Gadgetory

everyone and welcome to another ask GN as always leave your questions in the comment section below for next week's episodes we're shooting three today one is going to be on the patreon available side of things so you go to patreon.com/scishow his exit stubs out and get access to that extra episode bonus episode the other two will be here on the main channel we ran probably some pretty close to each other within 18 hours or so and we have several good ones for this GG 10:30 follow-up questions mostly legal ones and a questions about not all gigahertz being created equal things like that also a really fun flow question for fans before that this video is brought to you by thermal takes view 37 case the view 37 focuses on highlighting custom PC builds with its full panoramic window and tinted front acrylic and our thermal testing the view 37 performed reasonably well when considering its looks focused build which is partly thanks to the airflow design and the removal of a bottom power supply shroud for a balance of looks and performance check the link in the description below for the view 37 before we get any the first question as I noted in the other episode that's going live this shirt we've hauled through all of them already so we're be closing pre-orders in about a week when we run out of our inventory that we've ordered if you want one get your order in now on store documents access dotnet it's a limited edition shirt foil print shirt of our 10 year anniversary GN tear down logo design once they're gone they're gone we're not gonna make this shirt again ever so the pre-orders closed in about a week if you want to get it get it now first question is from dr. guns for hands who has one of my favorite names I think this was on the patreon discord dr. gunsther hands says just watched the GT 10:30 video and I have a question that I'd actually liked to direct at the lawyer guy you had on about the Nvidia NDA if he's willing to would having these two different products is referring to the GT 10:30 with gddr5 and the 10:30 with ddr4 we're having these two different products that perform wildly different to the tune of two acts in some cases with the same branding and hiding the differences in the product number constitute a misrepresentation of the product by FCC think probably you mean FTC guidelines NBA deceptive practice that they could end up on the hook for so first thing yeah FTC at FCC get confused all the time it's FTC that would handle something like that and the answer I think very briefly is no I don't think there is going to be a legal problem there so obviously we as a media outlet take issue with the GT 1030 ddr4 and gddr5 cards the fact that they exist with the same name and have such hugely different performance is something we didn't like and we said that in the video and made it all very clear and all that stuff the gt-r five versions actually a decent pretty good performer for the price but the ddr4 one is awful and it actually wouldn't be such a big deal if it didn't have the same name as I said before because I mean then it'd just be just be a low-performing card for display out but having the same name is what makes it misleading as you said and I absolutely agree it is it is definitely misleading now I am obviously not a lawyer I have spoken with two but but I had a pretty good feeling I could answer this question even without talking to them I did get some some legal oversight though just to answer your question as best I can so my feeling and and the one that the lawyers we spoke with basically confirmed was that although it is kind of a just a not great business practice and potentially misleads consumers to buying their own card although those points stand I don't think this is necessarily a legal problem and here's why ultimately it what is the what is the phrase caveat emptor it's all buyer beware everything is on the spec page so it says ddr4 it says ddr4 somewhere in the name even if it's kind of like a weird serial number basically that a consumer might not understand but it's there and it's in the spec sheet so really they've done I think the legal requirement to disclose what the thing is and you look at other companies do this all the time like SSD companies they change components yearly for the same SSD so you might buy it same for a actually some memory there's a reason a lot of these companies motherboards to VRMs well sometimes we'll ask this is more true for low on components high-end so a lot of time with low and budget components we'll ask like what PRM are using what nand are you using or what memory dyes are using and they will often the companies will say something to us to the effect of like well we can't answer that and of course my response for years was always but I can open it and find out so why don't you just save us all the time tell me and the reason I finally got an answer from someone on that because I would open them and I would find out and so and finally told me actually it's because because it's such a budget component we might change suppliers on that stuff and as long as it performs within probably their rough guidelines they're fine with changing suppliers but they don't like to make a hard