between marketing PR and people who are more engineering side not once not once did John use the word Corsair in this entire giant block of text so it's not even trying to market something to you which is pretty cool we very much appreciate that thank you John for the very straightforward answer and I guess if you do want to follow up on Johnny guru he's still got someone writing reviews over there so next question is from raid the road she said Steve a scene that the 8700 K or 8086 K is the bottleneck on the new GPU is that 1080p gaming would someone have to push for one of the enthusiast CPUs in order to reach say 240 Hertz gaming or we at the point where the current TV offerings just won't allow us to break into that 240 FPS for the majority of games so the the problem with h.t.t.h EDT cpu is like the Intel enthusiast CPUs is that as you increase core count you're gonna decrease the frequency and a lot of games still care more about frequency than cores especially be on something like six or eight cores or twelve or sixteen threads so it running a seven and eight exe would expand your head room for 3dmark but for a lot of games that drop in frequency will impact you in a negative way more than the cores will impact you in a positive way so a lot of games run maybe eight threads they might have game physics thread game logic maybe AI processing definitely a render thread that'll get hit really hard and as long as the render thread is one thread entirely on its own and not split between more which is common for a lot of these modern game engines the frequency will matter more because that single thread is going to hold the entire pipeline if it's not fast enough to keep up with all the draw calls all the render operations coming down the pipe and then physics of course is also heavily impacted in some games so the answer is to an extent HDD can help like the seven 100 accent maybe but in 8086 Karen so fast it's just it's gonna depend on the game and not many games will use 10 cores 20 threads so these instances frequency will help more 8700 k 8086 k you see rise in 2000 series really caught up a lot so frequencies not everything architecture matters too but even then a high end rise in cpu will make more sense in a lot of games than like a 32 core threader per CPU the the frequency hit will hurt you more than the core count gain in almost every gaming scenario that exists today so what's the limit then yeah sure nadie 86 sometimes did bottleneck the GPU we were testing at 1080p but the problem with that it wasn't always F 1 2018 didn't really bottleneck until we overclocked and even then we're not sure a bottleneck because we don't have anything better than that to validate so it doesn't always bottleneck at 1080p with the 20/80 TI you can do stuff like blast anti-aliasing buddy free goal is 240 FPS that's not gonna help so I don't know 240 FPS is hard to hit anyway I would say an 80 80 6k 7900 ax one of those would put you closer to that goal but it's gonna depend on the games you can always drop settings like if you're bound by the CPU not the GPU you can start dropping things like geometry so there's any setting that changes how much geometry is on the screen or the details of the meshes sure it'll help the GPU to but it'll primarily help with the CPU and dealing with a lot of draw calls and LOD view distance those things would impact how much the CPU limits you as well so anything like that game physics would impact you those are aware the the CPU tends to get hit harder than the GPU in games but no I don't think if it doesn't bottleneck at that much it just it depends on the game like like I said F 1 2018 didn't really bottleneck on the TI which is kind of impressive but so wasn't 240 FPS there's only 211 that's that's not enough for you I don't know tell ya there's not a lot more that you can get trogdor 8 Freebird does that near t.j.maxx actually hurt the CPU for example a 7700 K at 95 C 24/7 that sounds bad to me and probably many people's ears but is it actually bad or does it just sound bad because typically hot equals bad 95 C is pretty high we'll be some thermal degradation on parts so example of this if you're working on well degradation to VRMs is pretty common if they run like 125 C or something like that you start D rating but capacitors another really good example caps over 105 see if the rated for 1 over 5 C it's typically for 10,000 hours five thousand hours two thousand hours something like that and once you go over that 105 C point every 10 degrees you have the life of the capacitor so those are instances for sure or it matters but the caps even on the Intel CPUs it's not like they're on the die they're far enough away they're not going to be running at 95 C it will be significantly lower so that's not really your concern typically the bigger problem with CPUs as thermal cycling not sustained load so when when I worked at Dell we would do thermal aging tests so the goal is to see like how long does this lithium-ion battery for a laptop survive under these adverse thermal situations and we did ruggedized PC testing stuff that would go to like Iraq for the war so you test laptops under extremely intensive scenarios for heat because they're going to the desert and dealing with sand and so the way that a lot of that was tested this is all pretty much public knowledge at this point the way it was tested is by simulating age with heat cycling so you put that stuff in there was a where I worked there was a building that had a what was it called the name of the lab but had a shock and vibe that also had thermal chambers and so you put the stuff in there and cycle heat loads so it go really hot and cool down really hot some of them even had chambers that could go cold to like below ambient and that's typically used to simulate age and the reason I bring this up is because if that's used to simulate age rapidly because you have to think the product is being developed it's coming to market let's say you're advertising it as hate can survive five years in these conditions you haven't tested it for five years you don't really know so one way to know is to simulate it and if they're doing that to simulate age and this is pretty well known to you know for sure that thermal cycling is worse for it than just a straight 24/7 load so that's the theory and practice it could be different but getting back to the heart of your question running at Intel does a bit behaves differently than AMD at 90 C 95 C Intel as long as it's not a t.j.maxx it's not dropping clocks will age can't really answer that firmly we don't have