Gadgetory

fractal venturi fans love radiators yes they do what is up everybody welcome back to the channel and the third installment in my first five video series the series that's insanely popular with people with short attention spans who don't like any task where there are more than five steps involved in today's video I'm going to be talking about the first five things that come to my mind when I'm setting up to do some benchmarking because that's what I've been doing recently with the msi r9 390x gaming 8g I'm working on that in a future video so don't forget to subscribe if you wanna stay tuned for that one I also have some fury X's in the pipeline maybe some more work with the 980ti s but I wanted to start this video off by saying a huge thank you to those of you who gave me feedback and likes and even the dislikes on the last couple first five videos on a motherboard - the one on Windows installation it is because of your feedback that I've decided to bring this series back and even several of you actually asked for the one on GPUs and benchmarking so that's where this one came from now I'm talking about GPUs here when it comes to setting up for benchmarks but this could really apply to CPUs or SSDs or any other component in your system that you might want to benchmark as I was getting my list together for this video I realized that these are more about preparing to benchmark but that's really Gordon setting up and preparing because benchmarking can be very time-consuming so when you're preparing you can help eliminate any pitfalls that might cause you to have to redo work later on and that can save you a lot of time in the long run but on with the show thing number one is to start with a plan and since my focus was mainly on the r9 390x gaming AG here from MSI this kind of fell into place with how I arranged the rest of them Tamiya benchmarking setup now I actually haven't even plugged this video card in and started to benchmark it yet you might notice that I have a 980 on the test bed right now and then I also have a 970 here that's because when I realized as much marking the 390x I wanted to plan to give some comparative benchmarks that were against other video cards in the price range and the United seventy is a little bit cheaper 980 s a little bit more expensive the 390x falls right in the middle so obviously this is a GPU I'm gonna be benchmarking some video games with it but even beyond that there's a bunch of other variables that can be involved what resolution are you gonna focus on for example am I going to be doing 1080 1440 or 4k or some other resolution as well am i benchmarking one game or am i vet benchmarking multiple games setting that stuff up at the beginning can really help you focus your efforts in order to save time in the long run so to get your plan down from stage one you should know what hardware you're gonna be benchmarking what software you're going to use to benchmark that hardware and what your test bed configuration is going to be now I have a set test bed right here if you don't have that you can use your computer at home but keeping that test bed consistent throughout the testing is very important and I'll come back to that in just a moment thing number two is to update everything before you start benchmarking I cannot stress the importance of this for instance last night I went ahead and refreshed my windows 8.1 installation on this test bed I actually went for a from scratch installation and added all the windows updates and everything if you do this frequently it's definitely gonna be worth your while to get Windows installation done all the drivers installed and then make an image using imaging software using the built-in stuff within windows to create something separately you can load back onto that the key element for getting consistent benchmarks over time with your test bit is to have it as clean as possible you don't want other stuff gunking it up so every so often at least every month or so you're probably going to want to refresh that operating system installed to make sure that there's not too much other stuff going on so after Windows is installed and run all your dates to make sure Windows Update is updated because you don't want that kicking off and trying to run it while you're running your benchmarks and then you want to go ahead and get all your drivers updated so chip said of course key among the drivers you'll be installing is going to be your graphics card drivers so make sure you get the latest version from Nvidia or from AMD's website and also keep an eye out for if a driver has launched recently if it's been awhile say a month or two or if there's an upcoming triple-a game launch for example you might want to hold off and see if they're about to drop a new driver nothing's more frustrated than running through a bunch of benchmarks with the existing graphics card driver and then a day or two later then watching a new GPU driver and you being like well now my benchmarks are not irrelevant but they're a lot less relevant in that situation beyond the driver updates you want to also make sure your software is updated properly so that means all of the games whether you have Steam installs or otherwise just run each one first make sure it's patched and up-to-date and that those are all good to go the last thing I look at is my monitoring applications that I use some popular lightweight ones are going to be CPU Z as well as HW monitor or Hardware monitor you can also get GPU Z and I've run all three of those there's other ones out there so those are not the end all and be all I 264 is another one that I like to throw out there that's a little bit more heavy I'd say than some of the basic monitoring programs so just make sure you have those set up you're gonna want to monitor stuff while you're running the benchmarks and you want those to be up-to-date as well thing number three is be consistent how does one maintain consistency while benchmarking well done great way to do this is to start out with a plan like I mentioned in thing number one so doing that you're already kind of off to the right step but there's some other things to keep in mind because you want to eliminate the possible variables as you're switching from one graphics card to another or you're switching between say a standard setting as far as GPU frequency and an overclocked setting you want to make sure that the only variable between those two things is a graphics card that you're changing for example so for that reason you don't want to go ahead and be doing any crazy overclocking or changing your overclock settings in between benchmarks if you are overclocking your testbed make sure that overclock is engaged and make sure you've done some stability testing on