claim on what's in there because it might change so that's kind of a different issue but that kind of stuff happens all the time though and you don't see lawsuits over those normally but anyway then that's because they they define aspects like maybe IOPS or maybe memory frequency and timings and as long as they meet the frequency and tie means then they've done what they need to do to meet that because they weren't promising a certain type of memory die or in probably in Sam's on P die or whatever but anyway this is clearly different because this is an instance where there are two products not one and they have pretty close names but yeah answer it again I am NOT a legal expert I have spoken with some and but I yeah I think this is really a non-issue in a legal sense in a moral sense I think it's a problem and I don't like it but I don't think we're gonna see on video dragged through a lawsuit over the GT 1030 it could be wrong you can sue anyone don't have to be right or win or anything but I don't I don't expect it it's a little bit it's different from like the gtx970 practices because this is a it's disclosed it's just up to you to figure that out which again although shitty I guess is not a legal hook so next question is for and and thanks to our lawyer friends who helped out with that next question Louis Beck actions quash that hey I've got a question for Sen can you perhaps give some insight to the old adage not all gigahertz are created equal I deal in older hardware a lot and can can 100% confirm that sometimes a slower processor will outperform a higher clock counterpart I know that some of this comes down to cache size yes but how does the actual IPC compare between generations and if possible between products lines are there any instances of modern processors where the older generation or a specific older part compared to its specific newer counterpart would be a genuinely better choice great question a good example of this is how the later generation 1 gigahertz Pentium 3 processors would outperform the early 1.3 gigahertz Pentium 4 CPUs just by p4 having higher clock speeds yes so that's certainly one of them the next one I would point to is probably feed on to feed on to was a crazy good processor phenom ii x4 and it's unfortunate that i had to deal with Nehalem or sandy or whatever came out right around as we talked about in our phenom to revisit that we did it's unfortunate because Intel kind of came out of nowhere almost with massive improvements the Sandy Bridge and the Heylin was pretty good but phenom ii x4 was it great or was it x6 i think it was hack 6 wasn't it sorry my memory it's a long time ago I think it might have been x6 either way the point is phenom 2 X whatever phenom 2 was really good and it was better than bulldozers or in a lot of instances I think IPC was better on phenom 2 and I'm I'm pretty sure it was versus the original bulldozer launched bulldozer was pretty messy when it first came out and later on in life for that instance but a literal bulldozer I don't just mean like the name people use as kind of the overarching name for FX I mean like literal the first bulldozer processor was bad and phenom 2 it was better than it in a lot of instances so I would say that's a great example of what you're talking about other ones I'm sure there are other examples that's kind of the one that sticks in my mind I'll think about it and let you know if I come up with another one but that's probably that's the best one I have yours pretty good too other than that so some specific examples you're talking about I guess these are counterparts but not from the same company this is a Dressen you're not all gigahertz are created equal the best example of this is different architectures so same architecture like risin risin two or whatever they will share basically all of the core features that make Zen the Zen architecture but if you're looking at different architectures do you have things like you said cache size is a big one it depends ultimately on the kinds of things you're processing so if you're doing something that is really thrashing the cache then having a superior cache and one architecture versus the other would lead into that adage of not all gigahertz are created equal because if it's cache thrashing and just hitting constantly then having a faster cache or bigger l1 so anything that's closer to the CPU will be helpful you talked about like l1 l2 l3 cache it's kind of it's basically names like how close in proximity physically is it to the CPU so different cache sizes for each of those will impact performance in great ways not necessarily contingent on the frequency as much as just like the cache physical proximity to the cores and and the size of the cache and speed of the cache so that's one example Intel when Rison one launched was about seven and a half percent ahead an IPC this is something AMD said even so that's an example of where they had an IPC lead intel's rise into caught up in a lot of ways but Rison one there was about a 7.5 percent difference there so that's an example of where clock for clock you still had an instructions per clock lead with Intel at the time obviously the threaded multi-threaded performance is