hi i got a 300 check from the government and nvidia said i need to put it into a video card so i can mine over the years well that’s fantastic a really smart decision young gamer we can put that money into an rtx 3060 video card and we’ll reinvest the earnings into ethereum currency wallets with taking interest and it’s gone uh what it’s gone it’s all gone the video card it didn’t do too well it’s gone nvidia at this point has realized that its products don’t have to be exciting they just have to exist the rtx 3060 is supposedly out today and supposedly purchasable and regardless of how true that statement is we’re going to be benchmarking it and reviewing it we’ll be looking at thermal’s frequency and the validation of frequency claims specifically for the evga 3060 xc black edition this is supposed to be a 329 dollar card so adjusted for inflation scalping and bitcoin that’s somewhere between about 900 usd and actually it’s infinite there’s actually a floating point limit we hit before the price stops rising so somewhere in that range and this card is uh it comes in below the ti obviously and previously the 2060 the 2060 super came out uh but the 2060 in particular is still kind of available depending on what region you’re in and what manufacturers in your region have decided to keep some 2060 skus alive so that means this is really between a ti and a 2060 and that’s what we’re focusing on most for today before that this video is brought to you by asus and the asus tough gaming b550 plus wifi motherboard ready for amd ryzen cpus the tough gaming b550 board comes in atx and micro atx variants with key features including a wi-fi 6 module 2.5 gigabit ethernet a fanless chipset heatsink for quiet operation and a focus on stability and up time learn more about the tough gaming b550 plus wi-fi motherboard at the link in the description below basic overview for the 30-60 then so 330 dollars is around where the price floor should be it’s 35 84 cuda cores if we’re going to use nvidia’s marketing speak as compared to the 360 ti which runs 4864. it’s over a 30 difference or so but it’s not a one to one percentage change in gaming performance depends on what you’re doing memory is the most interesting so the 3069 ti is running 12 gigabytes of gddr6 as compared to eight gigabytes on the 360 ti also gddr6 but there’s a a memory interface change so this is 192 bit whereas the other is 256. that means that memory bandwidth is reduced on the non-ti and to be very clear here amd is 16 gigabytes of memory on the rx 6800 series is basically marketing you don’t need that memory you will not benefit from it in gaming scenarios versus say 12. or in a lot of cases eight depending on the resolution you’re playing the texture quality things like that but 12 will get by pretty well the reason a company might choose 16 versus 12 or whatever the case may be 10 11 is largely it boils down to how many memory controllers are on the die on the silicon and the architecture of the memory subsystem which we’ve all talked about in the past but 16 versus 12 for modern gaming not particularly useful where amd’s 16 gigabyte frame buffer comes really in handy is if you’re doing a lot of 3d art something like blender it can easily use that memory and we’ve seen it firsthand when andrew puts together some of the actually really cool uh 3d art that we’ve been working on lately we’ve seen the memory get used but for gaming not particularly useful so 12 versus 8 falls into the same category where you are going to become bound by other things on the gpu first rops for example and the ability to stretch to 12 gigabytes actual utilization on this outside of 3d art in gaming it’s very unlikely you’re probably going to hit some other limitations first and that’s fine but the reason i bring all of this up is that nvidia is putting 12 on this to combat amd on marketing nvidia might know something that we don’t know yet about amd’s next card whatever is going to be competing with the 3060 and the 3060 ti and they might have decided based on the information they know or they think they know that 12 gigabytes will do a lot better job marketing and selling this particular card then if they drop down to eight or more likely six is probably what this would have been when compared to amd’s price to price competitor in the future so anyway that out of the way the 12 gigabytes is genuinely very interesting because it creates a really screwy hardware stack where instead of just going like that it’s just a line segment straight up uh you get memory capacity on a chart that looks more like this and we haven’t really seen that before for performance we’ll be looking at standard gaming performance rasterization performance and ray tracing performance we are not looking at rbar for this one we’ve looked at our bar before particularly with the 6800 and xt launches and that’s in a separate bench video if you want to see the amd rx series performance with resizable bar but because amd and nvidia aren’t competing in the same price class we’re looking at it in a standard bench and then we’ll come back for rbar later with nvidia uh just because well this has got enough content in it already okay let’s