CHEAP does NOT mean GOOD VALUE - Budget GPUs Explained

it's easy to look at a GPU that costs a thousand dollars and think all of that for games what a bad deal but what if I told you that cheap GPU next to it is even worse don't buy it I'm not shaming anyone for being on a budget I've been there which is why I know that the less you have the more important it is to shop smart and this video is going to arm you with everything that you need to avoid the pitfalls of the extreme budget segment but without blowing up your budget the best part is that you can use what you're about to learn for years to come because these principles have been valid for as long as this channel has existed and it doesn't look like they're going to change anytime soon like my cringe Segways to sponsors like Zoho one Zoho one brings together sales marketing Finance analytics and support in a single unified platform to run your entire business try Zoho won today for 30 days using the link below [Music] let's begin with a couple of good reasons to buy a basic GPU let's say your PC dies the night before a deadline and grabbing the cheapest thing on the Shelf at Best Buy is your only way to get back up and running I would do the same thing or maybe you're hacking together a low power Plex box or a utility PC from old Hardware if you only need some form of video output budget gpus can end up being your only option thanks to their compact size and low power consumption but if your main consideration is the cost these good reasons start to disappear very quickly because now what matters is the value of what you're buying which for gamers fundamentally boils down to one thing for every dollar I am spending how much performance am I getting and typically for a GPU that's measured in fps or the number of frames that it can render per second for many of you watching this channel this concept might seem downright intuitive but people who aren't enthusiasts might not have the context to interpret a graph like this and they shouldn't be shamed for that I mean I don't know everything there is to know about tires for example and think about it if you spent your whole life gaming on consoles compatibility is binary it runs or it doesn't run you've never had to think about how well it runs and how that might vary depending on the rest of your setup and the games that you're playing making matters worse for new buyers it's not obvious whether a given graphics card on the Shelf will deliver an acceptable gaming experience and I lay a lot of the blame for this firmly at Nvidia and amd's feed both use naming schemes that work a little something like this your first number or numbers is the generation of the product the next one is its performance class and the final is a placeholder that sometimes gets used to denote an in-betweener product you've got your suffixes like XT or super that typically denote a more better version of what is fundamentally the same thing the issue is that neither of them treats their own conventions as anything other than rough guidelines and both of them have shown that they're willing to break these conventions outright in the interest of making a subpar product seem more desirable take the GT 1030 and GTX 1630 from Nvidia for example from the numbering alone we might conclude that while they're in a similar class there should be a huge performance difference between these two cards and there is but it's not as much as you might expect while the name suggests that the 1630 should be six whole Generations newer there's actually only one generation separating these two cards and even within its own generation I consider the 10 30s naming to be very misleading consider this a 1070 is roughly 80 percent as fast as a 1080. a 1060 roughly 70 percent as fast as a 1070 which kind of makes sense so far right but then a 1050 is only half the speed of a 1060 and a 10 30 Falls to just half the speed of that making matters more complicated Nvidia has made headlines multiple times in recent years in recent days actually for releasing gpus that use the same numbering scheme but with downgraded memory configurations that are catastrophic for performance as an innocent victim I mean Shopper how am I supposed to infer that higher performance tiers have large price gaps and small performance gaps while lower performance tiers have small price gaps with big performance gaps even as someone who does understand the numbers I would be totally lost here which is why the Savvy Shopper then always comes back to frames per second let's talk about how to read a graph like this this performance graph for in-game performance it's the result of data collected over the course of Benchmark and the bars you see here indicate how smooth the overall display is each item is split into three bars here average five percent low and one percent low frame rates averages exactly that the average performance but it doesn't tell you the whole story because how fast the game runs at any given moment differs depending on what's happening on screen that's where the lows come in and these represent the lowest frame rates that were recorded five percent and one percent of the time respectively which paints a more complete picture of the performance that you can expect from scene to scene what frame rate is acceptable or playable it is very subjective but generally speaking Gamers often want to see 60 or higher on these minimums in order to ensure animations and motion is as smooth as possible on the typical consumer displays that can display a new frame 60 times per second even then a high average frame rate with low minimums indicates a wide swing in performance that will be very noticeable during gameplay unless both are well above 60. console games will often lock their frame rate to 30 if they can't reach 60 to ensure those swings don't happen and while that's often controversial for current gen owners it crucially ensures that things feel consistent from frame to frame but don't be fooled into thinking that you'll exactly match a gpu's performance shown in graphs like these the system is installed in will have a massive impact on the level of performance and this phenomenon is often referred to as bottlenecking where the maximum performance is held back by the weakest link this is why you have to pay attention to the system that a Reviewer is using to test the GPU one piece of information you might notice is missing is the price and that's something that we've tried and unfortunately failed to incorporate into our graphs in the past the reason is that the actual price is often going to be very different from what the MSRP is especially laid into a product's life cycle and the used price is going to change the calculus even further and is not consistent but performance mostly is this is where it's up to you to check the current market conditions and weigh the performance data against the costs today it's also up to you to make sure that your performance data is as accurate as possible by using multiple different reviews well no one wants to make an incomplete review testing for reviews are often done in the span of a week or less and none can be truly exhaustive