I've had both a 2070s and a 5700xt and they consume roughly the same tdp. Either way seems NV is pushing up wattage reliable sources so I dunno.
-
Ever wanted an RSS feed of all your favorite gaming news sites? Go check out our new Gaming Headlines feed! Read more about it here.
-
We have made minor adjustments to how the search bar works on ResetEra. You can read about the changes here.
Rumored specs for NVIDIA's upcoming "Ampere" RTX 3000 series graphics cards (RTX 3060, RTX 3070, RTX 3080 and RTX 3090)
- Thread starter Deleted member 3812
- Start date
- Rumor
- Status
I've had both a 2070s and a 5700xt and they consume roughly the same tdp. Either way seems NV is pushing up wattage reliable sources so I dunno.
Zen's design made it cheaper to manufacture compared to Intel. In the gpu space it's not the same thing. The 7nm node is dense but not necessarily much cheaper per transistor, so Navi 1 was barely able to undercut nvidia while lacking several features of Turing.
My experience with Arma 3 illustrates this pretty well. That game is highly dependent on single-core CPU clock speeds so in many situations it'll always run like shit, but upgrading to a better GPU allowed me to push some effects like total draw distance and shadow draw distance further than before without the performance getting any worse. Having something like a CPU bottleneck may cap your performance, but it doesn't necessarily mean pushing the GPU harder will immediately drop your fps further.
When I went from a 760 to a 1070 I kept the same 550W PSU -- which was originally from around 2011, because I found the 1070 actually drew less wattage than the 760. Even today, as this mid-tier PSU comes up on 10 years old it's still perfectly fine.
If I can build a new rig with a 3070 though I'll move up to 650W or something.
Oh i plan to get a PS5 however for me i've done all my multiplatform gaming on PC simply because of guaranteed (more or less) backwards and forward compatibility in addition to the price proposition generally being better on PC storefronts imo. My consoles (Switch and PS4 Pro then PS5) have been purely for their exclusives with the occasional multiplat on Switch for the portability also however that's always been as a double dip. Currently i've been going the gaming laptop path (on my second one now that i got mid PS4 gen as my initial one had motherboard problems) and for this upgrade i felt that building a PC outright was a better idea and gives me a greater performance ceiling vs price compared to going a 3rd gaming laptop. In terms of choice i would likely bite the financial bullet for say the 3080 over the 3070 because as i see it, a bit of extra financial strain now could potentially save me a whole lot of pain in the future from that additional performance carrying me just a bit longer down the line.
The first thing i will grab beforehand though will be the Ryzen 3900x and then slowly build up the components i want till reveal day. Well the components that couldn't necessarily be completely thrown into flux by said gpu reveals.
Well 1000+ is less daunting than 2000+ i suppose, I'll still be over the moon if the 3080 is closer to 1k than 2k lol.
I was really hoping for $800-900 but maybe I was being silly.
Yes true, but I was thinking that it hasn't stopped Nvidia releasing a 1080 and then a Titan and then a 1080 Ti in the past. There were plenty of people who went from 1080 to Titan, or a 1080 to 1080 Ti. I don't think they would release new cards within a month, but 6 months I could see that happening. Especially now that they're happy to release a variety of cards in the market with their Ti/Super or 3rd party OC branding.
To your original point about whether a PC could last 5 years - I think there's a fair bit of uncertainty in the market right now in a lot of areas. I need to buy a new PC too, so I guess I'm going to just close my eyes and pick the best parts I can, but imo its a really bad time to buy right now from an objective point of view. I think I'm almost prepared for the fact that I might have to build another PC say 3 years down the track, one with perhaps SSD improvement, better 4k, a CPU with more than 16 cores etc etc. The PC i get at the end of the year (maybe a Ryzen 4900x) and a RTX 3080 and a 1TB SSD with read/writes of 3.5gb-5gb/s will be great, but PCs will have jumps every year while the console just took their big jump.
It is cheaper per transistor or they wouldn't do it (this is a fundamental 'law' of the business) - it wouldn't make business sense. Simply by shrinking the node you get more chips per wafer, which directly reduces cost proportional to the shrink, provided your yields are decent. Even if yields start so-so, they'll improve to where they'll make greater profit. Otherwise you just stick with whatever node you're on.
Navi 1 wasn't able to undercut Nvidia because their GPU design isn't as advanced.
The reduction in cost per transistor has been slowing. 7nm cost per transistor was estimated at roughly 70% of what it was on "10nm", but this also comes with added thermal considerations, and the cost per transistor on a larger chip would have been even worse than that because this is an averaged value that doesn't scale linearly with arbitrarily sized dies. AMD had public slides showing that the cost per mm^2 (which is a correlate of cost per transistor, although they obviously aren't identical values) on 7nm was not very favourable.