get into the review we’ll start with some frequency and thermal validation go into the performance and talk if it’s worth it or not we’ll start with frequency validation to determine how the rtx 3060 performs at a lower level there’s no fe model so we’re using the xc black as baseline with emphasis that any more advanced cooling features or higher power targets in v bios would flatten or heighten the frequency line as a result of those changes the rtx 3060 fe plotted a jumpy line between about 1900 megahertz and 1990 megahertz the fact that it’s jumpy is normal for gps dating back to at least pascal but it means that we’re bouncing off of various limiters in this instance it’s power but every couple of degrees will also cost about 12 or 15 megahertz per tick the rtx 3070 ran overall lower frequency in the range of 1850 to 1930 megahertz this is typical of a larger class gpu it’s obviously the better card despite the lower frequency but that lower frequency is what allows it to run a bigger core or is at least part of it as for the rtx 3060 non-ti this model produces a perfectly flat line that’s not because the 3060 gpu silicon itself is somehow magical it’s because of the 3060 xc black specifically this model of card we don’t have much of the usual initial frequency decay that’s experienced from a gpu warming up and that’s thanks to the cooler the 3060xe black holds 1905 megahertz for the entire burn and generally speaking we do actually prefer to see this behavior it’s easier to overclock this way too and it’s easier to predict but it’s often limited to cards with a high enough power target that something else becomes the limiter first like for example voltage thermal’s at steady state position the rtx 3060 xc black at about 64 degrees celsius and that’s tested in a 21c constant ambient temperature but this isn’t delta t value 64 is good and typical of an evga temperature target but remember that an auto controlled fan or auto controlled temperature doesn’t really tell you too much it just tells you what the v bios temperature target is and in this case it’s 64. that’s not necessarily an indicator that the cooler itself is good you have to look at fan speed for that hotspot temperature ran at about 75 c also fine fan speed sat at about 1900 rpm which is average among video cards we’ve tested lately particularly these evga xc models overall the xc black looks okay here we’ll defer to our tear down video separately for more qualitative analysis of the cooler assembly and that’ll go up after this one total war 3 kingdoms at 1080p had the rtx 3060 at 71 fps average positioning it just above the gtx 1080 and rtx 2060 ko stock carts the rtx 2060 super held about the same performance as the 3060 but the 2060 super wasn’t particularly exciting to begin with when it launched so not great positioning generationally the rtx 3060 has improved over the rtx 2060 and the ko for example which is the 2080 cut down that was priced at about 300 so that performance uplift is 11 here that’s not great for a 10 increase in price with the 30 60. we typically like to see more than a one to one generational improvement with price increase otherwise there’s not actually any progress the 3060 ti offers a 30 uplift over the 30 60 non-ti at a 20 price hike for instance the 5700 xt also outdoes the 3060 by about 8.6 percent here although it lacks other features like ray tracing hardware that are present on amd and nvidia cards from this generation overclocking wasn’t bad on the 360xc black the improvement was about 7 with low is also improving at a disproportionately positive pace so that’s a nice change but seven percent isn’t too different from what we’ve seen on other cards for this past generation at 1440p the rtx 3060 drops to 44 fps average and maintains the same 11 spacing from the 2060 ko as before and the 3060 ti maintains a similar 32 lead over the 3060 as at 1080p the 3060 would make sense for most games at 1440p but our graphics settings are high enough here that it’d be ideal to run a card class to 1080ti and up the 2060 super is again about the same as the 36 the albeit slightly ahead amd’s closest option is the 5700 xt but until amd has something of a 300 class the 6800 remains significantly distant in price and performance alike 4k plunges the rtx 3060 predictably down the ranks but overall spacing is similar to previously it’s ahead of the 2060 identically at 11 as before and between the gtx 1080 and rx 5700 xd this isn’t really a card for 4k gaming in most instances but it’ll work in some like maybe csgo or rainbow six each the reason we included it here is to see if the percent difference scaling changes at all but that’s not so for other recent rtx cards as the arc is actually remains the same between the current generation and similar enough to the past generation that spacing is similar amd stumbles a little bit at 4k when compared to 1080p with its newer offerings but the 6800 and xt just like the 3070 3080 aren’t particularly relevant for this price class they do offer some perspective if your budget increases or if one falls off the back of a gpu scalping truck we run strange brigade with both vulcan and dx12 but the hierarchical scaling