Gadgetory

today we're reviewing Nvidia's new gtx 1660 and technically the one we tested is an EVGA 1660 but and I'm sorry Jacob but we're in MSI headquarters so Emma sighs the backdrop for this they also have 1660s we're gonna be looking at the 1660 performance today stock and overclocked versus primarily the 1660 ti versus the rh 590 Vega 56 and basically every other card at nearby in the stack so today's focus is on Nvidia's new $220 price floor video card the gtx 1660 before that this video is brought to you by the coarser 1i 140 compact gaming PC the Corsair 1i 140 is a small form-factor PC outfitted with a 9700 K RT x 28 e 32 gigabytes of RAM and a 480 gigabyte nvme SSD all housed within a 2 millimeter thick aluminum chassis the Corsair 1i 140 is a 12 liter system fit for desktop use with the same size I 160 counterpart with higher-end parts learn more at the link in the description below MSRP for the 1660 is tactically about $220 that's the price floor and video sets but the partner model we're looking at is going to be about 250 for the EVGA 1660 dual fan that we reviewed these were checking the price on will put them in a title card on the screen if we get them in time so this is the gaming x16 60 from MSI there's also an armorer 1660 which should be the cheaper version probably but armor OC and one thing that we learned in the process of tests in the 1660 was that if you get a model like we tested with a higher power target the overclocking Headroom will actually allow you to bring the 1660 up in performance to about 1660 TI stock levels which is very interesting because it's it's a study and buying one price to your cheaper of a component and then you just put maybe 20 minutes of manual effort into it and it becomes the next price class up so if you start at a 240 250 dollar price point for or 1616 on Ti and then overclock it you get to about the same performance levels as a 1660 ti stock card at the $280 price floor which will probably have a like a 130 watt maximum board power so and and the 1660 that we tested also gets up to about 130 watt board power and and will look into these MSI ones later perhaps so let's get straight into the charts Patrick ran all the tests for us we're remote we're still in China but we're almost done here and he ran the tests we have thermals power and gaming to look at today and then we'll talk conclusions and if it's worth the money especially at the $250 boosted price point of EVGA is dual fan card that we're looking at today apex legends is still new to our charts and not as heavily populated with cards as our later games this one is important for its relevance but also because it's built atop a modified Source engine with DirectX 11 as the API giving us a title that represents performance on much of the games on the market we retested the 1660 Ti in this game but we haven't yet retested all of the other cards recent updates have improved performance on the 1660 TI n AF x legend so this is Nvidia driver side we also need to revisit some others later once we're home from China will use our multiplayer test course that we heavily detailed in our previous content tested first at 1080p this test placed the GTR x16 60 at 90 FPS average with high settings or 101 fps when overclocked for reference the 1660 TI XC stock performance was about 99 FPS average stock or about 102 FPS average went overclocked we did not see much performance uplift or scaling with the 1616 TI's overclocking apex Legends after the update at least we didn't see as much as before when we first tested the game and some of this we'll look at again in 1440p we tested this a total of three separate times including full driver wipe spore validation and confirmed this performance all of those times the 1660 was the closest to the TI and apex legends with most other games showing 13 to 17 percent performance advantage stock to stock between the two the rx vagin 56 card for reference ended up at about 106 FPS average with the RX 590 card at about 89 FPS average tested at 1440p the 1660 stock card still pulled about 61 FPS average although Loza started dragging behind performance is reasonable and the game can be played at about 1440p even with this card it's just that the settings should probably be dropped slightly to better accommodate heavy combat areas and the initial drop in the 1660 ti for what it's worth was that about 68 FPS average posting a lead of around 11 percent over the 1616 on TI overclocked in the 1660 got it 272 FPS average which is where the 1660 T IOC landed the two are within margin of error cyber elite 4 is still our best example if we could DirectX 12 implementation we tested this one with both 1080p and 4k resolution but please note that the 4k testing is meant to be treated as more of a synthetic representation of performance from card to card because we have so many cards on the chart it's easy to make those comparisons this is done because 4k gives us a full range of performance without CPU bottlenecks at the top end of the stack so we can compare the cards all to each other without any performance limitations from the game engine side starting with the more synthetic 4k workload the GTX 1660 posted 39 FPS average with low is very close by at 34 fps 1% and 33 FPS 0 1% these are pretty consistent overall comparatively the GTX 1660 TI XE under stock conditions posted 45 FPS average or a lead of about 16% over the GTX 1660 the 1660s outperformed marginally by the RX 590 which held a lead of a few FPS average and is not too dissimilar from the gtx 1066 here by its 37 FPS average overclocking the 1660 to 169 hertz offset got at 245 PS average matching the performance of the gtx 660ti stock card roughly matching the rx 590 Fatboy overclocked results of 46 FPS average and allowing the 2060 stock card a lead of 24% moving on to 1080p testing the GTX 1660 pushed at 110 FPS average stock allowing the 1660 TI XC a lead of about 14% the RX 590 fat boy ended up at 120 FPS average pushing closer to the 1660 TI than the 1660 in this title sniper is compute-intensive and allows AMD to leverage its architecture more than most other games do overclocking the 1660 got it to 126 FPS average