• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

GeForce GTX 1180, 1170 and 1160 coming in August. Prices inside

Page 13 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Status
Not open for further replies.
https://www.tweaktown.com/news/62855/msis-new-geforce-rtx-2080-ti-gaming-trio-teased/index.html

62855_01_msis-new-geforce-rtx-2080-ti-gaming-trio-teased_full.jpg
 
Yeah, unfortunately AMD won't have anything close to these GPUs until 2020+ and by then NVIDIA will have 7nm behemoths dialed.

Yup, it's a huge die and they're pushing clock speeds through the roof but I bet they're super efficient if kept at lower clocks. I wonder how they will mine...
For myself, I'm waiting for 7nm, both CPU and GPU to do a full build. I want to go 4K HDR and all stars should be aligned by then, especially with the predicted fall in RAM and NAND prices.
 
For myself, I'm waiting for 7nm, both CPU and GPU to do a full build. I want to go 4K HDR and all stars should be aligned by then, especially with the predicted fall in RAM and NAND prices.

Absolutely. Late 2019/early 2020 should be an incredible time for PC gaming. Hopefully 4k HDR high hz monitors are widely available by then so I can upgrade from my 1440p 144hz monitor.
 
By the way, if this will be a short generation, time wise, then ALL of the models are going to be released soon. 2070, 2060, 2050, etc. No drip, drip, but a full release to maximize sales in the time available.
 
By the way, if this will be a short generation, time wise, then ALL of the models are going to be released soon. 2070, 2060, 2050, etc. No drip, drip, but a full release to maximize sales in the time available.

It's almost pointless to do the smaller cards since they won't really but Turing with the the RT/Tensor HW. So you can just do some rebranding for lower end:

2060 is rebranded 1080 family
2050 is rebranded 1060 family.

Done.

Though I don't believe it will be a short release.

My bet is 2000 series will have a 2 year run, with 7nm die shrinks phased in.
 
It's almost pointless to do the smaller cards since they won't really but Turing with the the RT/Tensor HW. So you can just do some rebranding for lower end:

2060 is rebranded 1080 family
2050 is rebranded 1060 family.

Done.

Though I don't believe it will be a short release.

My bet is 2000 series will have a 2 year run, with 7nm die shrinks phased in.

We don't know if the CUDA cores for Turing offer any gaming advantages over Pascal yet. If the cores offer better performance-per-flop in games and better performance-per-watt (also helped with 12nm) then they could still release Turing x106 and x107. It is curious that the current rumours are 5GB 2060. Even if GDDR6 at 160-bit will still be faster than GDDR5 at 192-bit, why would they need to cut the ROPs and memory bus in the first place on an old Pascal card?

Maybe that rumour is nonsense, or true and they are simply adding a 2060 Ti that is the full card.
 
A few thoughts. If 2080Ti is dropping this soon, you know they got a 7nm refresh coming fast. This seems to be a GTX 480 vs GTX 580 refresh situation, but perhaps better because of the shrink. Good thing my 1080Ti is easily beast enough to get me to 7nm. Also, without HDR monitors available, I see little reason to increase GPU power just to get higher frames on old, out dated, washed out looking SDR monitor tech. They showed us the promise of HDR, so they shouldn't be surprised to find people holding off on a new GPU upgrade without an appropriate HDR monitor to pair with it.

Ray Tracing + HDR is the way forward, clearly. And right now, no one has proper HDR ultra wides or 1440p versions either (which is what most people want). Unless you want 27" 4K @ $2000 (lol) then you get to wait. I can't buy a faster GPU just for SDR. Its like taking a bath with socks on or something. Just seems pointless.

Lastly, did people seriously pay $3000 for a TitanV only for Nvidia to release new stuff with "RT cores" which, conveniently, the TitanV doesn't have? Good thing these folks can easily afford it, because they got burned to a crisp this time. This was an unapologetic torch to the face of Titan buyers by Nvidia this round. No RT cores. Dang man, that feels bad, right? Cruel IMO. Just cruel.
 
So if 2050 is a rebranded 1060 on 12nm, does that mean that it's still powered off PCIE? I'd have to think on the refined process and with lower clocks it's possible. There's a fair amount of demand for those cards since many systems don't have pcie connectors.
 
It's almost pointless to do the smaller cards since they won't really but Turing with the the RT/Tensor HW. So you can just do some rebranding for lower end:

2060 is rebranded 1080 family
2050 is rebranded 1060 family.

Done.

Though I don't believe it will be a short release.

My bet is 2000 series will have a 2 year run, with 7nm die shrinks phased in.
I (tend to) agree that we might see some Pascal GPUs step down a tier and renamed 2xxx, especially if they don't come with the RT additions. As to the cards lifetime, if the 2080Ti models are soon released, you're still stuck on 2 years? Remember the old adage.

A wise man changes his mind, a fool never will.

For my case, I generally support AMD and its products, underdog, good performance/$, etc. In this case however, Nvidia have the means to bury them for quite a while and I don't see why they will forgo all of the major benefits of 7nm to stick to some arbitrary 2 year release cycle. Stick the knife in deeper AND twist.
 
Pretty obvious fake to me, IMO Nvidia will definitely not reduce the amount of VRAM. Especially not if that card would be a 1080 performance-wise.
 
Pretty obvious fake to me, IMO Nvidia will definitely not reduce the amount of VRAM. Especially not if that card would be a 1080 performance-wise.

Amount of RAM is believable, but performance is not. It is a midrange product, so price must be reasonable- and 5GB is cheaper than 6GB and will perform the same at 1080p. 5GB also means 160-bit bus, which when using cheaper GDDR6 12Gbps will provide 240GB/s bandwidth.So we get perfect scaling of bandwidth per cuda core: 616/4352= 0.141 for 2080Ti, 448/2944=0.152 for 2080, and for 2060, which can be expected to have half the hardware of 2080- 2GPCs with 1536 cuda cores- 240/1536=0.156. Of course, smaller dies will boost higher- that is why giving them somewhat more bandwidth makes total sense.
 
Status
Not open for further replies.
Back
Top