Gadgetory


All Cool Mind-blowing Gadgets You Love in One Place

Replying to Comments: The Most Complete Used GPU Comparison Yet

2019-01-12
welcome back to harbor unboxed roughly two weeks ago now i put together the first ever replying to comments episode that was really well-received you guys seem to enjoy it quite a bit and there were loads a requests for more episodes so i plan to create one every two to three weeks depending on what content we've got out and if there's anything worth following up on recently i released a video I'd been working on for a little over a week it was my used GPU comparison covering about 60 different GPU models now that video was extremely well received and I can't thank everyone enough who liked the video and took the time to leave a nice comment I very much enjoy reading all your comments and your feedback and all that sort of stuff I wish I could have responded to all of it but at least I was able to read every last one so again thank you very much appreciated there were however a few comments that stuck out and I felt warranted more of an in-depth response so I've decided to make another reply in the comments episode so that I could do just that so let's get into it right so as I said a lot of you really liked the used GPU comparison and I noticed a lot of requests for the same sort of video but with a focus on CPUs I really hate to do you guys content a lot of the videos we make our subscriber requests and so we certainly appreciate your feedback unfortunately though a CPU version I don't really see how it's possible basically use CPUs are significantly more difficult to make this kind of content for than used GPUs and there's just a number of reasons for this firstly you're not just buying the CPU at least in most cases you're not just buying the CPU I'm more often than not secondhand shoppers are in need of not just the Seaview but also the motherboard and memory so it'd be more of a platform comparison and with that pricing just it can vary massively so can availability and then the quality of the parts that'll also vary quite a bit as well then you have to take other things into account such as overclocking headroom power consumption platform features security issues CPU features and so on in the case that the GPU just slots into a PCI Express slot on any old system you might require one to two external power connect but that's about it so for CPUs I do prefer to review them individually test out their platforms and all that kind of stuff and I do quite a bit of that each year right now though I recommend most of you second-hand shoppers invest in well the am for platform there's loads of cheap first generalising CPUs on sale and if you buy a be 350 board you'll be able to take advantage of these n2 processes that we were released later in the year the only recommendations I really make is I personally would avoid the AMD FX range like the plague as well as any really old Intel CPUs even those cheap x58 Xeon models just not worth the drawbacks unless you can get them for basically next to nothing anything has well or newer from Intel is worth considering at the right price but again I would recommend focusing your attention on cheep am for deals ok next up we have a comment here from Andre and he pointed out this is well it was a bit of an opportunity miss let's say and that was to sort the graphs by the cost per frame honestly I had considered doing this in fact I did actually do it I sorted the graph by the cost per frame and had a look at it but for me it didn't really make sense it wasn't really the best way to analyze the data even so this video does give me a chance to include that graph now the reason I didn't include it in the original video is because while a GPU might offer a low cost per frame which is certainly good the actual performance could well suck making it pointless take the r7 265 for example it offers the best cost per frame but you'd be world's better off spending $22 more on average on the 285 as it offers over 50% more performance and a significantly better gaming experience we see that the GTX 580 also ranks really well but you wouldn't touch that thing with a 50 foot pole the GTX 570 is in the same boat here and there are a few others as well anyway it's not a bad way to arrange the data I just felt it would be less confusing to the viewer only to show the massive graph once and arranged by performance makes it much easier for you to locate a cost-effective GPU that will deliver the kind of performance you deem acceptable anyway certainly not a bad suggestion and I appreciate the comment and I hope I was happy that I was now able to provide that graph moving on for the next series of comments I'm not going to call anyone out as that's not really necessary but it is an interesting it's an interesting point of view so I thought I would discuss it at least give my two cents on it so a few people were criticizing my choice of using the medium quality settings at 1080p rather than ultra one of the arguments I saw was that reviewers test it's an EP primarily using ultra quality settings so using medium patent unrealistic picture just on that note if I had a dollar for every time one of these arguments ended with words such as misleading unrealistic or pointless I have many more dollars than I do anyway at its core it is a fair criticism and believe me as someone who often tests PC hardware under various conditions with a wide range of settings having to land on just one yeah that's something I organized over for quite some time before I got benchmarking in the end I considered the massive list of graphics cards quite a few of them packed less than four gigabytes of vram and many were around six years old after all we're talking about used graphics cards here not the latest and greatest tech hell I don't even review brand new entry-level or budget GPUs with ultra quality settings so it doesn't really seem right to do that with half decade old hardware or even older graphics cards moreover I feel medium was highly realistic who's honestly buying a GeForce 600 or 700 series graphics card and running with the ultra quality preset enabled in games such as battlefield or Tomb Raider really ultra would have broken so many of these GPUs completely skewed at the cost per frame data meanwhile the medium preset did nothing to hurt or skew the cost analysis of the higher-end models such as Vegas 64 and the GTX ten-eighty for example having said let me know what you guys think is medium the right choice for this kind of content or next time should I upgrade to high or even the ultra quality settings I just quickly hear a shout out to my man Timmy Joe he makes videos on the internet about PC parts he also picked up an iron I'm 270x for $6 canadian which is about $45 u.s. and that's $6 below the average selling price that we found in december so he must know what he's doing if you haven't checked out Timmy