Gadgetory

welcome back to another episode of ask a GN as always leave your questions in the comment section below if you have them there were a lot of good questions last time we're not gonna be able to get to all of them but hopefully get through a good amount of these and the very least before diving into the sgm content it is again brought to you by and the free synced devices including this lg monitor which is a 34 inch display 34 40 by 1440 and has the usual assortment of letters and numbers that is difficult to remember but I think I got it I think it's 34 um 88 - P nailed it doesn't even have the check haha that is a that's an $800 display 34 40 by 1440 with free sync will have a link in the description if you're curious to learn more about it so let's get to the first question first one is brought to you by foreigns gaming tech who says Steve a new question concerning a high performance cards on low resolutions where you see a performance ceiling we saw this in your Titan XP review and you talked about this in tgw which is tech gaming weekly that is hosted by Joker productions I was on there a couple days ago I'll be on there again on Wednesday this week so he's foreign Gaming taxes could you explain why a Titan XP and a gtx 1080 have the same performance at 1080p IE why even a 69 50 X maybe bottlenecked in this situation so we were not using a 69 50 X I don't know how that would perform exactly the reason that 59 30 K it's a bit lower around than a 69 50 X but ignoring that in our specific test case where we had a 59 30 K I don't know what game I was talking about at the time but may have been GTA 5 at 1080p probably with a Titan X and a GTX 1080 I think with GTA 5 we hit somewhere around the same 134 FPS range or something like that so that could be a few things one depending on well the obvious point is that there's obviously a bottleneck elsewhere in the system this is almost always the CPU some games very specific ones like I think f1 2012 does actually have some memory bottlenecks to some points which is pretty interesting but not and normally CPU bottleneck and the reason that happens is either you've enabled a CPU intensive setting in the graphics maybe the game does its physics processing on the CPU or something like that or some other thread management job allocation function and the engine is not really particularly friendly with the CPU you might be stacking one thread or one core a lot more heavily than anything else in the system and that could cause a choke point things like that which are just optimization items on the game side normally that can cause it but most commonly it's something where because the CPU has to do work for every frame created there's going to be a limit at which the CPU can effectively no longer keep up with the GPU so you could put a Titan X in there you could put maybe three Titan X's in there if if I had them I do this to demonstrate it but if you have like say let's keep it single cards for easier demonstration you have a Titan X versus a GT X 1080 we showed that maybe even have other ten 80s at varying frequencies in there they might all perform within a couple frames of each other which shouldn't really happen at least with the Titan XP and that is I mean obviously demonstrated when you increase the resolution to 1440 or 4k it starts becoming more GPU bound and so now you can see the differences between the GP is more easily this is also why we try to use as high and CPU as possible for VP benchmarking but you're still going to run into limits and when the CP is trying to keep up with the GPU and creating these frames one of the things that it most commonly does fall behind it is the draw calls so every time for DirectX 11 OpenGL old your API is basically that are higher level and have these abstraction layers between the components and the software you have these abstraction layers in between every time primitives need to be created or Poly's or whatever the GPU is effectively it's communicating with a CPU for every one of those draws that's called a draw call and that's a problem that's why low-level API czar promising we still need to see better implementation but you get the idea where level API is can bypass some of that so that's normally where the bottleneck but there are other reasons as well like physics or something like that if it's cpu-bound next question is from Ryan Razer who says would you consider putting the gtx 1070 micron vram issue out there none of youtubers say anything about it whereas forums are filled with posts every review on YouTube has Samsung version of memory could you do a comparison of the 10 serie Samsung versus micron so I don't have both of those that's that's why we haven't done a comparison it's as simple as that as far as talking about it I suppose the the issue that's in reference here is basically a supply-side change so some gtx 1070 is shipping now have micron memory whereas the ones that shipped out initially had Samsung memory and this happens pretty regularly it's not new to the 10 70s and the reason it happens is just the suppliers and manufacturers trying to keep up with demand so in this specific instance Samsung doesn't output the same quantity of memory modules that is required to produce the GPUs in demand by the buyers and so Nvidia or whomever EVGA a big partners have to source their memory from multiple vendors in this case they're introducing micron to the mix so they have two sources of memory they can try and keep up with a stutter step of the supply basically it's pretty as more card so that's the idea now as for the performance difference I don't know the technical reasons why this is happening at a low level but basically as I understand