marking it as equivalent to the stock 1660 TI XE and within margin of error it also allowed the card to outperform the RFI ninety overclock but this is insignificant beyond declaring that the two can be made to be equal the rough to FPS difference can't really be called observably better although it is measurable but it can certainly show a minimum of equal performance everyone 2018 is next for this one tested first at 1080p the GTX 1660 90 I ended up at 92 FPS average with lows at about 50 fps and 31 FPS for 1% in 0.1 percent lows respectively the performance allowed the 1660 TI XC elite of about 13.7 percent overclocking the 1660 to a 160 megahertz offset pushed it to 104 FPS average which is the exact same performance we got out of the 1660 TI XC even the lows are roughly identical sure overclocking the 1660 TI gets another 5 FPS average but at some level it may be better to step down to a 1660 and settle for TI baseline performance and keep in mind too that the 1660 TR we tested had a TDP of 130 watts or 10 mobs over stock but it did not have any power offset allowance unlike some of the other cards like the 1616 on TI that we're reviewing today if you wanted 1660 TI stock performance the 1660 could be purchased and overclocked for the same performance level unless money Vega 56 hid 115 FPS average when stock for reference with the rx 598 gigabyte card holding at 82 FPS average one stock and just to clarify there saying for reference means for your reference not necessarily with the reference card the Vega 56 and rx 590 cards ran tighter frame time consistency than Nvidia does in this game with Vega 56 advantage to a point of low frame time performance averaging to about 60 FPS 0.1% lows whereas these 16 60 TI and 1660 both said in the 30s at 1440p the 1616 on TI operated at 70 FPS average giving the 16 60 TI elite about 10 FPS average or about 14% once again overclocking ties things up between the OC 16 60 and the stock 16 60 TI the 16 60 is fully capable of playing at f1 2018 at 1440p and could be made more consistent in frame times with some settings tweaks performance is about at the same level of the gtx 1070 stock card when overclocked and better than a 1060 six gigabyte and also not too distant from an overclocked archetype 90 which is no relax for arch 580 which is an overclock to our X 480 far cry 5 is next at 1080p the GTX 1660 ran an 81 FPS average of putting it as the closest in performance to an overclocked rx 590 and allowing the GTX 16 60 TI stock card a lead of about 13 FPS average over 16% this is consistent with other results Vega v6 maintained a strong lead over both the ti and obviously the non TI for this game but pricing also structures the 56 and 1660 at much different price glasses the 1662 overclocked planted it once again at TI level performance at 1440p the 1660 card stock still managed to nearly hit the magical 60fps mark that everyone wants and is close enough that it could be made to work reasonably well at 1440p with some settings tweaks at 56 FPS average and 64 FPS average when overclocked the card nearly hit 16 60 TI performance even at the higher resolutions the gtx 980ti was also roughly tied with the overclocked 16 64 reference with the stock rx 590 ending up functionally tied with the gtx 1660 shadow of the Tomb Raider at 1080p plays 16 60 stock performance at about 74 FPS average allowing the 16 60 TI a lead with its 85 FPS average or about 15% once again this lead is consistent with other titles and seems to be about the average percentage increase overclocking the 1660 with the 160 mega Hertz offset allowed it to match the 16 60 TI stock card which then climbed another couple FPS when it too was overclock the 1660 OC performed about where the gtx 1070 is with the stock gtx 1660 at about rx 590 performance including that lows that were within margin of error or close enough at 1440p the gtx 1660 fell 250 FPS average roughly tied with the RX 5 90s 51 FPS average and low performance is within a margin of error of the RX 590 the 16 60 T I ended up 14% ahead of the GTX 1660 with in range of our previously potted differences while the 16 60 OC pushed it up to 1660 T I actually stock performance the 1660 TI XE overclocked permitted another 5% performance over stock but the 1660 overclocked permitted 15 percent performance uplift this massive headroom is what allowed for the tie making performance note however that the gtx 1660 were reviewing has power target increase availability or allowance of about 23 percent whereas the 1660 TI XC reviewed at the low end of the price stack remained capped at 130 YTD p about the same as what our 1660 was doing and 130 watts is still 8 percent higher than the stock 120 watt reference TDP of a 1660 TI the 1660 TI XE if it also had a 23 percent performance allowance power allowance could be pushed a little bit further and would increase the gap that you're seeing here it's just that for these two cards the 1616 on TI can be made to match the ti up until a point of overclocking with minor differences emerged finally for the GTA 5 performance at 1080p the gtx 1660 ended up at 85 FPS average allowing the 16 60 TI oxyelite of about 10 FPS average or about 13 percent overclocking as the last several charts katli 1662 about same performance as the 1660 TI XE stock card the RX 590 ended up far behind here with Vega 56 also stuck at lower performance levels then seen in other titles relative to the Nvidia cards the way GTA 5 is built tends to favor and video devices which isn't for any reason aside from how much compute is used in each game and so the 1660 does favorably here 1440p plates to the 1660 stock are dead at 60 FPS average but it's frame time performance became more variable and dropped to about 32 FPS here 1% close some settings reductions would be recommended to improve frame timing consistency although the card is overall capable of running GTA 5 at 1440p with these settings overclocking pushed it to 70 FPS average or about the same as the 1660 TI XC once again and about the same as the vegam 56 red dragon card from power color for that overclocked we've been showing Patrick Stump the card up to its maximum power target and saw