is mostly the same between the two of them in this 1080p test the rtx 3060 ended up at 165 fps average with lows at space proportionately and without any major frame time spikes the 3060 ti keeps a 29 lead over the 30 60 here with the 5700 xt about the same as the overclocked xc black 3060 and the overclocked 3060 offered a 7 leap over the 30 60 not overclocked thus far most of the percentages are the same as previously there’s one meaningful change on this chart and it’s that the 3060 is now 17 ahead of the 2060 versus 11 in previous tests that’s at least more meaningful than in total war the 2060 super is again about the same as the rtx 3060 here at 1440p the 3060 slightly increases its lead over the 2060 and scales to a 20 lead the 3060 ti maintains the same 29 spacing over the 3069 ti and the 3060 oc keeps about the same seven to seven and a half percent improvement older gpus particularly the 10 series and back should slip further down the relative rankings as the resolution increases and the same is true of amd gpus the key difference is that modern amd gpus will scale disproportionately well at 1080p and 1440p to some extent which may mean that amd will be particularly competitive at these lower price classes where 1080p remains the most relevant solution of course until a card in that price class exists from amd there’s nothing to talk about at 4k the rtx 3060 holds on at 70fps average a testament more to the asura engine than anything else the card falls just below the rx 5700 xt which is flanked by the 360 oc results at 74fps average the same 7 jump as before over the stock result compared to the rtx 2060 ko the 3060 is about 24 the heck and compared to the 2060 super it’s again roughly the same so it’s at least better than we saw in total war versus the 2060 ko but if you end up with a super card previously there’s really nothing to upgrade to rainbow six siege is a popular competitive title that also happens to spew frames at 1080p the rtx 3060 holds 288 fps average marking it ahead of the 2060 super by about five and a half percent and above the 2060 ko by a predictable now 20 the 3060 ti has a 23 lead in this one with the rest mostly matching the previous scaling this game could be played on a gtx 960 even today as we showed in our recent revisit but the competitive positioning of the game itself increases the importance of higher frame rate performance for some players not everybody if you’re not good enough to care then well cool but if you are looking for higher performance than rainbow six siege is the type of game you would use to figure out what gpus are capable of those higher frame rates at lower resolutions at 1440p the 36d ti falls to 237fps average which is about the same frame rate as the 2060 ko in the previous 1080p chart the 3060 non-ti ends up at 186 fps average just behind the 1080 ti again this card the 3060 shouldn’t be an upgrade ambition from 20 series cards in general although some of them are ahead of it but especially not from a 2060 of any kind super or not it’s just not a big enough jump for the price but that’s normal it’s more sensible if skipping a generation and moving from say a 1060 or a 960 although since the 3060 is 100 more expensive than the 960 at launch which is more than inflation at this point it does feel like it’s a different price category than it used to be the rtx 2060 super ended up just below the 30 60 xc black once again with about a 10 fps deficit 4k results show the 3060 following the expected ampere behaviors higher resolutions tend to benefit the newer ampere 30 series cards as compared to most older nvidia architectures or amd’s rdna too and so the 3060 secures a spot alongside the gtx 1080 ti some things never change like the fact that the 1080 ti remains the goat the 2060 super trails the 3060 as does the 5700 xt and the 2060 ko allows the 360 a lead of about 21 here the ti the 30 60 ti that is and up would be needed if you want to hit those higher refresh rates while still playing at 4k but anything above 60 here is fine it’s just not maybe as competitive red dead redemption 2 is up now this one is among the more demanding titles for gpus as a counter to rainbow six and it’s useful as a measure of gpu constraints even at lower resolutions the rtx 3060 pulled a 92 fps average result with lows paced predictably behind we didn’t see any issues with the frame time pacing while running this card at least not any outside of the usual comparative deficit in total war 3 kingdoms the rtx 2060 ko held 77fps average previously with a gtx 1080 at 73 fps average generally speaking users of gtx 1080 and better probably don’t have much of a reason to upgrade to a 3060. you’d either need to climb the stack higher or wait for another generation for scale the 3060 is 19 ahead of the 2060 ko the 360 ti is 23 ahead of the rtx 3060 and just for an amd comparison the rx 6800 is 49 ahead at 1440p the rtx 3060 does 71 fps average about equal to the gtx 1080 ti we’d reprise our old title from the 2080 super review the one where we said we get it nvidia you can make a 1080 ti but the 3060 is actually significantly cheaper than a 1080 ti it’s an improvement despite equal performance