Gadgetory

everyone welcome back to another episode of ask GN I think we have some good questions this week but I can't I can't presently see them so you just have to trust me pretty sure one's about Nvidia there might also be one about Intel and AMD to give you a bit of a preview I guess if you have questions for next week leave them in the comment section below or if you're on the patreon discord you can post them to ask GN chat and also she'sshe's supposed to do that out of the box they're almost saying before that this video is brought to you by thermal Grizzly makers of the conductor hot liquid metal that we recently used to drop 20 degrees off of our coffee leak temperatures thermal grizzly also makes traditional thermal compounds we use on top of the IHS like cryo not and hydro not pastes learn more at the link below so couple things starting out first of all and ext saw our hashtag please puck from last week where we basically jokingly although some of you didn't didn't seem to clipping on that jokingly said tweeted NZXT and asked him for a free puck now I guess it was a half joke because they have given out free pucks in the past but no one was really expecting like 500 tweets so they're not gonna give away 500 sorry guys if you got one awesome but if not they started a contest from it because they were like hey your please Park hashtag has 6 million impressions on Twitter which is more than any of your single videos has ever gotten so we should give some away so yeah I I don't know the full details go to NZXT Twitter page all I know is that we will be giving away a small box of Gian March with their prize so I think the first place person for whatever the contest guidelines are will get I'm pretty sure a puck and a puck gamers nexus shirt which I have worn previously and also a small box of our merch so I think they'll probably be like a shirt sticker or something like that let us know what your shirt size is if you win but yeah so check that out I guess didn't really expect that to happen also very briefly before the first question our mod mats have a date kind of on the next restock so the first run sold out completely ran out of them we took all the feedback we mostly been that we need to order more of them and those should be exiting production sometime I think we're planning to be able to ship them by end of March to early April that's the current timeline for it so that's after shipping production all that stuff we should have more available if you've been interested in one store documents Nexus dotnet slash mod matte to backorder one will have probably enough to get everyone who back orders one in the first shipment unless it goes absolutely crazy but I think we've ordered enough to prepare for that so yeah check that out thank you for supporting it but let's get into the first question so the first question is from Brom who says with the rumor of new NVIDIA GPUs being announced end of March at GTC what are your expectations for the new GPUs regarding crypto do you expect a little availability hiked prices things like that first of all I'm not sure that those rumors are substantiated by anything for it GTC I don't I'm not sure where they originated but I have no idea if a card is coming out at GGC sometimes they do sometimes they don't the last few years it's been more science community driven than it has been gamer driven I would I don't know I I'm not expecting a card for gaming at GTC I'm really not it could happen and that'd be great and we'd all be happy for it but it's not presently something that I'm expecting given the state of the current market now as for availability and I hope I'm wrong on that but it will see there'll be a new card at some point obviously I just I don't think it's gonna be that soon but anyway availability I typically with with the Pascal and video launch there was a lot of availability at launch so if they follow that previous trend and if they're in production now there should be a lot of cards available and I guess it'll depend on how good they are at crypto mining but I don't I I know what NVIDIA ships for volume with Pascal and previous architectures and it's a freaking lot of cards so really unless the the demand is that insane I don't think it'll be impossible to get one like the first day or two or three they might sell out after that that's not too atypical these days but I don't think it's gonna be as limited as the cards are today I certainly don't think that'll be the problem at least not at an initial launch and I really hope it won't be the problem as for high prices I'm assuming the question here is is basically if I may rephrase it is basically do you think that the current price ladder or hierarchy of GPUs will increase with the next generation in other words do I think that a $700 10 80 TI will become $1,500 1180 TI - whatever they call it there hasn't been a lot of price creep in GPUs over the last few years other than what's going on right now but as far as MSRP it hasn't really crept up that much so if there's an increase I would hope that it's inside of a range of 100 to 150 dollars and I don't know if NVIDIA will do something like they did with a 1080 launch where they had a higher priced reference card than partner cards but they really backpedal on that so I think that might remain one MSRP for partner and reference cards as for high prices though the only thing I really foresee changing the prices in a really solid like reliable definitely will change the prices way is the memory prices less than the crypto miners I really don't think crypto miners are going to impact the price for MSRP all that much I think more likely what's gonna happen is the 20 to 30 dollar increase for 8 gigabytes of memory will increase MSRP that's probably what we'll see more than a crypto impact and there might be a crypto impact I don't know I it could be that the next generation is completely insane with mine if it's anything like Volta it'll be decent but it's not it's not gonna now can I flip the industry on its head or anything for mine so but yeah at this point all I can tell you with any level of confidence is opinions and my opinions for your questions are well this one wasn't a question but with the rumor of new NVIDIA GPUs being announced at end of March at GTCC I don't think there will be a new consumer GPU I hope I'm wrong if there is a new GP announced I think it's probably going to be data center enterprise or science driven