Gadgetory

everyone welcome to another ask Jian episode as always leave your comments in the questions section below kind of flip that around if you have questions leave them down there I'll get them next time if I can we're still gonna go forward and try the two episodes per week thing we're basically gonna try and eventually gap them about 12 to 15 hours apart and if it goes over well we'll keep doing it otherwise we'll go back to one so leave your questions if you have them this week we have a couple good ones on chipset differences which definitely need some clarification so I'll be talking about that before that this video is brought to you by thermal Grizzly's high-end thermal paste and liquid metal thermal Grizzly's cryo knot is an affordable high quality thermal compound that doesn't face some of the aging limitations of other pastes on the market cryo not has a thermal conductivity of 12.5 watts per meter Kelvin focuses on endurance is easy to spread and isn't electrically conductive making it safe to use on GPU dies thermal grizzly also makes conductor not liquid metal which we've used to drop 20 degrees off some temperatures than our dee-lighted tests buy a tube at the link in the description below quick update on the mod mats first if you haven't back ordered one you definitely should because we still have them in stock though we do tend to sell out of all of them right around when our orders come in that next order is already on the way here it'll be here before we leave for Computex which is next week from the area of this video so they will be shipping out around then it'll be about a week from now that they ship out to all of you then adding the time for your regions and I think they'll be in here with just enough time for me to autograph the ones that you all wanted to autograph so we should be good there if you haven't backward one though get it in because they're shipping out in about a week next our first question is from olivia turner who says what are the differences between z 370 and z 390 from a mainstream game or PC users and point of view well actually you can look at it from any user's point of view because e3 70 and 90 are really simple and it does I agree that it requires some clarification because I needed it to not long ago we talked about this I think in a news video but XIII 70 and 90 if you look at two block diagrams side-by-side which we did in that news video it's basically identical the only difference is the wireless AC support so Intel has integrated wireless AC support in the chipset now chips level rather than just a board level and that's on Z 390 other than that same lanes same everything else there's I think I know there are a couple more USB 3.1 options with Z 390 but that's not really relevant to most people because you're talking about like six versus ten or ten verses fourteen USB ports so does it really matter no not a whole lot it's not like you're getting more lanes for graphics which generally come off the cpu anyway you can pull them from the chipset but you don't really want to but you're not getting more Aegis i/o lanes high-speed i/o lanes and HS io lanes if you don't know are basically chipset lanes that the motherboard manufacturers to some extent can control so say you have 30 HS io lanes available to you made available via the chipset by Intel the motherboard manufacturer can decide we're gonna we have these ones that are required to be components X Y Z we're gonna pull the other we're gonna pull four of them and allocate them to a PCIe slot for M 2 devices we're gonna pull another 4 and allocate it to whatever Paul Cal for more and give it to Gigabit Ethernet and so forth so ancients IO is assignable by the motherboard manufacturer to some extent but you don't get more of those with 390 versus 370 basically it's the same chipset so in terms of functionality nothing is really changing in a meaningful way other than wireless AC which is not that meaningful for a lot of us for desktop use so you end up in a situation where probably the only reason you're getting a z3 90 board is to support the 8 core CPU is if they have new power requirements that the existing six core CPUs don't have and we don't really know exactly what's happening there yet I know that some of the higher-end Z 370 boards have V around that are still powerful enough for the eight core CPUs coming out but I don't know anything else beyond that that's really the extent of my information that said we'll see if anything comes about at Computex which again officially I think it starts June 4th or 5th and ends on the 9th typically there's some coverage before the show and a little bit after it so make sure you tune in and we'll see if we can find more on the new chipset sometimes it comes out whether or not Intel wants it to as for other chipsets it is a little bit kind of a confusing market right now with and the more or less having copied Intel's names to the point where Andy's launching a z4 90 I get why I am Dee did it that's fine I do find it annoying that even on the review side I can't keep straight which motherboard is from which vendor without looking at the socket anywhere that's pretty annoying but if you're talking like B 460 B 450 be 350 B 360 whatever there are a lot of it's very easy to mix them up now so if anyone's confused about the chipset differences between a chipset that has a 10-digit difference like 350 versus 360 the difference is one of them is for AMD and ones for Intel and you can't really cross compare them so you know you pick the one that fits with the CPU you're buying but as for intra company differences if you 7390 basically the same thing x4 70 X 370 more or less the same thing as well except a couple things we talked about in our 2700 X review if you're curious check the review I think I talked about it there not too much has changed though for any of these new high-end chipsets next question is from X Cleo who says how does increase in VR I'm switching frequency help overclocked in and what way is performance affected by increasing it higher is better question mark this is a setting if you're not familiar with it it's a BIOS setting and most