Joey's PC tech channel I recommend you do he's a bit of a wild man and his contents always a lot of fun something I did look over was the fact that eBay often offers 5 10 or even up to 15% discounts to try and entice buyers and they did a lot of that in December because it's sort of the holiday thing leading after Christmas they try to move as much stuff as they can on the platform so yeah it explains why I was scratching my head so much wondering why so many graphics cards say the gtx 1050 i sold for around $700 or even a little over $700 but if you apply a 10% discount then you're getting down towards $600 and then with a 15% discount you end up below $600 so yeah not really a bad deal after all also I should note that cannot throw out my cost per frame numbers can throw them out a little bit but given the discounts were universal across all graphics card sales it's probably not having that much of an impact still something to be aware of next time when I do update this video and thank you everyone who pointed it out yeah that was a bit of an oversight on my behalf I thought I've made another oversight when exactly thought not sure about that one and Patrick pointed out a bit of an odd result the GTX 780 was better than or really on par with the r9 290 and GTX 970 the 780 also knocked off the three gigabyte GTX 1060 so yeah that didn't seem quite right however I did go back and I checked the results all over again and no mistakes were made so it does seem like the medium quality settings really helps you out - the GTX 780 our performance did get a bit crippled when I went back and tested those tiles with ultra was it horrible but it didn't stack up nearly as well against the more modern cards and then some of the other cards like the r9 290 so this made me think I need to start testing these older GPUs at 1080p using two or three different quality presets normally I just use sort of an ultra or very high quality preset so yeah something to think about when I do the revisits of the older graphics cards I also checked out some other games for honor Forza horizon 4 and even fortnight and again found with media quality settings that the 780 does really well so yeah very interesting okay so viewer wicked a ribbon has a word of warning for us and it's something I did consider getting into with the used GPU guide but in the end decided against it I really just wanted to stay on topic the core topic as much as I could and that was just to focus on price versus performance anyway his warning is to be careful that buying graphics cards from minors cryptocurrency miners we're not talking about young children here he purchased two rx 570s that came with modified biases and he found this broke driver support with newer drivers of course it is possible to reflash graphics cards so this isn't an end of world type scenario but you do need to be a little tech savvy as well as be willing to deal with these potential issues in the first place generally I wouldn't really worry that much about buying an X mining graphics card if you can get one for a bug and then yeah I would just go for that a secondhand shopping is a risk period and I know plenty of people who have bought cheap mining graphics cards over the past few years and they've had no issues whatsoever of course if you are in a position where you can test the card before you buy for example you can drive over at the sellers house place some load under it in a game listen out for things like excessive coil whine and keep your eyes peeled for any artifacts then ya suggest doing that that would obviously be an ideal situation for buying a secondhand product generally though my own experiences buying secondhand graphics cards have been very good I scored a nice comment here from ferns with a request to see behind the scenes video showing the process of making sort of the used graphics card guide I have had these sorts of requests before but honestly it it really is extremely boring yeah it's seriously boring it's about on par with watching paint dry if I'm honest our benchmarking itself is extremely boring it's not really a it's not a great spectator sport and then looking at eBay listings over and over again well that's that's torturous that's that's really bad viewing so I think I'll spare you guys the behind-the-scenes video for this kind of testing and analysis last comment this one is from aku but I did see quite a few comments like this so I thought it was probably worth addressing so the claim is that graphics card pricing is very volatile and the figures in the video will be inaccurate by the time it's released it's certainly true that PC hardware pricing is volatile especially for graphics cards that said the demand from mining is over now and we've pretty well settled into the RT X release cycles so yeah it's not really shaking things up there for the use market so prices seem pretty steady now the prices provided were based on the average selling price in December in the case of the gtx 1070 that figure was 248 dollars u.s. I just tallied up the latest dozen 1070 sold on eBay over the past few days and the average selling price came out to 247 dollars u.s. so pricing seems pretty consistent so far and I don't expect the artex 26a to shake things up much and if anything that will drive the gtx 10:17 1070 TI and 1080 pricing down so obviously that's a good thing for buyers you can expect the next big pricing checkup to come once a midi release their next-gen mid-range GPUs and that's expected to come sometime later in the year and I think that will do it for this episode of reply in the comments again I seriously appreciate the overwhelming a positive feedback on that video everyone that stopped and gave it a like and a comment and all that thank you very much it was great to read over all those comments and even hear your thoughts and feedbacks and suggestions ideas all that kinds of stuff it was really cool and yet all that great positive energy will certainly help motivate me to do it all over again in three to six months from now whenever it's required but yeah not looking forward to it but since you guys appreciate enjoy it so much I will get through that one also coming up on the channel in a few days I have a massive a a truly massive benchmark comparison between the RX 570 and GT x 1050 Ti and many of you have pointed out that I've been recommending the RX 570 for months now but I haven't created an updated benchmark all the stuff I have is quite outdated so yeah since that is the case I will be correcting it and that is going to do it for this one if you enjoyed the testing be sure to hit the like button subscribe for more content and if you appreciate that weird of a box then consider supporting us on patreon thank you for watching I am your host Dave see you next time
We are a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for us to earn fees by linking to Amazon.com and affiliated sites.