that is not fact I don't have any tips from anyone that's my opinion the next one expectations for GPS regarding crypto we can assume that if the next generation is even remotely close to a mix of Pascal and Volta we can kind of extrapolate the performance there's already Volta mining benchmarks there's Pascal mining benchmarks I wouldn't expect the next gaming architecture to do better than either of those two do you expect a little availability no and do you expect hiked prices yes insofar as memory cost has gone up and NVIDIA might or AMD or anyone might capitalize on the shortage a bit and increase the price a little bit but I don't expect a 2x cost flagship GPU and if there is one I really hope there isn't if there is one that'll be big news and we'll certainly be talking about it but I really don't expect a 2x cost on that that's that's crazy that's a gigantic jump and I've seen some reports of that so that's not what I expect but we'll see this next question is sacrum Goyal who says should you deal it a lock to CPU so you can reduce the heat and keep the CPI maximum turbo clock say for example deal it in a high-five 8400 and use a better thermal compound and keep the turbo at 4 gigahertz locked CPUs for Intel like the i5 8400 or you can look at the turbo tables for them and basically the lock comes down - it's not a case queue they don't let you over clock it pretty much end of story it does not have to do with thermals the lock is limiting your multiplier adjustment to the maximum single core turbo of that processor so what you could do deal it to answer the question immediately no I don't think it's worth deleting an i-5 8400 at all unless you you really need it for some reason and further if you wanted to push that max turbo clock all the time by a motherboard that will allow you to adjust the multiplier and it won't let you adjust it past the max single core turbo but you can find one that will let you adjust multiplier so let's say the multiplier is something I don't I don't know top my head with single cores but let's say it's 44 or 45 X you could apply 44 or 45 in this case to all cores all the time through bios with a motherboard that supports it which is most of them so that would be the best way to do what you want you go into BIOS you set the multiplier to the maximum number that it permits you to do and assuming you have a reasonable non Intel stock cooler on their thermal should really not be a concern and if they are them there's a bigger problem with the case or the airflow or the CPU cooler so that's how you would you would get I guess you're saying here you're going for 4 gigahertz if that's what that turbo is for the 8400 then you just set it in BIOS to 40 and tell it to apply to all cores and you've pretty much solved your problem so yeah and also if you're curious you can open up XTU it's an Intel utility and they have the turbo tables in there so you can reference them I think it's on the right side of the screen and see what one core to core 3 core for core turbo is because contrary to popular belief I certainly don't have it memorized and it's slightly different for different bins of CPUs so you can it's a really good utility I would recommend it if you don't have it on your system and you're running an Intel processor it's cool just to see what information it offers because it's like it offers a lot of stuff like hardware info does less kind of information overall but it offers the more critical aspect like are you power throttling throttling current limit issues things like that it'll tell you in big yellow letters they can fix it in BIOS next one is motorsport who says Steve why don't you use a smoke machine or incense to show the airflow in cases think people would really like to see some claims about how air is moving realized physically we've done that so we've done smoke tests in the past I used smoke pellets for it and I'm sure we have footage somewhere that's years old of the GoPro shot I did I like 120 FPS at the time that was fast but so the problem with smoke testing with cases it's a really cool concept on the surface and we did a pretty good job of it but ultimately you really only have a couple of frames even with a reasonably high speed like 240 FPS because we're not gonna go to a thousand or something so even at a reasonably high 240 FPS you still only have a couple of seconds where you can actually see the air channels forming once you start blowing smoke into the system with the method we use there's there might be other ways to do it I used smoke pellets I'll explain why in a moment but with that particular method which I did at the time think was the best option the smoke would fill the system so fast that you could kind of see a channel forming at the beginning and then that was it so you can't like 5 to 10 seconds after slowing it down to 30 or 60 FPS where you can see the channel forming it's not that exciting it was a lot of work to set up it took me like probably two hours to set up each time I did it originally I'm sure I could do it faster now but it was pretty disappointing for me because I was expecting more and I think we could do it better today than back then and I don't know it might be worth trying again but smoke testing is it's without like a proper chamber to do it it's really hard to get the effect that everyone wants and there are also some pretty important side effects of it so if you use a fog machine you're now leaving really residue on all the components you could electrically short something and the components that oil doesn't really ever come off speaking from experience like it you you can kind of get it off but it's always gonna be there to some extent and we'd basically need a special smoke testing rig that were okay with damaging or killing components through electrical shorts things like that you could use something that doesn't have water vapor in it because fog machines do so you use something that doesn't have water vapor that'd be a smoke pellet I guess you could probably find like really anything you could burn and use that but a smoke pellet is a good option and it's literally a pellet that you light on fire and it puts off a huge plume of smoke that's a good way of doing it without the water vapor problem you still have smoke going all through the computer having done this it smells like smoke basically forever afterwards all the components the case all of it like you can get I left a case airing