higher-end at least motherboard BIOS options and I actually I posed this question to build Zoid because I haven't personally messed with VR I'm switching frequency too much and build Zoid said builds lloyd's from actually hardcore overclocking YouTube channel if you don't know it could check it out he says vrm efficiency goes down if you this is with increasing the switching frequency BRM efficiency goes down voltage regulation accuracy goes up and ripple goes down so that's pretty direct answer if you have voltage ripple concerns then it sounds like it's changing or increase in the switching frequency would assist with those regulation accuracy as he said goes up so you end up with probably a flatter voltage that's closer to what you want more or less and then you could validate it at the back of the socket if you really wanted to by checking one of the caps for the vrm the vcore vrm and yeah i think that that more or less answers there very directly though next question from nori who says what's the difference between dual channel and quad channel memory clearly there is some hardware difference check out these two sticks of course here are GB thirty-two hundred megahertz bottom is quad channel rise in specific per quarter actually there is no such thing as dual channel or quad channel memory there is dual channel configurations there are dual channel configurations for motherboards there quad channel configurations for motherboards it's actually determined at a platform level so technically it's incorrect to say dual channel memory but everyone will know what you mean if you say it if you buy a kit of four sticks of memory it could be quad channel memory if you wanted to call it that it's not really the accurate thing to say but you have four sticks of memory if you have a quad channel motherboard then you could put your four sticks of memory into a quad channel motherboard and call it I guess quad channel memory as long as the board and the platform support it if it doesn't support it if you're using like a z 370 or x3 70 board that's dual channel only then and you have four sticks then it's put into a dual channel configuration it could still be run as quad channel but you need a platform that supports it so the point is there's actually no such thing as dual channel memory or as quad channel memory even though yes new I will list things like that sometimes so if you ever are looking for we've had people ask in the past why's quad channel memory so expensive often it's because if you go to a retail site and select that you're looking for quad channel memory as they specify it you're gonna get a really limited list of things that are more expensive when you could search for just any kind of memory and as long as it's for sticks it'll do quad channel assuming that the motherboard has validated that supported obviously as as we discussed in another recent video stn video so actually that was a patreon asked Jian video so if you're not a patreon backer patreon.com slash cameras and access all tiers get the SP ends on the patreon upload it's it's basically a bonus episode of this and I talk about memory support actually on that episode so you can check that one out for more but anyway the point is there's no such thing as that that said I get your question so you can have things like dual ranked versus a single rank memory you can have basically more modules on a given PCB to achieve the same amount of capacity so your density might change per module and that does impact performance it is more sensitive on some platforms than others and that's I think what your question comes down to it's basically you're talking about how many modules are on it to achieve the same capacity which comes down to a question of the density per module in terms of gigabits how many gigabits per module on the stick of memory basically and then as noted some of them are dual sided versus single sided and that also has some impact on on how they branded or what its performance is or how it looks if you're looking if you take the cooler the heatsink off said yes it'll look different next question Xanadu man says will AMD be releasing a second generation threader for line 29 xx I'm looking to upgrade soon and I've had my eye on thread Ripper but if they plan on doing a second gen soon I'll wait they are doing a second gen as I'm aware the current plan is August I think that was the last update and they put out and I or I think they put it out I think it was them I don't know we weren't at the recent 2003 s launch event but from what I understand it was discussed there whether thread referred to would be coming out and the answer was probably August but I don't know if that's changed and I'll certainly let you all know when we find out so that's my understanding whether or not you want to wait that long is up to you generally the way I look at this should I buy or wait questions is are you happy with the PC you have now today or is there something really specifically that's making you unhappy with it to a point where you can't work that efficiently or you don't enjoy playing games on it if it's the latter case then just just upgrade now cuz if it's making you unhappy it's not worth waiting because you're talking often single-digit percentage changes it's not like thread report versus a fax it's threaded professor starter for two and there will be an improvement certainly if you are really core starved then maybe wait because I think thread Ripper two might have more cores but I'm not positive on that if you're not and you just want a high-end platform and you kind of need something today just buy it today if but yeah I mean just depends on how desperate you are next on Red Mage Cecil says for your mitx case testing do you think there's value in having a standardized expandable test set eg full-length GPU full-size tower cooler used in your standard case reviews that you can use to flex particular mitx cases that can accommodate larger hardware to give a more real-world representation of what the cases are capable of where do you think in muddy the data too much the answer is yes it it would be beneficial in some ways to have say a