The economic downsides of using 7nm for large dies were a major factor in why the Radeon 7 (331mm^2) was unable to meaningfully undercut the RTX 2080 (545mm^2). It's a much smaller die, so usually that would imply big savings, but 7nm was just a much more expensive process, even at that scale. Later, the 251mm^2 Navi was only able to undercut Nvidia by a modest margin, even without dedicated RT and AI acceleration on the wafer.
This is not just technology of AMD versus Nvidia, this is also the reality of using modern processes to produce large dies. It used to be that every shrink was denser, cheaper, faster, and more power efficient. They're still getting denser, but the other three factors are in constant contention with each other.
Just assume late September and you won't be disappointed I think.
Rumors are basically anything 3xxx series will do that faster than your 2080ti.
I think it will end up being:
Announcement mid-late September
Release Late October\early November
In preview build of Legion a 2080ti was locked 4k30 with RT and presumably max settings according to DF. They couldnt elaborate on how much headroom there was.
I think we might get some info on the cards in September and a paper launch later in the month.
Only for the 3070 and 3080 however. 3060 and 3080ti (or equivalent) will be later in the year, maybe in December.
12-pin power connector is real, but like the strange cooler we saw earlier in the year, it is a prototype and maybe they won't put it in the cards.
3070 is within striking distance of the 2080ti, if a little lower in raster-only workloads, but better in raytracing. I hope it ends up being $500, but I think it will end being higher than that.
With them being made in Samsung's 8nm means they're gonna be hot.
I'm listening to rumours Ampere is a power hog (3080 Ti 300W and 400W overclocked) and 'only' offers 40% more performance over 2080 Ti. Yet the rumours of AMD's big Navi are 50% perf over 2080 Ti.
Could it be happening? Has Nvidia dropped the ball going with Samsung 8nm?
Could it be happening? Has Nvidia dropped the ball going with Samsung 8nm?
Probably. If raytracing and dlss weren't a thing this generation would go to AMD 100%. (well there's also the issue with the drivers).
I have a 750w psu, but even I don't want to think about the power consumption of the new cards.
They played too hard with TSMC and they lost. I sure hope they go with them for the cards after these ones. Or Samsung gets a good node.
I think almost the same... But
How many "aa/aaa games have dlss"
2?? 3???
The ray tracing always off for the fps..
But I don't know. If amd does not launch something good I will go with a 3070/3080
As far as I know the rumours of big Navi are of it being on the level of 2080 Ti, not 50% more. I still think AMD needs a bit more time and there's also the issue of the drivers.
And at least for people who render through GPU, AMD is still not an option.
My 860W Platinum PSU could use a workout, but I doubt these rumors are true.
12 Pin connector is targeted towards OEMs/System Integrators and reference cards. Has been in the works for years. No, you don't need a new psu.
What options did they have? Navi 10 consume 250W being a 300mm^2 chip on N7. It just wouldn't be able to compete with Turing at all on a less advanced process, going with N7 was the only viable option for RDNA1. And it certainly wasn't cheaper per transistor than the old 16nm node Turing uses.
Too bad that wafer prices are completely different for different processes so you can't directly compare the price per mm for chips made on different processes.
There's being pessimistic and then there's this....lol! Seriously though, I'm about 95% sure top Big Navi card will be faster than a 2080 Ti, it would be pretty disastrous if not. The next gen consoles performance level tells us what ballpark.
Rumour is Nvidia TSMC 7nm 'refresh' of Samsung 8nm Ampere cards next year.
I'm just sad because I can't render with AMD :( And to be fair at this point I've read so many rumours that I don't know what to believe anymore. @.@
Still, I hope they do extremely well. Anything that makes NVIDIA falter is good.
Oh god, the wait continues
No way they do a refresh on an entirely new node, thats nonesense.
I mean you're worried about your PSU, but it's probably the cheapest discrete part of your build (maybe with the exception of your case).
5700 xt already is on the level of a 2070 super.
So you believe they are launching a new card in the end of the year that only matches 2080 ti? Do u think this make sense?
Xbox series x sits between a 2080 and 2080 super and you think their big chip is 15% faster than that only? Really?
A new PSU would still set me back like 100 euros. I'm happy to not spend that.
Out of interest I looked up what the largest capacity consumer PSU is that you can get. Seems like 2000W is the limit. But you can also get this 9000W badboy. Guess I'll go for that one, just be safe.
I have a RM550x PSU and I'm sticking to it. Nothing can go wrong, right?
I don't understand people saying this at all, unless your CPU is using like 250 watts (spoiler: it doesn't) then a 650 watt PSU is more than enough for at least a 3080.
I'll be running it on a 550w PSU without issue. PSU requirements are always massively over exaggerated.
it's a PITA to swap out though - maybe slightly less if you have a modular setup and buy the same brand?
Yup, I also don't really want to deal with taking it out and connecting everything back up.
I had to switch a cooler the other day and I'm good with banging my head against this PC for the next year or two.
Anyway, seems like 620W will be fine and the same connectors.
I had to switch a cooler the other day and I'm good with banging my head against this PC for the next year or two.
Anyway, seems like 620W will be fine and the same connectors.
- Status