it and I don't have one of these cards as I understand it the micron memory doesn't overclock quite as high as the Samsung memory and that's kind of the crux of the issue Ryan razor here is saying that some people report stuttering checkerboarding BSO DS I think that's mostly happening with overclocking but I really can't validate it personally so let's just go on that for now there is a BIOS update out there if you apply it from what I've been told by partners and by some of our users the BIOS update should resolve the issue where you can get those overclocks without the same stuttering or BSOD or checkerboarding issues basically artifacting that's being seen on the non bios updated cards so that's as much as I know that's really that's the start in the end of my knowledge of this right now until I have a device in hand I will not comment any further on it there's no point because I can't test it so if we get a 1070 device with micron memory maybe we can test it do a BIOS update see if this is actually an issue or if it's resolved with that BIOS update but that's the idea that's what's going on if anyone's curious as far as I understand it though it's not a huge issue if you are not doing memory overclocking but if you've had issues with it please feel free to leave a comment let me know because I'd be good for me to know about next comment is from Alex gray who says hi Steve you mentioned in previous videos that 18 inch MSI laptops are the same design as fan books is MSI buying fan books and up in the specs are the fan books tab stack them aside machines or are they just using identical third-party chassis what's the relationship between MSI and fan book fan book is a brand but or a model I should say by CyberPower CyberPower is a system integrator or an SI they compete with origin high power folks like that so they're the si si is often source their notebooks from somewhere else they don't generally actually make them so origin and cyber power and IRI power just as three kind of larger companies all use MSI for one of their suppliers for laptops so MSI makes the actual unit generally they do actually have supply the masalas 2 though by MSI is effectively the the OEM and is rebranded by CyberPower origin whomever the reason you might buy one or the other is because as I think is pointed out here is sometimes there's different specs that are only available through one partner and maybe not through the source and the other reason is add-ons so in the case of some of the SIS like cyber power and Iowa power and I think also origin blast I tested they don't really do all the bloatware that you'll get straight from MSI so there are some value-add to be had but that's the relationship basically MSI makes the units and CyberPower the other is buy it and sell it as their on model laptop Cleavon Sega are also really popular supplier or om is for notebooks next question brand new tracks do you get bored of reviewing products I mean like does it get stale until you get the next latest and greatest not really to be honest because it's such a crunch to get all of the production done and testing every time we are I'm normally writing scripts for videos and right up until we film it and then I'm writing the review for it as it's being edited so it's pretty fast paced I don't really get bored of them I would probably say I wouldn't want to review 410 50s for example that would maybe get a bit boring or at least uninteresting if not boring but there's almost always something you can find whether it's a really cool design or a really bad design or something like that so there's always something to look out for and that's the fun part the part that would not be fun is if you are running into a lot of issues with a product that are not easily explained or again with a low end like a ten fifty hundred dollar card the manufacturers you know they're cutting corners they have to it's a hundred dollar card so they're not gonna have that interesting cooling solutions but they can still have interesting flaws and that's always fun to look for so no I don't really get too bored of reviewing things I would say more overloaded worth reviewing things is more of the problem than bored next question I would Kurt says hi Steve a while ago the pump of my 980 I broke down after fiddling with it seemed to start working again about it roughly 50% speed I wanted to the RMA the card temp seemed fine pumps less audible basically what should I do so I don't know how you're reading your pump rpm because the hybrid cooler the pump is connected straight to the video card and there's as far as I'm aware no way to read the pump rpm maybe if you plugged it into the motherboard but you need something to adapt the connector because it's not the same fan connector as you'd find from a fan to a board so that I'm not sure where you're seen your speed maybe if you're going based on noise but I think what's more likely the case is I don't know maybe some air pocket got sucked through the the radiate of the pump and you had noise reduction after that one through if you have your radiator with the tubes mounted at the top flip it around that would help but if you think there's actually an issue I would definitely warranty it before it runs out because you don't want it to fail again later if you don't think there's an issue which it doesn't sound like there's necessarily an issue but yeah check the the radiator orientation I guess is probably what I would do first famous killer 5 says do you know or do you test your how do you test your GPU overclocks is vally enough I've read everywhere on the internet and you've set it to that synthetic benchmarks are not real-world scenarios meaning that some overclocks