out outside for probably a month before the smoke was mostly the smell is mostly gone and the components same thing so it's like we could do it but you have to set up a special secondary bench that you're okay sacrificing I have ideas to get around it I don't know if we'll ever do it but yeah I've tried it a few different ways and best case you get some smoke in there you get a few frames where you can see the air flow patterns and that's pretty much it we've found that through objective testing by populating different fan slots we've used tape a lot you may have noticed we tape off a lot of panels and things and do all kinds of extra tests we found that that approach to case testing is better to understanding the airflow patterns than smoke it can be done with objective thermal measurements we can tape off different corridors of the case and see how it impacts the temperature and then through just deduction we can look at that data and say okay well we added a fan to the top but the CPU temperature went up why how does that make sense we added a fan its intake its front of the CPU and this is something that you'll see soon has happened the answer in that specific scenario was be found by taping off the top of the case and the answer was when the case was when we added a fan it changed the pressure dynamic of the case such that the CPU cooler the tower cooler was no longer pulling air in through the top even though there's no fans on the top the pressure would allow the air to just sort of slowly draft in and adding a fan to the front made it so that the air was actually going out the top instead of in and more of it was escaping out of the top before it got to the CPU then got to the CPU so the point is we've found generally that that kind of a testing is better to understanding the pressure and airflow situation than smoke I would love to do more with smoke because it's cool to visualize it but it's very difficult to get it the way I'm happy with so yeah that's that's pretty much that one the next one is from Robert Davidson Robert said I constantly see people saying that games don't use more than four cords on a CPU but I have a 1950 ex and noticed that several games use all 32 threads to those games benefit from high core count and if not will they in the future will the 1950 X ever be more value in a game at 3.7 gigahertz than eight threads at five gigahertz we kind of talked about this a little bit in our threader for review I don't think we went into as much depth then as I would now I understand it more now so some games Final Fantasy is a great example Final Fantasy 15 is one of the few games where when you launch it from memory at least with a benchmark it would show every thread more or less semi populated and the population of those the utilization which say of those threads would kind of decrease as you add more threads but it would scale across all of them which looks cool from from like a utilization standpoint but in some games we've tested we've found that it's better to have a higher percentage utilization on fewer threads than a lower utilization per thread across a lot of threads and we've demonstrated this with Rison and thread Ripper alike these are good comparisons because they're in the same architecture family and it's a core account change there's some cash some stuff like that frequency change but we've demonstrated it there and at least one of our reviews can't mention can't name it off stop my head we've demonstrated with the Intel family CPUs and the high core count CPUs so where you see like a seventy nine hundred x79 eight exe in a lot of games they'll underperform versus an eighty seven hundred K just like a thread Ripper CPU can underperform vs. horizon seven or five CPU it comes down to game programming some games will leverage the threads better from memory watchdogs is certainly one of them that we've tested where we would see better scaling on thread Ripper and high core count rise in CPUs versus the lower court count rise in CPUs or better scaling on HED T Intel CPUs which is the standard desktop ones so there are games that do it but you also have games like civilization for example where you would assume thread count would matter a lot because civilizations are really AI CPU intensive game it spins off threads for AI processing but it's ultimately in that specific game spinning off one thread for AI processing and the other threads are much more minor you'll have maybe a render thread that deals with your draw calls you might have a like I not physics in that game but you might have some kind of audio thread and what happens is because you're doing most of your work on one thread typically render maybe AI in this particular instance that thread just needs to be fast and it doesn't matter really anything a house so that's an instance where something like 5 gigahertz on fewer threads might be better but again watchdogs to shows the opposite so it depends on the game to answer your question more directly let's see well in 1950 X ever be more value in a game than eight threads the answer is yes but it depends on the game and certainly we wouldn't recommend thread refer for gaming just like we wouldn't recommend even a 7900 extra game which is a cheap H ADT CPU we'd recommend you know the $300 parts from AMD or Intel because those are more suitable for it right now but you can play games on them and there are a few game that will leverage those threads but you've really got to have another use for it yeah so the answer though why do threads scale this way look at something like CryEngine try engine has a while ago one of the last revisions I paid attention to before they lost their minds at Crytek CryEngine at the time when I was looking at it could do eight threads it was the first engine that natively supported spawning of or saturating eight threads on a CPU and so what you could do is as developer you could say I'm gonna push rendering off to this one physics here hey I hear something else too one of the other threads and you're ultimately going to be bottlenecked by your most heavily loaded or burdened thread which is probably render and if if you're falling behind on that one thread and it's not a really asynchronous pipeline you're gonna fall behind on all of them because you're you're stuck stuck processing whatever the most there is to do on one thread which is probably rendering so we have a really good interview with Sean Tracy from