full length GPU that we could test in off camera maybe the SG 13 silverstone case we talked about the eternal take core v1 both of those can fit longer GPUs it be useful to test and I would agree that there's there are definitely limitations to testing Mini ITX cases with a standardized test set of hardware because as noted in the review there are so many variables specifically with many ITX cases that don't exist elsewhere that it's really easy to kind of get lost in a data set and not see that you know you could put for example in the v1 a bigger cooler on it so like with the testing we just posted you have the cry rig otaku where it only fits like a 50 millimeter cooler and you have the v1 which could fit a mini tower cool or maybe I don't know 90 millimeters or something like that I'm not really sure but it could certainly fit a bigger one so yeah the the v1 could perform much better in that regard but if we're trying to standardize the test and you can't cross compare them as everyone knows so if we had a bigger cooler on the v1 but not the Taku it's not really a proper comparison so there's definitely a methodological question of to what extent can you synthesize performance and validate box a versus B for cooling abilities or for airflow if ultimately you could put a better cooler and box B because it's bigger so we talked about that a bit whether or not would it be worth it like he said I would like to add more hardware I'd like to add a full-length GPU or different coolers for larger cases the problem is it costs a lot of money to do that not because the products cost money to get necessarily but because we already can invest several days in testing a case and I have to pay people to do that we pay to produce the videos I pay to pay Patrick to do a lot of the testing or building I have a time cost as well and once you factor all that stuff in we start losing money really fast if I'm testing say two different GPUs and two different coolers even those to do the different fan configurations we do it gets expensive fast we're not careful and go kind of crazy with it so depend on how popular the content is it can be worth it like if we if we revisit it into the standalone content piece of just here's a bunch of ITX cases that can fit full-length cards let's test only those and nothing else it might be worth it if we're not switching cards but if we're switching devices just I mean realistically I have to make money to keep a business going and it cost me a lot of money to do that you're talking multiple extra days of work because if we take our current test set and say it takes two to three days and we multiply it by two configurations it's a linear so easy math four to six days you add another cooler configuration now you're looking at like more than a week and you definitely in our position would lose money doing that so certainly as the channel grows we can continue to expand like that that's why we added ITX cases because we've grown to a point where I can support it now it's just it's gonna have to be an incremental thing I do want to get there but you know just be patient with me as a channel continues to pick up speed we'll continue to add stuff like that because we'll be able to afford it more or justify it from a business standpoint because ultimately I have to pay people so but I do agree it would be absolutely something I want to do it just a matter when can we afford to do it or justify it next question Doctor Who who says this question is about rammed training we have heard the term used many times what actually goes into Ram training is this a handshake between motherboard and RAM then reported back to BIOS or the chipset our multiple boot records of frequencies that worked stored by the motherboard so upon next boot it knows what should probably work this is we actually have a piece coming up somewhat soon on this question I'm going to read an excerpt from the scripts for that because it addresses it pretty directly this is written by Patrick Nathan who works for us on some of that stuff and Patrick wrote Ram doesn't set its own frequency and timings ultimately the BIOS does with SPD XMP a starting point on Intel boards this element of BIOS is called the MRC which means memory reference code this is one area where board manufacturers can secure an advantage in performance because the sheer variety of RAM and the difference is between Samsung micron and hynek's those are the chip makers not the memory manufacturers they the differences between them make tertiary timings hideously complex to adjust memory and board manufacturers can work together to bacon optimal timings for popular kits but for the most part there are determined if they are determined if left on auto at post so that answers a lot of your question where they should remain unchanged unless they're a boot failures so during training the integrated memory controller the IMC will attempt different timings it tries other things too there's more than just timings but it'll try to pin timings try to boot and it might fail so if it fails a lot of the times the solution to a boot failure related to memory is to wait and try it again and the board will cycle maybe three times and if you can't do it it'll go back into BIOS but if you feel like sitting there and pushing the retry button over and over some boards have a white reach high button on it you can often get it to boot with timings that it's trained that's what training is to work and from that point on you should be good it'll mostly stabilize things every now and then if it's really borderline you might end up with memory failure or memory errors not failure memory errors during a really memory intensive benchmark but if you're just doing like time aspire gaming it'll be okay to be doing anything super memory intensive and calculation intensive then obviously fix it but so yeah during training by EMC tries a variety of settings to stabilize the system not all of which are timings and it retries on each boot until it succeeds so if memory isn't stable with XMP just allow it to reboot a few times before giving up basically and we'll talk about this more soon next one there's