the rest of the chart falls as predicted the 2060 run 61 fps average the 360ti runs 93 and the percent scaling is mostly the same as the aggregate of the previous results the 4k result for red dead redemption 2 has the 30 60 below 60fps now at 44. you could drop some quality settings to help it a lot but at this point 4k would best be paired with a different gpu a higher-end one if you’re not tied to the idea of 4k in every game played the 3060 might be an okay solution but there are times when it’s just not gonna cut it like here for cyberpunk 2077 we ran our standard test at high settings and 1080p the rtx 3060 ran at 80 fps average placing it once again right between the 5700 xt and the 1080 ti the 3060 ti has a 29 lead as a result of this that’s similar to some previous tests but a little on the higher side this configuration is playable on the 360 without any serious issues well aside from the existing issues in cyberpunk you can’t fix those horizon zero dawn’s up now and we’ll just do 1080p for this one to keep it short at this point we’ve established a pattern and percentage scaling expectation so we sort of know where the 3060 will fall on basically every game at 1080p the rtx 3060 produced an average fps of 113 tied with the 1080ti fdw3 and not distant from the rx 5700xt the 360 oc was somewhat meaningful here getting us another 7 and moving to about 121 fps average that’s not bad for an rtx 30 series card at the launch price and by at the launch price we obviously mean if you converted the msrp and usd to bitcoin at a one-to-one ratio we’ll stop there for now on the rasterized game performance let’s move on to rtx benchmarks and then we’ll talk power consumption quake 2 only has nvidia devices on our charts right now the team recently added wider vulcan support but since we ran all these tests with the rtx extensions previously we’ll stick to that for the remainder of this gpu bench’s life and then upgrade later this is fully path traced so rasterizing performance isn’t useful here at 1080p the rtx 3060 ran at 66 fps average about the same as an rtx 2070 super or an rdx 2080 fe the 3060 ti leads by 19 a little below the average leadership percentage in rasterized testing but overall roughly the same hierarchical position as before at 1440p the rtx 3060 held 38 fps average putting it in the same relative position as the 1080p results the 3060 ti here scales to 20 to 21 percent ahead and at 4k the 360 falls to 17 fps average unplayable really but maintains the same hierarchical positioning the only interesting takeaway here is that the 360 ti is now 22 percent ahead so it’s slightly gaining with each resolution uptick when relying heavily on rtx functions the rt testing for shadow of the tomb raider tends to balance better with cards that have stronger rasterizing advantages than they do ray tracing advantages like the 6800 and the xt cards because in this game at least with this test only rt shadows are present so it’s still mostly rasterized at 1080p shadow of the tomb raider with rt has the 30 60 at equivalent levels to the 2070 and that allows the 3060 ti a lead of about 27 the 3060 out does the 2060 ko by 26 when rt is involved at 1440p the rtx 3060’s 51 fps average has it again between the 2070 and 2070 super with similar distance from the 2060 ko and 3060 ti as previously control is next we have dlss on and off testing for some of these tests and amd’s on here as well but ultimately control is among the best dlss implementations so far and will almost certainly be played with dlss on with an rtx card the dlss on off testing is only useful for determining how much dlss impacts performance like for like but we did that previously so we don’t need to do it again the 1080p results have the 3060 running at 84fps average with dlss on the 3060 ti has a 31 lead over the non-ti but if we disable the lss the ti drops to 68 fps average and is similar to a 2060 ko with dlss on showing you the value of dlss when it works well finally a quick look at minecraft the rtx 3060 stock card ran 83 fps average when dlss was on which is basically required for minecraft rtx and it positioned itself again between the 2070 and 2070 super at this point we’ve established that this card will land between the two 2070 entries for rtx performance power consumption is last this is tested at the board level and isn’t power from the wall we’re testing only power consumed by the gpu and the board including the vrm efficiency losses with the power virus workload the rtx 3060 xc black pulled 169 watts when stock or about 185 watts when overclocked this aligns with the 12 offset offered on evga’s v bios but obviously would change depending on how much power a particular v bios allows the card to draw if it has multiple power connectors it’s possible that the manufacturer lets you pull more than what we’re seeing here the 3060 ends up below the 360 ti stock draw by about 30 watts here in a gaming workload the rtx 3060 pulls about the same power as in the power virus load when 100 loaded meaning we’re not waiting for the cpu here this is normal and it’s expected behavior but it helps to validate power numbers in a real world gaming environment as well the