the star citizen team where he talks about thread saturation on CPUs from a high-end game perspective go to our channel go to playlist I think there's a star citizen interviews playlist where like 40 of them with them and search for Sean Tracy tra see why it's a great interview probably one of the best we've ever done with a game developer and it explains a lot next one dubs says you mentioned that certain brands of motherboards will provide higher or lower voltage than what is actually reported by software is it really brand X or Y offsetting voltage or just the way the components are tested so some of this can come down to how the motherboards built not necessarily by the brand but specifically what I think I was probably referring to there is some boards like gigabyte boards will be a little bit more generous with their voltage provision than others we saw this with KB Lake when it launched so we saw this with coffee lake when it launched where auto voltage on either one of those platforms would result in voltage that was way too high you'd have a 90 degrees Celsius CPU and you could drop the voltage of something like one point one eight one point to something and sustain sixty degrees instead of 90 and it was still be perfectly stable so that's probably I was referring to there are auto voltage tables there LLC tables those things change between brands and that is more of a brand thing because it tends to apply on multiple C or multiple motherboards from that family razor says when will side panel fans make a comeback and try in a number of cases I've noticed vrm temperatures are consistently hotter by a small amount of cases which the left side panel fans does fractal have any designs making a side panel for the r6 I have no idea about fractals plans side panel fans making to come back stay tuned subscribe because we have something coming up soon on that and I'll leave that there because the rest will be answered in the content next question Keith bérenger says if a viewer had an older piece of equipment they'd like to seek you compared to a more modern piece would you consider it as a retrospective piece of content or do you prefer to work with modern components that have obviously more relevant metrics and to what people search for YouTube in Google this is a good question we do pretty well with with retrospectives and with post-mortems of things because we did this with a 2600 K in the 2500 K the i7 930 a bunch of old CP is the Phenom 2's we did last year and that was our ancient so was the 930 and the content does reasonably the reason that content does reasonably in those cases is largely because a lot of people still have this or had those CPUs rise and just launched coffee leak was coming up everyone wanted to know if they should upgrade or not so they worked well there we always will consider looking back at an old piece of hardware we just look back at the Raven oh - that's a case it's not a CPU just because it would be kind of fun so there are times when we'll do it just cause it's fun most of the time obviously hopefully it's relevant to people like the Sandy Bridge look back the Phenom look back because we do need to get traffic to actually make money on it otherwise lose money so it depends on what the component is you know transparently if if we're gonna lose money I'm looking at it then it's pretty hard to justify but sometimes it's worth doing anyway because there could be just some satisfaction and looking at for example the Raven o2 and seeing how it does today just cuz we really liked that case so that's an instance where I'd be willing to lose a couple bucks just to see how it does today but we can't do that all the time so it does depend on what the component is but let me know if you have an idea of something you want us to look at because we certainly consider it I think this might be this the last one I think this is the last one so kada lists said apologies if this has been addressed previously but I was actually wondering if I could ask about the people who aren't on camera right now I remembered names being thrown around like Keegan and Andrew and I was wondering who exactly are all the behind the scenes people IGN could you describe their jobs or roles at GN and one or two sentences for each person so somewhat chronologically Keegan has been working with me the longest he travels to a lot of the trade shows and conventions with me we do he does a some film and editing work at trade shows primarily and just in CES with us he's done copy text with me twice now in Taiwan so we're pretty used to traveling and working together his job primarily those events has been camera and editing work and lately he's been doing a lot more logistics help for me management on the like keeping schedule side of things Andrew you mentioned as well Andrew does all of the day-to-day video editing and rendering or filming editing he has does work on 3d models that we use so if you've seen obviously the intro to this video or any of the other ones and you did that through blender and basically does day-to-day video work and also product design stuff like that Patrick assists in and at this point leads a lot of the case testing so he's more or less autonomously doing case analysis and then I help as a test lead and we look at a case and say how do we want to test this and then he takes that test plan and goes and does it and he writes the review so we have the website gamers nexus not Annette Patrick's written version of the video that we produce goes up there he fuels a lot of the the build and appearances sections of case reviews he also programmed our recent a console benchmarking tool and then we've also got we have some some additional assistance as well outside of day to day operation Ryan helps with some business development stuff and Eric Hamilton helps with some news writing and that's most of the current core team for GM so questions for next week leave them in the comment section below again the mod map we should have them coming in and going out the door immediately end of March early April so get your backorder and if you're interested we've taken feedback I think we have some improvements in the pipe as well and future things will be talking about later at a later date but I subscribe for more Zoids patreon.com slash Cameron's next topside directly we have a separate patreon as Jen going up next and that's all for this one I'll see you all next time is it recording