that means that each of us has to prioritize our testing and we'll all have different priorities some reviewers test different games at different settings or compare different gpus and this is crucial for understanding the complete picture plus sometimes reviewers mess up that's why you should always check multiple reviews before making a decision now that we're familiar with how to interpret these numbers there are a couple of things that are going to stand out first and foremost is that the blank tees on lttstore.com are great quality and very reasonably priced second is that like I've been saying for years it should be illegal to call a low end GPU value or budget because it's people on a budget who can least afford to have their money wasted and these things are just that a complete waste of money across the board but just pointing out problems isn't useful what are the solutions I'm glad you asked the first one is the kind of thing that makes you kind of scratch your head and go really instead of buying an entry-level GPU like one of these buy the next tier up for almost the same amount yeah you heard me look at the pricing for the GeForce 1630 while you could treat yourself to a polished turd in a box of KFC the 20 price Gap to get double the performance with a 1650 seems like a no-brainer especially if you consider that even though both of these cards might run your games today the 1650 is going to last you a lot longer as new games come out the 30 Delta on the AMD side of things is a bit steeper but we've also seen promotions that have brought the 6500 XT down to literally the same price as a 6400 and what goes down almost always comes down again option number two is a little bit outside the box but can be an effective way to upgrade not only your GPU but actually your whole system without spending extra money let's say you're running a 4th Gen Intel CPU from about 10 years ago working motherboards from that generation are worth more than you'd think selling for anywhere from 60 to a hundred dollars a fourth gen Core i5 is worth about 15 bucks and eight gigs of DDR3 is worth anywhere from 15 to 20 dollars as a kit you might even be able to do a little bit better so at minimum you're going to be left with ninety dollars to spend alongside your hundred and fifty dollar GPU budget well guess what you can actually upgrade to a ryzen 5 5600g with 16 gigs of RAM and an a520 motherboard for about two hundred and forty dollars that gives you six much faster CPU cores double the RAM and an integrated GPU that guess what will perform on par with that GT 1030 that you could have bought you're also getting modern features like extra USB 3 ports m.2 storage support and the ability to run Windows 11 without any hacks I mean who doesn't like a fun little Facebook Marketplace Adventure it's kind of like your own little scrap yard Wars option number three though is my personal favorite why do you want a brand new card sealed in a box when a much better one is a bus ride and a quick dust cleaning away A good rule of thumb is that for any current generation GPU there's probably going to be a previous gen one that is either faster for the same price or cheaper for the same performance for example the 1630 is priced at about 150 to 190 US dollars for that kind of money you could get a GTX 1080. I mean sure it's from 2016 and it will consume more power while you're gaming but it runs at least twice as fast and there are other options why don't you save some money on an RX 590 from AMD that is a great performing card from 2018 that regularly sells for under a hundred bucks option number three is also great if you don't care about gaming performance at all since if all you're after is a basic working GPU so you can output video those can be had for so little that even if they die you could just buy another one and still have money left over compared to buying a brand new card on that note I get it not everyone is comfortable giving up a warranty but I would make the argument that much of that risk can be mitigated by using platforms like eBay with buyer protection or doing what I used to do and testing products when you meet up to buy them with the amount I've saved over the years buying secondhand Hardware it has more than covered me for the odd failure which is all fine and good that we have these options but there are still a couple big questions remaining like why do these crappy products exist at all and why don't AMD and Nvidia just price them fairly then do they hate poor people let's address that one first it comes down to the steps involved in building a card and getting it to your door every GPU no matter how low end or high end has some fixed costs the gold in the contacts the copper and the PCB the PCI bracket assembly HDMI port royalties quality assurance testing all of that stuff factors in before the cost of the expensive Parts like the GPU chip itself the memory the voltage Regulators so the lowest end gpus actually start to approach this fixed cost floor to the point where the manufacturing cost isn't significantly different compared to the step up and this is especially true when like the GTX 1630 and 1650 the two cards actually use the same physical chips we'll have these better value cards linked down below as for why they exist then the chips in these bottom of the lineup gpus are most often created first as mobile chips in a laptop the cost difference often really is a matter of just choosing which GPU to solder on because almost all of the supporting components are already part of the final product there's already power delivery circuitry there's already a cooling system there's no need for a PCI bracket and the outputs are already there a laptop is assembled and Q8 as a whole unit so these gpus can make economic sense in that market what makes economic sense for me is to segue to our sponsor Nord security if you keep up with the tech news you'll know that hackers are always looking for new ways to compromise everything from Tech Giant servers to Grandma's computer thankfully Nord Security's well-rounded protection package is there to help protect your files devices and personal data online like nordpass a password manager that helps you generate unique passwords across your devices and browsers or nordlocker a powerful file encryption and sharing service that's a great alternative to Google Drive cybercrime is everywhere these days so make sure you're taking the right precautions when you surf the web right now you can get four months free on all of nord's products so what are you waiting for head to nordsecurity.com Linus or click the link down in the description below if you're looking for something else to watch go check out the last time we did a video about how cheap gpus suck when wait Luke hosted it that was six years ago time sure flies eh

Channel: Linus Tech Tips

Views: 2,006,017

Keywords: gpu, video card, graphics card, cheap, budget, value, performance, price, cost, display, GT 1030, GTX 1630, GTX 1650, RX 6400, RX 6500 XT, amd, nvidia, naming, misleading, Pascal, Turing, RDNA2, Radeon, GeForce, upgrade

Id: 1BLU-tnFOXk

Length: 14min 42sec (882 seconds)

Published: Tue Dec 13 2022