performant old consistently at nineteen thirty five megahertz with few dips between this is often rare and it's a result of power limitation and positions restricting clocks as the frequency bounces off of that power limit but with no core offset applied and just the power limit increased the clock becomes consistent and holds steady this remains true at 100 megahertz offset where we're still not hitting a power limitation skipping a few steps we saw crashes at 175 megahertz offset on this card and so settled on 160 megahertz with a 15 mega Hertz increment being about the minimum stamping that can be applied to Nvidia cards at 20 70 megahertz there's not much to be mad at with this car 20 70 is a strong frequency to hold for a mid-range video card memory ended up had about 24 80 mega Hertz we often stick closer to 400 500 megahertz offset for gddr5 we didn't see any performance decay in games with this 960 offset the memory is good enough on this device the one that we tested and holds these higher clocks with that issue you'll sometimes see soft memory errors with this type of offset but in the games we tested those didn't impact performance if they occurred it may be more of a concern with professional applications although we doubt many people are using a 1660 with those types of applications so overall the overclocking on this card was good it was good enough to make it into a 16-6 TTI a stock when running with that overclocked power consumption testing is done at the wall but uses heavily controlled test benches for accurate tests run to run we're looking at a total system power consumption in this test so it's not the individual card power the gtx 1660 peaked at about 258 watts total system power consumption 1 stock averaging closer to 200 watts total system power draw our 1660 tea I saw about the same power consumption for reference but keep in mind that partner model cards can change power target beyond reference total board power provided by Nvidia spec table which is what happened here the RX 590 for further reference plotted at about 325 watts peak power consumption and remained more consistent in its higher power draw whereas the Nvidia cards will power throttle more aggressively during testing the RX 590 ended up about 70 to 120 watts higher than nvidias 16 series power consumption depending on the thermal torture testing though is up next under stock Auto conditions the EVGA 1660 that we tested kept a thermal target of about sixty degrees Celsius illustrated by the thermal overtime plot holding steady at sixty see this had a fan ramping to about eighteen hundred rpm against its maximum fan speed on the cooler of 3500 rpm and we allow this to be automatically controlled so we're letting the BIOS on the card dictate the fan speed which allows us to see how it performs and what's thermal targets are out of the box thermally the GPU does well and the fan rpm could be manually run at a less aggressive curve if lower noise were desired rather than lower thermals so what she over to the frequency chart we see one of the most impressive frequency lines we've ever plotted versus a results of the lower thermal target that kept this card under sixty degrees Celsius for the duration of this particular test the card operated at about 19 20 megahertz constantly when under stock conditions and with our 3d mark fire strike extreme fixed workload this illustrates good thermal performance of the cooler and that were not bouncing off of major power limits under stock conditions 1660 positioning then after all the charts after you've seen all the game performance it ends up we're at the $250 price point the card is it's obviously a replacement for the 1066 gigabyte clearly and it's a decent replacement at that it's a competitor to some extent with the RX 590 which is about the same price plus or minus maybe 40 bucks depending on the model and for the RX 590 to really compete well it needs to come down in price closer to the 1660 overclocked models are the ones that are easily overclocked anyway like the one we had so all right 590 is in a bit of a rough spot right now performance wise the 590 we'd like to see it closer to the to $20 range depending and there might be some out there at this point remember that prices on launch of cards the competitors of those cards will often change the prices so you might see the 590 come down a bit by the time this review goes up and we'll double check and see if that is the case an update News video what we think of so but 590 needs to come down a bit to be truly competitive 56 we'd like to see stay in the 282 maybe $310 price range at that price range 56 is a really good deal but it's often more than that and that's where the 1660 Ti starts to look really good especially the 280 our price floor for that one so 1660 really where it shines we think is if you buy something like the dual fan models maybe like one of these ones in the background here the armor or gaming extra something like it depending on where the price is if it's in the 242 maybe 260 range and you put 20 minutes of work into overclocking it you can get it up to 1660 TI performance when it's stocked so that's really not bad if you're willing to put in a couple minutes of work you can save about 30 bucks off the top of the 1660 TI price and for people who are on a really strict budget the 1660 does well in those instances so that's gonna be it for a look right now the 1660 does fine it's gddr5 so it's a bit different but performance wise once you kind of match the clocks it's not all that different in gaming because a lot of the games are more frequency bound on the core than they are bandwidth bound on the memory which is what we saw today on our gaming test where the 1660 brings up performance to 1660 TI class so that's it for now we'll be home soon and do some more testing on these 16 to 60 cards once we get back thank you for watching as always subscribe for more you can go to store documents access net to support us directly and also check back for our msi factory tour because it's gigantic and they make millions of video cards and month reports per month so check back for that i'll see you all next time