will pass some benchmark can be unstable in games how do you test yours so we test ours by I've shown this in the how to overclock your rx 480 video I normally launched the stress test for 3dmark and let it run while I tweak the clock settings and frequency voltage all that stuff and see if it fails during that process and it will eventually fail and then from that point dial it back to where it's stable in 3dmark and then start testing in games so I'll do a couple test passes in like shadow of Mordor The Witcher things like that where you get varying boost and load levels on the cards and then at that point after just a couple passes of those normally you've got a pretty good idea of if it's stable or not now ideally if you're a user applying the clock for effectively permanent use you might want to do an in-game benchmark like dirt rally or something and let it run overnight but that's the basics of it yeah I would not use just a synthetic test basically because it'll look a lot higher than what it's actually capable of often times and especially that's true with fur mark next question Josh Orenburg says is a big question you guys always say there's no such thing as a good or bad company like a good or bad product so to clarify that the reason I tell people tell writers the that phrase is because I don't want people writing in articles things like we trust company a company a is a good company because they everyone can screw up in the industry so products I go buy this product is good this product is bad not the company that makes it so that that stated that clarified the question is is there a koala variance in quality lifespan of a product like motherboard video card memory for different manufacturers or they all pretty much on level playing field can you trust all Strix cards for example or EVGA cards from any lineup any company to have similar build quality or lifespan among their line of of components so is it even possible to make accurate estimate of probable lifespan for each product type it's sort of not really um not for us that's that's more of a validation lab thing to do like something they would hopefully do internally where um TBF numbers not from media outlet though we don't have enough sample size but uh the question well except for things like pumps where you know the liquid will permeate the tubes within five years or something like that but the question there is variants and products they use different suppliers for their components you might have one video card just taking this if you have like Strix card versus a higher-end hall-of-fame or classified card or something like that just the supplier of the vrm alone will be different and impact how things play out and there's plenty of examples of where that's a good or a bad thing depending on what brand or what product you looking at motherboards there so much has moved to the CPU these days that they are less important than they used to be a lot of stuffs integrated a lot of controllers are integrated you don't need to solder them onto the board anymore but they're still variants it's still possible to buy a board that's more likely to fail then or more likely to overclock better than another board and that's generally dictated by things like the power design the components use pastors used I look just again on motherboards you might have killer networking versus Intel something that makes your life miserable versus Intel's networking solution I've had bad experiences with killer so yeah there are differences but there is definitely a point where something like memory which is more of a commodity at this point is is not quite as varied in what you're buying especially for gaming where it's not going to be as much of an impact or any way to your performance so that's the short of the the answer last question I'm amazed this channel isn't bigger after seeing subscriber count just join the ranks by the way what do you think about the NX so first of all stand up or shut up which is the name of the user by the way and not me same head to them the user's name just and appreciate up uh it is called an antenna switch okay it's not an axe anymore everyone knows that get with the program so my thoughts I think the Nintendo NX is pretty interesting will from a hardware perspective alone its Tegra I said it'll be Pascal will see I'm fairly positive that's right but it's been in development for a while some folks are suggesting Maxwell and videos text on their blog did say that the switch will be using something I don't know the exact quotes but paraphrase is something like the best or the latest architecture used in the world's most powerful gaming GPUs which is what NVIDIA says to me in Pascal so that was my assumption either way it'll have a Tegra SOC Pascal or Maxwell I'm leaning towards Pascal the the portability will be interesting from a battery life testing standpoint I think it's a very risky move and I think that's what Nintendo needs right now because they're getting a lot of pressure from PlayStation and Xbox in the traditional market and they hold the handheld market there's really not much out there for competition so merging the two means they can merge their development and hopefully have better output for the two types of player bases on one device with one set of coding so it's interesting we'll see how it actually works what the battery life is stuff like that but I've got more thoughts in the original NX / switch announcement specs video we did so check that out if you're curious or more patreon like a poster video helps out directly subscribe for more content it helps out a lot so do hit that subscribe button I'll see you all next time you thirty-four um 88 - pee