like enough of a sample size to truly know but realistically at 90 C on Intel yeah 90 C looks better I could say 95 that's getting kind of high for my my taste for a lot of the processors but 90 C leaves you enough Headroom where if you're under if your ambient temperature goes up 5 C which is not crazy you're still under TJ max where's 95 it goes up 5 so you're hitting TJ Max and that's gonna be a problem but 90 C Intel is not gonna drop clocks it doesn't look good it's probably not good for your OCD thermally if you like if you have that but it's fine AMD has an office mate optimal operating temperature below 70 degrees Celsius depend on if you're calculating in that offset they do so optimal temperature below 70 C and it behaves closer to GPUs and boosting frequencies so for an D rise in 2000 series processors if it's cooler it'll boost a bit higher the hotter it gets the lower boost so that is a limitation and also I think they aged a bit differently - but I don't they haven't been out long enough for me to really know build Zoid may know more about this you can go paint him on his channel actually hardcore overclocking but my my answer to you is in theory the damage should be worse from intensive thermal cycling between like 90 and 50 than from straight 90 all the time 95 I try to bring up it down a bit just because it doesn't make me feel good but I don't have any scientific like data to say 95 C is more likely to kill the CPU faster realistically for most of these processors they are likely to become obsolete just by nature of becoming obsolete then die from a thermal scenario they're pretty resilient and they have a lot of protections to shut them down their current protections all that stuff next question one of the last few tech professor said does case internal volume have significant impact on thermal testing beyond case form-factor primarily in that it takes longer to heat the extra material and reach steady state but our test is long enough that it doesn't matter we've factored that in and I guess you also have longer cooldown time too but and it's just it's more material it's bigger so it'll take longer to heat all that material up during a test next one Andre Scott says how is progress on the RAM timing series part one was very informative and helps to explain primary timings and how they reworked in relation to each other part two is done we are working on some animations for it just like the first one once those are done we'll have that uploaded and then we've got as in Therese ed just to be clear for new builders which way does it air go through a graphics card heatsink talked about this last week but correct me if I'm wrong the air is pushed in through it and exhaust into the case through the sides of the card with aiv cards rather than pulling it from the sides it's not pulled in from the sides it might be on like one card it's not it's almost never pulled in through the sides even on the blower card so there's Nvidia blower cards on the far right of the card if you have a PCIe slot oriented on the left far right of the card there's that grille and it looks like air comes in through there but actually it doesn't because if you take the shroud out underneath it there's the the blower fan and then it's got this hook around it on the outer edge of the card and that's to direct the air through that flow path so it actually doesn't come through the side of those either even though it looks like it does it's just it's just for looks those fins don't even do anything they're not contact its contacting anything so just we looked and in a partner card with like dual a triple axial typically yeah it works just like a case pan the air goes in down through the top of it and it gets pushed out wherever it can so it's not like there's not a flow path necessarily there's a flow path and so far as the fins are pointed at a certain way it's either horizontal or vertical that affects where it exits but the air just follows whatever path to exit it can which is typically out the sides and some can come out the back yes there's there's holes in the back for the reason for that reason two more elitist jerk who has the most appropriate name on the Internet it's just it's very fitting for I don't know a hundred percent of Reddit for example elitist jerk actually has a very nice comment though because elitist jerk says would GN consider doing D boxing videos in place of unboxing videos for unboxing embargo lifts that is to say destroy the Box creatively first while leaders jerk I like that you know you've I guess I should say your your father very disappointed you didn't live up to your name elitist jerk man what a what a missed opportunity should've seen the comments last week from TM so elitist jerk we had an unboxing video wherein we unboxed the box with an angle grinder let's just drop the b-roll of that in there and maybe we can do more of those if the companies keep doing this thing so you bring up a really good question that's been bothering me the nvidia embargo lifts did not for the US didn't allow disassembly of the cards and they didn't communicate that message to any other regions apparently so we posted ours anyway because other people started posting theirs from other regions so it is actually a problem that these unboxing embargoes are like it's just pure marketing everyone knows that linus didn't do an unboxing of the RT x cards linus was the king of unboxings when his channel was new and in speaking with him he's been public about this too if it's just a straight open the box embargo it doesn't add really any value and he knows that it's just a marketing play and I don't know if they're just trying to get marking his you can't even take it apart so I have a problem with unboxing barrios and if they continue to restrict the the loopholes we find in them where we can do other stuff with the card then maybe we will start doing destruction embargoes instead just angle grind through it accidentally cut the card and have something like that then oops guess we did a teardown by accident last question mark Gucci said not really an ass gen question I love the to stuff to animal cats in the background videos can you start moving the two cats in different places in each video like a mini Where's Waldo nope those are from the patreon backers though thank you for some of them hold on and as always leave your questions in comment section below go to patreon.com/scishow gamers next it's got the bonus episode and story I can't access that night to pick up a shirt like the one I'm wearing here and I will see you all next time