a different story and another example of not all gigahertz being created equal because if you have a low frequency processor with a ton of threads then that frequency is going to be less of an impact on things that care about threads like blender or something like that and frequency matters more for things like premiere but ultimately just having straight frequency if the rest of the architecture doesn't line up doesn't matter a whole lot because if you have 8 threads or 8 cores at 5 gigahertz versus 4 but the process of the product only used is for then doesn't really matter the product in this case being like Adobe Premiere I mean software product so that's an example another one I took a couple notes here let's see yet so I PC I would say is more deserving of a standalone video we might do that I'll talk to some people bottlenecking is also interesting here so take Vega or Titan V GPUs we're on both of those use HBM but the HBM is kind of on Vega is a major bottleneck point so as you overclock the core more and more again not all gigahertz created equal you keep pushing the core maybe you push it another 50 megahertz or something like that at some point major diminishing returns are hit if you're not also increasing the HBM clock in most gaming applications not everything but most gaming so in that instance it's a case where you can keep pushing the frequency but at some point having a frequency that's 50 15 whatever 100 megahertz higher then the choke point on memory it becomes irrelevant and that's because the memory is now the bottleneck and that's just because big architecture happens to be very memory bandwidth intensive you look at the GT 1030 this is another great example so let's talk about this one I set the GT 1030 this gddr5 I think this is the g5 card I sent the GT 10:30 with gddr5 the same frequency as the one with ddr4 core frequency and you know the same core frequency it was a massive difference like 50% differences with the gddr5 card being faster so you're talking 2x performance from baseline ddr4 card at the same core frequency and that loss is because of the memory bus and because the memory bandwidth which is derived from the memory bus and the memory speed so that's another great example of the core frequency gigahertz not being equal because ultimately you ran into some other barrier and I am NOT an expert on like really low level architecture I'd love to be but it's it's you know I kind of I know my place and I've been focusing on obviously learning as much as I can but a lot of my job is talking to people who know more than and getting their insight and sharing it with you so what I plan to do is hopefully someday we'll I've got some experts in mind who are brilliant and we'll talk to them about IPC specifically so I can expand on that in a more competent and collected way but I hope those examples help answer your question next one Gigi chb gig GHB says which moves it more air one fan speaking of experts one fan at 1500 rpm r2 at 750 stacked together or in side-by-side config what about noise and energy efficiency what if the front fan slower than the rear fan when stacked like in jet engines so I'm pulling out my phone for reason here I didn't speak to someone about this so this is a fun question basically let me boil it down to this if we have two fans side by side that are 350 LPM linear feet per minute of flow versus one that's 700 is there a difference and how much air is actually going into the case we kind of have to do the physics thing of like assume a spherical cow we have to assume a whole lot of things here like we're ignoring that the case impacts air flow or ignoring a lot of things let me read a response on this from VSG from thermal bench PSG is one of our experts we speak with he works he he runs the thermal bench website it has great resources for open-loop testing that's and fantastic as well which we've been talking to him about fantastic - so I would encourage you to check it out but he said feel free to paraphrase I won't read it but honestly it comes down to the specific fan and airflow restriction in play the PQ curve of fans isn't linear and airflow restriction is certainly not so for higher restriction applications such as radiators it is possible that a single fan at 2,000 rpm will fare better than two fans and push pole at 1,000 rpm and this is assuming other things effective in pitch and air field combination are not a big factor to give you an example I had tested this with the Corsair h75 AIO it came with two fans so I used one first and then both and push pull and VSG gave me one of his graphs we can share with you from his website so these graphs are little little complexes a lot going on but they extremely helpful in detail PWM signal percent is on the bottom and then the rest of the data on the sides rpm on the left and let's read the rest of his response he said for a less restrictive application to fans and push poll will generally have lower noise for the same amount of air pushed through all other things being equal so that's why it was popular to do so with water cooling back when radiators were optimized for high air flow and performance now that operating noise has become a priority in