only a couple left here Jason Blasi says my brother put conducting on his 580 and now it won't boot guessing he shorted it anything we can do to get it running again already tried cleaning it off and reapplying Arctic silver 5 instead yeah that's this is why we don't generally do liquid metal on a bare die for a GPU so you could do stuff like I mean first of all for your situation clean it like really thoroughly and not just the dye area but the whole board just liquid metal has a tendency to kinda jump around as you're cleaning it if anyone's ever pushed liquid metal around with like a q-tip or a paper towel you'll see that kind of balls up and it's really easy for it to just jump it'll kind of like almost almost in a magnetic fashion grab on to a nearby component so it's possible that there's just some that you missed I would say rubbing alcohol works the best hit the dye and everything around the dye be really careful with applying fingernails like if you try to put pressure around the die area to scrub stuff off and you're using your fingernails and you hit an SM like a small surface mount device SMD like a pastor it's really easy to rip those off so be careful but Robbie and alcohol should fix it use a q-tip where there's really tight areas that you can't really get with paper towel hopefully that fixes it you could try compressed air I guess anti-static compressed air should be fine I would prefer to just clean off with rubbing alcohol because then you know you're not blowing it around into other components but if you really have to you can try that as far as preventing this in the future probably use the last liquid metal and then assuming it's a horizontally mounted GPU just apply like a nail polish to the area around the GPU or if you don't want to do that you can use different types of tapes like captain tape which is a thermal tape and that'll protect it in case some of the liquid metal spills off the sides but otherwise yeah just clean it and be really careful make sure it's absolutely clean because you don't want to end up in a situation where it's it shorts something to a point that you fry a component if it's just short circuiting that like a direct short you can normally still fix it get it to boot just fine once it's clean is this the last one this is the US from the same guy Doctor Who doctor who says again I went back in time to ask another question this question concerns power supplies and maybe something you can integrate into your upcoming testing if it is in fact if it is in fact different depending on the power supply but who actually makes the power supplies are they designed in-house components delivered and built at a known Factory are they designed by the company but actually making the product is outsourced or are they completely outsourced for a specific unit is designed and built by and then the logo slapped onto it the answer is yes to basically all of those so the first question who makes the power supply's like most things like this liquid cooler this is made by ASA TAC technically they make the pump they assemble the radiator base but they work with NZXT to specify everything so in the case of a liquid cooler NZXT specified with a stack that they want this braided sleeve on it they NZXT built a custom PCB to go in here for RGB lighting all that stuff not a nun's exe ad just saying that's the differences that they do so ultimately the performance of this thing that is dictated by ASA Tech for the most part other than the fans which to some extent are made by the manufacturing partner power supplies are the same way so power supply manufacturers or suppliers include C Sonic channel while technology super flower there's a couple other ones great wall sir tech Flextronics so they're cut and some of those are much bigger than others but see Sonic's common one channel while technology is pretty popular as well and so it's super flower so there's kind of three of the bigger groups they are the suppliers for a lot of it and then the companies like Coursera EVGA etc will work with those companies see Sonic making their own kinda like Samsung makes their own memory devices they'll work with them to specify what they want Corsair does stuff like puts in special PFC chips and things like that so there's some level of customization certainly by the manufacturing partners it's made in a factory often owned by one of the suppliers I just listed but that doesn't mean that EVGA Corsair and all of them aren't making the power supply in the sense that they are still designing them and they saw the engineers working to solve specific problems that their users have and the suppliers work to the spec now it's also true that a supplier can have a platform just like AZ Tech has a Gen 5 platform offer it to the partners manufacturing partners and say this is what we got here are the few things that you can change on the list is maybe RGB LEDs or maybe the fan or some other small cables so if if it's a cheaper power supply there's a good chance they're working off of an existing platform and customizing very little if it's a high end power supply like an X 1600 I or something like that they've probably customized quite a lot just it really depends on what it is and then also finally to answer your last question do they design stuff and then or do they just slap their logo on it they're definitely power supply's where they just slap that logo on it just like there are with coolers to the for example is I'm the original Arctic coolers the whatever they called it liquid freeze or whatever they were some of those were basically just out of the box is tacky and it's with almost zero customization so you can have that too it just depends on the price the unit and we'll talk about more soon so that's it for this time as always there's a patreon special episode we'll have another ass Jen coming up soon if this isn't the first one it'll be a second one obviously go to patreon.com/scishow and XS to get the bonus episode go to store gamers nexus net to pick up a mod mat like this one or one of our 3d laser engraved logo cubes back here subscribe for more thanks for watching I'll see you all next time