Good price for a classic midrange GPU. Ampere being launched tonight, but not likely there will be any mid-range option before 2021.
FWIW, this gives around the same performance as a GTX 1070, which still go for around $350 used on eBay.
Good price for a classic midrange GPU. Ampere being launched tonight, but not likely there will be any mid-range option before 2021.
FWIW, this gives around the same performance as a GTX 1070, which still go for around $350 used on eBay.
I'm buying whatever AMD RDNA2 comes out at 150w TDP … cant stand these super hot power hogs that the 3000 series is looking to be
3000 series likely to be more power efficient unless clock pushed up to the limit, will use 7nm process.
I'm sure both companies will have reasonable offerings at ~150w TDP. But not sure how long we will have to wait for these more mid-range or budget-range cards. I would really like both companies to release more budget-range options. For example I'd be happier if they had released the 5300xt to retail.
I would really like both companies to release more budget-range options. For example I'd be happier if they had released the 5300xt to retail.
For various reasons, including mining and the associated GPU boom, there's just been a massive amount of stock of last-generation mid-high end cards that make releasing low-end cards not financially feasible.
Even cards like the 5500XT and 1650 Super don't really make any sense when the RX 580 is still going around for less. If you look at the 5500XT and 1650 Super prices, they've been cut drastically over the past few months.
@p1 ama: I will admit I was dreaming. I understand that from the point of view of AMD & Nvidia it makes sense for them to try to make the cheapest gaming GPU to be mid-range. The higher the cost on the cheapest gaming GPU, the more profit they will both make. It's reasonable to be honest; just not what I would "like" (as a consumer with low budget and low requirements)
RDNA2 is unlikely to offer more performance per watt overall, but there are other factors here, like DLSS adoption, raytracing adoption, and FP32 performance of Ampere.
Even then, NVIDIA would likely still hold their typical architecture advantage, and are using a 7nm process. It's either equal to the process that AMD is using, or better.
Expect AMD to seriously push clocks, resulting in stupidly high power draw at stock. The current rumour for Ampere is performance per watt is 1.9x gen-on-gen, but that's with RTX features being used for both.
I'll generally agree. We all hope that AMD will deliver more & better; but hoping is different to expecting.
@pronoun: I think they've been staggering their architectural jumps too much, and they now have way too much work to do as they push towards a chiplet design.
Given the work they've done for Ampere, NVIDIA are going to be coasting for 5nm and 3nm designs.
Who knows what xx50 is going to be called lol. It could be the GTX 1750, the GTX 2650, the RTX 3050…
It'll be the 3050, because the RTX performance leap will see every card in the stack being capable of delivering acceptable performance with raytracing on, even cards that won't perform as well as the 2060 in traditional rasterisation graphics.
gives around the same performance as a GTX 1070
Which i bought in May 2016 for $445 AUD, crazy how GPU price/performance has legit stagnated for half a decade
I bought a GTX1070 in June 2016 for $700 and that was the cheapest I could find. How the hell did you only pay $445?
EDIT: Was 555 via Amazon US, just went way back in my emails haha
That's still pretty bad for price to perf over time though.
@Void: And all while nVidia's share prices are skyrocketing.
@Void: How?
That's still pretty bad for price to perf over time though.
What are you talking about? $555 to $358 is a 36% decrease in price…
Either way, this is pretty representative of the long-run trend.
The 760 was $249 USD at launch, the 960 was $199, the 1060 6GB was $199/$249 (3GB/6GB), the 1660S is $229 USD MSRP.
Nvidia's product stack has moved more and more upmarket over time, primarily (IMO) due to two main reasons:
(1) AMD has largely been uncompetitive, meaning that Nvidia is able to use more cut-down chips to maintain the generation-on-generation performance increases that they've been seeing over time.
(2) Nvidia's generational improvements have largely outstripped the increase in graphical requirements for games and the desire to play at higher and higher resolutions. FWIW, 1080p is still the most popular resolution today, and it has been for like the last 10 years. 1440p has gradually increased in popularity over time, and 4K hasn't reached any level of popularity at all really. Though I suspect this is due to poor DPI scaling in Windows making 4K displays really difficult to deal with.
The net result of this is that in many generations past, you really needed high-end GPUs to get playable framerates in most games at reasonable settings, whereas today, you get a similar gaming experience from very mid-range cards.
I've been buying GPUs for a long time. Back in the day, if you wanted to game at high settings, you would be looking at the very least a 6800 GT (or even 6800 Ultra), then maybe a 7800 GT, 8800 GT…etc. Those "top end" GPUs cost around $500 USD back then. Today, $500 USD will get you probably something like a 2070 Super, which will give you a similar relative experience to those cards back in the day.
The point is that Nvidia is practically a monopoly, but they are still making remarkable technological advancements which are just outstripping what is required.
If you upgraded from a 780 Ti to 980 Ti to 1080 Ti, for instance, yes you paid more and more each generation, but you quite literally doubled your performance year on year as well.
@p1 ama: 36% doesn't sound all that impressive for 4 years.
@p1 ama: The 760 was $249 USD at launch, the 960 was $199, the 1060 6GB was $199/$249 (3GB/6GB), the 1660S is $229 USD MSRP
Wish you could use AUD instead of USD. The 229 USD Supers were marketing gimmicks. They were non existant even in US. They all star at 400 AUD/290 USD.
Kotaku has archives for GPUs launch prices. This is the worst GPU prices ever. Glad to see the back of this gen. RDNA1 was a flop. NVIDIA investors were disapointed by RTX sales.
https://www.kotaku.com.au/2019/10/nvidia-gtx-1660-super-aust…
Unfortunately since there is no competition from AMD, once again have an open goal and thus will charge an RTX tax. 400USD / 700AUD will be the cheapest RTX tomorrow, the 3060. Prices are only going up, NVIDIA has no incentive to do otherwise.
You're using current exchange rates to present US pricing based on AU price history.
You're not removing GST.
The nature of the Turing architecture means we haven't even seen its best performance yet. Unfortunately hardware comes before software when it comes to certain features, and people were making the same shortsighted comments when the 900 series debuted.
If you go back and look at the price history on this website, you will see really well priced 1660 Supers dotted every couple of months right up until March this year. Please don't be part of the brigade of idiots who can't look past their flawed napkin math.
How does it compare to this? RX 5500 XT at 312~
maybe wait for rtx 4090, rumour the performs is 5 times to the just announced 3090
But It'll be 80% slower then the 5090
That's MSRP.
In terms of traditional mid-range, the 3060 will be launched this year. The 3050 though, likely in January alongside mobile parts.