rtx 3060 ti pulled 201 watts here so spacing is about the same as before the oc again boosted the 3060 by 10 to 15 watts depending on when it was tested and this aligns with the power offset finally as we close out here just a couple quick overclocking notes on this it was pretty easy to overclock it’s the same process as the other 30 series cards we’ve already talked about all that in the past and have hours and hours of content on it one key note though interestingly this particular card and not clear how much this is the xc black and it’s the bios versus all 30 60s but this particular card it stopped scaling at a 100 to 175 megahertz offset so anything from 100 to 175 for core offset did not produce a frequency change which is not completely abnormal but it’s a little bit odd because normally you start crashing somewhere in there but instead it looks flat and so the first reaction we had was maybe this is some sort of driver level let’s not allow it to boost higher than is stable so that it doesn’t crash thing but that wasn’t the case what we ended up doing was typing in 250 megahertz offset 200 megahertz offset and so forth and we found that eventually it got past this weird wall where it was not boosting at least in the 3d mark applications we were validating with and finally at 200 and 250 it boosted higher 250 was not stable 200 was as the offset but kind of interesting and the only real takeaway here is if you want to play around with one of these for overclocking if you see it stop scaling at whatever offset doesn’t matter they’re all different because the baseline’s different but let’s just say it stops scanning at 100 but it doesn’t crash at 125 or 150 it might be worth keep increasing that until it eventually does crash because you might hit some weird trigger point where it increases as for a quick conclusion so in ray tracing performance we found that it’s somewhere around a 2070 to a 2070 super in equivalent performance and obviously if you own basically any 20 series card this isn’t really an upgrade at best it’s a lateral move to side grade if you had say a 2060. uh versus the 2060 super and rasterization performance on the 2060 in rasterization performance it’s anywhere from 11 to about 20 percent versus the 2060 ko so that would be the cut down formerly 2080 die that was turned into a 2060. and it’s uh anywhere from the the 20th to 30 percent lead in the 30 60 ti versus this 30 60. so that gives you the scale where it falls and ultimately for value it’s we’re not like we’re not particularly positive or negative it’s kind of meh and the reason we feel that way is it’s pretty simple so we can’t ignore the value argument it is compelling to ignore the value argument right now because as a reviewer you know that the pricing isn’t necessarily going to be what it’s set to but the reason we in particular are not ignoring the pricing of the card despite the fact that the price will not hold that launch is because the pricing of the card is what nvidia sets and it indicates nvidia’s behavior and nvidia’s expectations and what nvidia is trying to do with the cart and regardless of the wider market that’s something we’re very interested in what is nvidia trying to do with its positioning of this device and our takeaway is that the the price to performance increase is simply not exciting enough to in any other generation make a strong recommendation for this card it’s fine it does well overall but it’s not exciting and in this market that’s all you need to sell all you need is it’s fine in some of these games we’re seeing a one percent dollar increase for a one percent performance increase which is not progress it’s stagnation and so that’s why we’re met on it in a couple other games it does a little better than that but not enough to really make us do a headline that is extremely positive about the device so overall we don’t really have any major complaints at least not about this particular xc black model but it’s also somewhat disillusioning to review devices when we already know that they’re in such short availability versus the demand regardless ultimately this part of the conclusion is for the future as this card comes into stock what you need to know is 2070 2070 super for rtx performance is where this falls and then it’s roughly a 1080 ti to some of the 2080 class cards in rasterization performance with a 3060 ti offering about a mid to mid 20s to low 30s percentage point increase over the 30 60. so if you’re comparing should i buy this or the ti that’s the range to look at it’s about 30 higher bump in rasterization pretty reasonable and for the most part probably worth it if they’re all at msrp amd doesn’t have an offering to compare directly to this right now the closest thing is the 6800 which is nearly 2x in price uh so we’re not going to make that comparison but you can obviously pause on the charts if you’d like to make it yourself that’s it for this one thanks for watching subscribe for more as always you can go to store.gamersexes.net or patreon.com directly and get behind the scenes videos and we will see you all next time

Leave a Reply

Your email address will not be published. Required fields are marked *