Gadgetory

welcome back to harbor and boxed in today's video we're recapping what happened in our 48-hour NVIDIA GeForce r-tx 20 atti founders Asian stress test which a surprising number of you guys watched for some reason you wanted to see a boring screen of a benchmark looping for hours on end a lot of you were hoping to see a die or catch fire midstream but at least a healthy bunch of you enjoyed the Q&A we ran at the start of the stream for a couple of hours so a quick recap of what we were doing in the stream and why we did it so basically we ran our 20 atti founders edition card and stock clocks in a unit in heaven loop for 48 hours to stress test it and see if the card would fail we wanted to test this after reports that hard OCP had a fair resolution card died under similar circumstances our card showed no signs of failing up until this point so no Space Invaders artifacts or any other issues but there was a slim chance that running at high temperatures for a long period of time would cause a failure due to one of the components degrading again I'll setup had the card running in stock a lot of people were questioning why the card wasn't overclocked but that really wasn't the point of the test we merely wanted to see if the card would die under standard operating conditions we know cards can die when overclocked the interest was whether the card would die in and out of the box type configuration the card was in our core i7 8700 k test rig with the cpu also not overcooked and fan speeds were set to order anyway at this point you're probably wondering did the card die and the outs of people that understandably couldn't be bothered watching until the end of the stream is not sorry to have people who helped want the card to die it was very unlikely to happen and in the end it just didn't happen there were no artifacts spotted in the entire 48 hours of live streaming and the card still functioned perfectly after rebooting the system in fact it's running here it's all fine no problems this was despite the cards sitting at around 84 degrees Celsius for a full 48 hours except for one minor interruption with backplate temperatures measured through a thermocouple comfortably sitting in the 70 to 75 degrees Celsius range that's certainly quite for a GPU but not hot enough to get it to malfunction or die the card not dying does present a bunch of interesting discussion points though neither Steve no I expected the card to die during the test in fact we mentioned we'd be shocked if something broke under these conditions but it seems a number of people in the chat wander the impression that there was a very good chance this sort of test could kill a card like the RT X xx atti so let's talk about that firstly despite the negative press surrounding the RT X xx hei and its various failures not every card is destined to fail even though our card worshipped to us before launch and presumably came from an early batch before any issues were identified it was never a situation where you buy one these cards and you're guaranteed to receive a dud everything we've heard so far still suggests the failures have been a minority of card shipped a concerning minority for sure considering the amount of people reporting issues which seems higher than past flagship GP launches but a minority nonetheless in our case the RT X xx atti in videos shipped us to review is a good unit that works perfectly fine under standard operating conditions I'm sure there are plenty of other people out there in a similar position they've bought a founders edition model it arrived and luckily it works without any problems especially if you don't see any space invaders artifacts after a few hours it's highly likely the card is a good one and will remain functional for years so point number one is simply that not every RT X 28 Ajay founders edition all died simply because it's an RT X 28 I found his Edition point number two is that something I guess that cropped up numerous times in the chat as a bit of a misconception that running a graphics card at hot temperatures will cause it to fail even if it's stock clocks which is not the case our card ran for the entire 48 hours between 84 and 86 degrees Celsius on the GPU and a scorching 70 to 75 degrees Celsius on the back light from ambient room temperatures between 19 and 25 degrees Celsius and reasonably slow case fan speed set to their silent mode the backplate reaching the mid seventies is certainly far too hot to touch and GPU temperatures in the mid 80s isn't great you normally want the card running a bit cooler to get the most out of invidious GPU boost algorithm but simply running at hot temperatures like this isn't going to cause the card to die unless you modify the card settings through overclocking or other me in fact the card is actually designed to be run around the 84 to 86 degrees Celsius mark it's the point at which the cards temperature protection algorithms kick in and begin reducing clock speeds to prevent any higher temperatures it's all a full thermal throttle like you might see on a CPU hitting t.j.maxx but NVIDIA GPUs begin gently clocking down or at least not boosting as high around that 86 degrees point no doubt the cards can actually run hotter and in fact do if you up the thermal limit in your overclocking software this is a basic fact about modern hardware design if you don't touch anything a component running under stock conditions will manage itself to prevent damage GPUs are one such component where this is true but you also see it across CPUs motherboards and all sorts of other PC hardware of course there are lots of reasons why running a GPU at a cooler temperature is better you get less power leakage so the card will run more efficiently and waste less power at lower temperatures you'll have more overclocking Headroom because of this and in the case of NVIDIA GPUs their GPU boost algorithm will push the cars higher when temperatures are lower with even something like a 5 degree difference allowing for higher clocks cards running cooler will also degrade slower we're not talking about short-term degradation here though we're talking about after years and years of use you might get a bit of extra life out of a card that's run cooler than one running near its thermal limits again GPUs are designed to be run near the limit for years on end though so simply running a card at hot temperatures isn't gonna cause an immediate failure and this is clearly the case without our text when a heavy ti here running the card at high temperatures for 48 hours well basically did nothing for it short term usability still works fine in fact nothing has really changed compared to when we first use the card will this test impact the 2082 ice overall lifespan well probably not in any significant way and it certainly be too early to tell if it did the final point is simply that yes you can run a card at a hundred percent utilization for days on end and run into zero issues miners of course have already proved this but a few people seem to think that you need to let the card rest and cool down periodically as a typical gamer would between gaming sessions but GPUs are designed to operate continuously and there's really no magic timer dictating how long you can use it before it needs a rest the good news for us is that our art x20 atti founders Edition has now essentially been validated to work perfectly so we can continue using it without any fear of immediate death or needing to go through an RMA process it also shows to you guys that yes if you buy an art x20 atti with founders edition PCB there's a realistic and highly probable chance that it will work just as well as any other GPU you've owned in the past at the same time this doesn't really give any additional insight into why some art X 2082 eyes have been failing and we didn't expect the test to provide any additional information anyway one card you know it's never going to be enough of a sample size to generate any sort of concrete information and it still seems at this point we really don't know what's going on we certainly don't expect Nvidia to provide us with a reason at least publicly if you are wondering whether your and Vidya r-tx card is fine we recommend running a stress test like this for around 12 hours the art X 2080 the didone are started to show artifacting around six hours in so if it passes a half-day test it should be fine from that point onward you shouldn't have to worry about the card Ram we die at least you shouldn't have to worry about that any more than with any other GPU and that just about wraps up this live stream experiment we had a bit of fun chatting with you guys live and hopefully we'll be doing more live streams in the future considering how well this one went and of course we'll continue to have our monthly live streams with both myself and Steve exclusive for our patreon members so if you want that more casual type of stream where we just chat about cool stuff in tech sign up to our patreon and you'll get access to our exclusive discord chat while you're at it time to switch off the test system and give the artex 2082 eye a bit of a rest I said it wasn't actually necessary but my power bill will thank me later that's for sure because it is subscribing for more fun or in this case boring content straight in your inbox I'll catch you in the next one