other regions than just Europe radiators are designed differently so push-pull is a waste of money unless it's an aesthetic choice or you are absolutely limited by radiator space side by side versus push-pull can't really be compared since the airflow field is different so hopefully that helps you out and that goes back into what I was saying talking with experts vs she does great work I I would encourage you to check out that chart you rewind it if you need to it answers your question pretty damn well next question is from lo1 or LOL i guess ii 101 says how is the thermal conductivity of the foil use for the shirt of the foil years I would like my shirt to conduct heat away from my body also do you think the KB like our mobile CPUs were in effect of Rison as well as impending rise in mobile conductivity of the foil we use I'm not sure maybe we can get snowflake to test it she's the one who oversees our testing operations so I'll certainly ask her if we can scrape some of it and find out what the thermal conductivity is using some high-end thermal conductivity testing solution as far as the question real question I suppose or the secondary question because let's face it this shirt is pretty awesome and you should consider picking one up on Stuart exit on that next question do you think the KB like our mobile CPUs we're an effect of Rison as well as impending rise in mobile came they are the whole everything to do with AMD and Intel right now is really interesting so first let me say this I think Hades Canyon which is them working together was a a really big kind of warning shot to Nvidia and video dominates that low-end sort of not even low and just let's call it HTPC devices look at ZOTAC there's small PC boxes almost universally they've used and video mobile GPUs or modern Nvidia desktop GPUs now they've gotten topics the laptops pretty much all used on video GPUs they have like 90 plus percent market dominance for GPUs not in Taiwan's in laptops and so Hades Canyon I honestly think scared and video quite a bit and seriously I do think it was a concern of theirs and that I think was a response from Intel and AMD to Nvidia so kind of different part of the question of of Nvidia's market dominance and Intel doesn't particularly like Nvidia and vice versa so I would say that hades Canyon and that product is a response to Nvidia KB like our mobile CPUs things like that I would say yes I do think that Intel in general not even just with this product but in general is responding to AMD it's it's really exciting to see we haven't seen this kind of action in the CPU market in a long time so yes other examples of Intel responding to AMD so one thing here before I get into that when people talk about like this is a response to X first keep in mind that a lot of the products we see silicon especially are in development for years so like even just a case some of those take two years to make so if you have two cases launch that are pretty similar to each other maybe five months apart just remember the second one might have still been in development two years ago and they may have never known about it until the other one launched and the same goes for silicon for CPUs and GPUs and so there are responses but for things that kind of launched immediately after rise and like coffee like coffee like as an architecture is not a response to rising it was going to happen but what I think is a response is things like the stack in terms of pricing for how many cores you get or marketing especially gigantic response there and the bigger thing that Intel has done is pull in all their launch schedules X 299 was supposed to launch at least a month later than it did three weeks two longer and they pulled that into Computex of whatever year that was last year I guess and that was a response to thread Ripper absolutely the 28 core CPU that was shown off prematurely that was a response to thread report to coffee Lake the you might remember the low-end platforms non Z we're not launched with the Z series they were launched later and that's because the Z series was pulled in with the 8700 K in 8600 K I think and those were launched ahead of time as a response to rise in gaining popularity for high thread count users so those are the responses now architectural e I would say sort of late this year and next year might be when we start seeing the beginnings of architectural responses not just like marketing and pricing and timeline responses so hopefully that that gives some insight I think we'll also see changes for spectrum meltdown and stuff obviously but the answer is yes I think there are responses but just remember that architectural II that stuff it can be years so they're probably pretty locked in by the time the stuff is shipping for a long time at that point they're just doing fine-tuning and there's still can be responses but a lot of them are going to be less physical to the hardware itself next one zib Zoolander says if I put six Noctua NF a 1225 fans and of course our 280x will the build collapse into a black hole time-sensitive question please respond next question Knight Rider 21 says hi Steve you guys talk about overclocking but doesn't make sense to buy an unlocked CPU to