Gadgetory

what's up guys how's it going and welcome to my first ever radeon RX vega review AMD's new GPU line it launches today with the Vegas 64 and the soon to be available they get 56 I'm going to be focusing on the Vega 56 today which is retailing for $399 and is intended to compete with nvidia gtx 1070 under the vega 64 as well which is going to retail for $499 for the standard version at least it looks like this what i have here is actually the brushed metal finished limited edition version that you can get if you purchase a Radeon pack but I will put links to the Radeon packs as well as the individual cards in the description once they become available and another day or two I'm going to be coming back with a video follow-up on the Vegas 64 benchmarking numbers for today that I'm going to be focusing on the Vegas 56 versus the NVIDIA GeForce GTX 770 so let's get right into the testing set up my test bench is composed of an Intel Core i7 7700 cave cpu running at 4.8 gigahertz the motherboard is a gigabyte z 270 horas gaming 5 the water cooler is a corsair h1 sent to under an 18 millimeter all-in-one liquid cooler and for memory I have a 2 by 8 gig kit 16 gig solo of Kingston HyperX predator ddr4 running at 3200 speed cast latency 16 in dual channel mode for storage I'm a Windows 10 operating system on an Intel 600 P 512 gigabyte nvme SSD games of video files are on an external san difficult or to one terabyte data SSD connected to a USB 3.0 not only for power I've of course are HX 1000 I 1000 watt 80 plus platinum power supply and it's all on an open testbed so for comparison I have the galaxy gtx 1070 EXO c sniper and this is a manufacturer overclocked third-party reference design version of the gtx 1070 this is the ultimate benchmarking question when you have a new graphics card like the vega 56 what should i compare it to Andy of course would prefer that I compare it to the gtx 1070 founders edition but that's running at stock frequencies and stock GP runs at about 18 50 megahertz peak 17 50 megahertz stable under gaming load but just seems a little bit slow given how well the 1070 actually overclocks and how many third-party designs are on the market that you can buy that already come pre overclocked so I chose the Galax gtx 1070 xoc sniper for this reason it has a healthy manufacturer overclock it takes just shy of 2,000 and was running at about 1962 megahertz under load given that the 1070 is out for over a year already I think this modest concession to the third-party versions that are available is fair but feel free to use the comment section down below to tell me if you disagree as always the GPU driver I'm using is the latest from video version three eight four nine four and I have pressed provided drivers for the Vega cards I'm using hardware info 64 and the AMD Radeon settings utility to monitor statistics ambi also delivered a new driver and software build on Saturday with some updates for overclocking supports and although we weren't supposed to use that build for comparison testing actually got some pretty solid results with it so I have included them too now with early hardware and software monitoring can be touchy so here's how I overclocked with Radeon setting using the wot man utility under global settings manually entering the frequency just wasn't working it also used the percentage slider I started with a 10 percent overclock that I eventually dialed back to 9 percent to due to some instability I maxed out the voltage where possible I bumped up the eight gigs of HP into memory from 800 to 900 megahertz frequency I set a much more aggressive fan curve maxing out at 4500 rpm and I set the power limits plus 50 percent I'm not completely sure what frequency I was hitting again due to quirks with early software and numbers showing up that pretty positive weren't actually accurate but I think the GPU is operating in the 1662 1680 megahertz range topping out at just over 1700 I did encounter some instability as well while I was going through these benchmark numbers but I've still decided to include them because they did improve my performance by about five to ten percent when it was working and I wanted to give at least a starting out idea for how performance might improve with Vegas 56 when third-party cards with better cooling options arrived with all that said here are my benchmarks so guys I know I didn't talk you through the benchmarks this time but I think I will do that for the next video when I cover Vegas 64 honestly though these were some of the closest numbers I've seen with two similarly priced graphics cards generally speaking the 1078 kind of shoots the gap between a stock and an overclocked Vega 56 which would make them very competitive at the same price there's a bit more to the story though starting out with Tower draw now I measure power draw at the wall I got an average while gaming with 3d mark and I also did the peak power draw across all the testing that I did Vega at 56 overclocked got pretty power-hungry hitting as much as 487 Watts max power draw that is while overclocked though now Vega is a higher TDP graphics card though with two eight pin PCI Express power connectors and varying board power settings depending on whether or not it's water-cooled what power profile you're using and if you switch to the secondary v bios which is tuned specifically for lower power usage I ran my Vega 56 with the main v bios and the turbo profile because I'm more interested in performance and efficiency right now that means I'm using 190 watt power profile but all of these values kind of go out the window when you start overclocking as you can see from my average in peak power draw numbers the ramp up is pretty significant when you overclock I haven't tested these lower power settings but my impression was that Vega can be very power efficient as a GPU just not while running it to frequencies that also produce good benchmark results when you compare it to Nvidia as for temperatures both of these cards got pretty warm hitting 80 degrees or more on the GPU while benchmarking the temperature increase while overclocking the Vega 56 might seem minimal here but remember that I was also running the fans at close to max meaning that they were also pretty loud at the same time speaking of loud though unfortunately I didn't have time to run official sound meter testing for this video but anecdotally I can tell you that with the base turbo setting the fan noise for the Vegas 56 was just fine not quiet but well within acceptable noise levels for gaming when overclocking however well a blower of style fan running at 4500 rpm might as well be a small vacuum let's just put that way that's for frequencies again there was some weirdness with the monitoring software but my card was able to hit the same frequency from what I could tell as the Vega 64 liquid-cooled when overclocking which is supposed to be about 16 77 megahertz but my card was at least up in that range when not overclocked it was pretty stable as far as I could tell at 1590 which is the reference speed so overall now that Vega is finally here is it's living up to the hype I would say not really but that's mostly because there was too much hype and it was drawn out over way too long a period of time it's also totally fair to point out that AMD is competing with Nvidia cards from last year the GTX 1080 and 1070 launched at May 27th and June 10th 2016 respectively and these cards watching 15 months later are competitive when it comes to performance and pricing but not quite so much when you also consider power draw and launch timing I also noted that the 1% and 0.1% lows weren't looking quite as good for Vega in my testing it did improve when overclocked but hopefully those numbers will also smooth out as we see some driver updates as the drivers we can also assume are pretty early at this point on the plus side though it is very good at least to show that the Vega 56 is very competitive with the gtx 1070 at all resolutions so if power draw is less of a concern for you and it's a very viable option even more so when you look at the complete picture which includes not just your GPU and your computer but also monitors as freesync monitors are just way more affordable than in videos g-sync alternatives often buy one hundred four hundred dollars or more if you're buying a GPU and a variable refresh rate monitor the precinct option what they get is just a much better price to performance choice if we're really looking at the whole picture though we have to also include cryptocurrency mining and that could very well destroy any hope that the common gamer has for getting one of these GPUs at the actual retail price there's still a lot more to be said and question when it comes to this Vega launch I'm sure in the next few days and weeks but for now I'm going to cut this one off and continue my work in the Vega 64 benchmarks if you guys have any comments or suggestions for that coverage please leave those in the comment section down below subscribe to my channel and don't forget to turn on those notifications too if you don't want to miss when I post in the next video it's a thumbs up button on your way out if you enjoyed this one thank you guys as always for watching and we'll see you next