that because that's another thing you don't want to do is run some benchmarks and then suddenly get a blue screen or something and you realize you've overclock your cpu too much and then that might affect all your old benchmarks that they might be go over and redo things that's really what it's all about avoiding another way to maintain consistency is to run each benchmark at least three times that's kind of the bare minimum standard and if there's much variance between those three runs then run them even more than that oftentimes people will run five times and then they'll throw out the highest and lowest numbers but three times is generally a good standard to go with you can also do reality checking as you're running your benchmarks so find someone else online who has run a similar benchmark configuration to you see what tests they have run and then see if you're getting numbers that are within the same general ballpark now granted as already mentioned lots of little variables can affect benchmark numbers but generally speaking if I'm running say for instance a 59 60 X like I would be here versus a 59 30 K back there that's not gonna affect it too hugely but it's a good idea to go online and make sure that the numbers you getting aren't too far below what the expected amount would be because that might mean that something's wrong if you do encounter a situation where you don't feel like you're getting the right numbers that's where you might go over and lean on your your benchmarking monitoring apps you can see if the temperatures may be too high reality checking is very important as you go along and don't just run benchmarks and run benchmarks until you come back to them later on and be like oh these numbers really aren't what they should be thing number four is a little bit out there so bear with me here I want you to go vertical or horizontal sounds sounds exciting right but no what actually mean here is that you should focus focus what you're going to be benchmarking so that you don't spend too much time again time is just it's a time suck it will suck the time out of your life and it will give you nothing in return except benchmark results I suppose but what do I mean by going vertical or going horizontal well the idea here is that if you're gonna be benchmarking say a 390 X and you want to benchmark it against a bunch of other GPUs like a 980 and a 970 and a 290x and whatever other graphics card you're testing it against and then you want to also do each of those graphics cards with like seven eight or nine different games which is great and people do that so if you want to save a little bit of your time then going horizontal I'm gonna say means testing a bunch of different graphics cards with one game so you might say here's my GTA 5 testing with a bunch of graphics cards and then people or interested in gta5 tests can go look at that going vertical at least in this explanation means you're testing maybe one GPU or maybe just a few two or three at most but you're testing that against a bunch of multiple games going for lots of games and lots of GPUs it's just a recipe for sucking away two or three weeks of your life which if you've got the time and go for it a lot of people do that and and I like reading those articles but for me it's that's too much time and finally thing number five is to give comparisons I see lots of benchmarks posted online and all they have is one set of benchmarks for the single card that they might be benchmarking for example not that that's like completely useless but without a frame of reference it's hard to say that that's good oh that's bad you could look elsewhere online you can do reality checking for example you can look at other benchmark sites at a benchmark but really you're the only person who can say for absolutely sure that you have neutralized all of the other variables so that you're giving a direct eight to be comparison especially if you've been following any of the advice through this video so for that exam for that reason I will say always give at least one set of comparison benchmark numbers now if you're in a situation where like you're like all I've only got one graphics card I don't have a set of them like you do that's cool and I understand so maybe do something like run your graphics card at the stock feed that it comes out of the box and then overclock it and run the benchmarks with overclock numbers and you have like point A to point B comparison numbers and that gives you at least something else to go with providing a baseline set of results so you can look at those numbers and say ah number B is higher than number a and therefore number B is superior makes the benchmark just a lot more satisfying to look at now the one last thing I'll point out here is that since you're running these mention marks on us since I'm saying give a basis for comparison you should also set up some sort of logging mechanism I like to use Google Docs just open up a spreadsheet there you can make it as complex or as simple as you want but it's a great way to keep things logged and then you can just go ahead and get a laptop or something punched in those numbers as you run all of your benchmarks and that is all those are my first five things for benchmarking a GPU specifically but of course applies to other things as well I know there's other things out there too that might come to your mind when you're thinking what are the first things that I think about when I'm about to do some benchmarking so post those in the comment section down below I'm sure there's gonna be some people out there who are like we didn't mention anything about like setting up to a measure power draw for example I like getting a kilowatt or some other power measuring device to do that you might be a little bit more concerned about temperatures may be setting up instead of on a test bed inside of the case doing something like that for more accurate results another extremely popular thing is just to unplug from the internet entirely while you run your benchmarks which I do sometimes but not always but again let me know in the comment section what you guys do or think about when you're setting up to run benchmarks on a graphics card thank you so much for watching this video don't forget to subscribe for the upcoming videos in the 390x fury x is more benchmarks more builds got all that stuff come in also if you're interested check out my store at store Paul's Harbor net I have my new shirts in stock available in three different colors and people seem to really enjoy them they're very soft he'll come up and touch me but thank you guys for stopping by and checking out to the latest in my first 5 video series I'll have more in this series coming soon and as always thank you for watching