undervolt and save on noise I haven't actually tried this in a while but a lot of the unlocked CPUs I've worked with like in laptops you can still undervolt them so I don't know that you actually need one to undervolt it's more of the ratio that gets unlocked and the base clock so I don't think you have to buy an unlock CP if someone has tried recently please correct me if I'm wrong in the comments but last I tried was seventh gen and I could still under volt below stock just fine and I think that same was true for 8th gen anyway is it worth it yeah if you want it to be quieter the way it would work is you drop the voltage then you can drop the fan speed significantly and then then it could be pretty damn quiet if you don't want to just invest a lot of money and trying to silence like with quieter fans or bigger cooling radiators whatever so yeah well I would say no it's not worth buying an unlock Seaview for that purpose it is worth buying a CPU and under-voltage because I think you can do that anyway it's like I said like you said we mostly do overclocking so it's not something I try a whole lot but yes I think pretty sure on most platforms there's no problem with doing that as long as the BIOS permits it because it's not like it's going up it's going down so yeah it's worth under volting to reduce thermals noise next one this name is awesome banff hammer says what a monoblock add more heat into the system and possibly reduce clean components that matter if the system is small it's going to depend on the size of your cooling solution the more liquid there is the more that heats going to be spread out it will ultimately yes impact things so a great example of this with thermal data we've tested it's not a literal mono blog was pretty damn close we tested like it was the FTW three or the gigabyte card or whatever one of the water forest cards where they had a copper cold plate it hits the GPU and it expands out and hits all the GDD are modules and some of them expand out further and go to the vrm it's a great way to cool all those components with liquid without actually having water flowing through blocks but it does it increase the GPU core temperature and the same thing applies here to what you're talking about so just just translate what I'm saying to CPUs but we did a test with a cold plate on the GPU and then we did a test with adding the expansion plate to cover the vrm and the GDR memory and the GPU temperature went up we'll call it but that's because the cooling solution is shared with more components so it's dealing with more heat you might have like say I don't know 100 watts 150 watts of heat or something like that coming off of a silicon processor and then you might have another 8 watts of heat coming off the vrm or 15 or some low amount and as long as your thermal solution can handle it it's not that much of a hike and it can be a benefit for the other components things like GPUs the trouble is that they are more frequency sensitive to thermals so an extra 5 degrees can matter there but with a CPU if you increase your CPU temperature by a couple degrees because you are now sinking the vrm as well it's not normally a big deal unless you're already like in the 90s so the answer is yes it will impact it but maybe not meaningfully I think we have to more quickly let's see Space Jam flam hey build Zoid and Steve and Bill joy too knew how to safely apply a liquid metal to Thane's video you use the motherboard mounting bracket as your only retention method for the IHS so no glue or anything do you think this is an adequate long-term solution for me it would be on coffee Lake horizontally mounted in the case yes the answer is yes that bracket is we we do the same thing I almost never put adhesive on deleted components because it kind of eliminates half the point of deleting them because you're introducing a gap now so I actually tried to avoid adhesive unless I'm shipping it to someone like if I ship a loose processor out to another youtuber who wanted us to delete it for them I'll seal it normally was crazy glue or something but otherwise I don't and that's because the the clamping force on that socket is significant and then you're gonna put a cooler on top of it so that thing's not gonna move it will not move so I would not worry about that aspect of things next question last question tambourine says you really got me with the get it before it's gone forever at the end there supreme marketing skills Steve ordering one and a Matta thank you very much for picking up the mod Matt our next round some of you have been asking it is almost here so email support I gamers Atkins dotnet if you need an update on that but we're pushing them to ship them end of this week as a filming so that would be week of the fourth so end of this week we're hoping they start shipping to us and we can distribute them thank you're picking one up and thank you for grabbing the limited edition foil shirt you can get a story to gamers next it's not net get the other episode on the main channel we film two main episodes they are both pretty good you get the bonus episode on patreon.com slash gamers Nexus and subscribe for more as always thank you for watching I'll see you all next time