Gotta keep those profits up ya know!Prices.So they call oficialy midrange 400+usd?Really great time for pc gamers..
2080super 700usd
2070super 500usd
2060super 400usd
https://videocardz.com/81149/nvidia...ies-final-pricing-and-availbility-date-leaked
https://videocardz.com/81105/nvidia...nd-rtx-2060-super-final-specifications-leaked
Differences:
2060 S: 2176 cores, 256-bit 8-GB memory, 175 W. Official claim is "15% AVG faster than RTX 2060" (almost no benefit from extra memory bandwidth)
2070 S: 2560 cores using TU104, 215 W (a bad sample), "16% AVG faster than RTX 2070"
Embargo ends 2019-07-02 06:00 AM PDT
You've been a broken record re 'mid-range' prices for years. The 2060 Super now has 256-bit bus + 8gb vram for only $50 more than the original 2060. Sounds fair to me considering the superior specs and performance todays mid-rangers have over the lower priced equivalents years ago. Are you expecting zero inflation, no rising costs for massive improvements to much more complex hardware for eternity? Get real man.Prices.So they call oficialy midrange 400+usd?Really great time for pc gamers..
2080super 700usd
2070super 500usd
2060super 400usd
https://videocardz.com/81149/nvidia...ies-final-pricing-and-availbility-date-leaked
Prices.So they call oficialy midrange 400+usd?Really great time for pc gamers..
2080super 700usd
2070super 500usd
2060super 400usd
https://videocardz.com/81149/nvidia...ies-final-pricing-and-availbility-date-leaked
Are GPU manufacturers making higher margins on todays products vs years ago?If each GDDR6 chip costs 10$ after volume bulk rabates - memory subsystem for 256 bit memory subsystem costs 80$, alone. Add to that GPU, add to that PCB, shrouds, etc, and suddenly you may wake up with 200$ manufacturing costs for single 445 mm2 GPU.
GTX 1070 vs 2070? Most likely - not.Are GPU manufacturers making higher margins on todays products vs years ago?
5years ago we had GTX960 4GB for 200usd7 years ago, we had the 660ti ( 2gb DDR5, 192-bit bus) for for $300. Based on production costs (I presume someone KNOWS the production costs of todays RTX 2060 6gb DDR6 and 8gb models) if they are going to argue that these cards should maintain that $300 point forever. Well, lets hear it.
7 years ago, we had the 660ti ( 2gb DDR5, 192-bit bus) for for $300. Based on production costs (I presume someone KNOWS the production costs of todays RTX 2060 6gb DDR6 and 8gb models) if they are going to argue that these cards should maintain that $300 point forever. Well, lets hear it.
The R9 390 was a refresh from the R9 290 that was released @ $399. You know why its gone down to $330 in the R9 390? The GTX 970, thats why:You could buy a 8GB GDDR5 512-bit bus R9 390 for $330 four years ago.
You could buy a 8GB GDDR5 256-bit bus RX 480 for $240 three years ago.
Gaming is not the sole revenue source for todays more diversified Nvidia. Also the windfall of the mining boom may have had something to do with record profits a couple years ago.The RTX cards since release have been poor in sales ("lower than expected" official statement).Nvidia's record profits and margins are also a big clue that their card prices have moved much quicker than manufacturing costs did.
As tech advances, generally things do get cheaper, sure. But this is NOT a blanket rule for everything tech related. GPUs are very different beast than CPUs which only consistent of dies.Usually, as technology advances things get quite a bit cheaper. For example, see what a $330 Ryzen 3700X gives you, compared to a $315 3770K Ivy bridge seven years ago, or even a $305 7700K released 3 years ago. You're welcome to compare the Ryzen 3900x and 3950x compared to 12 and 16 cores as well from 3 years ago. However, in the last 5 or so years on the graphics side, we haven't seen this happening at anything close to the usual rate.
Nvidias high end pricing, esp the 2080ti is OTTP, sure, even disgusting, no arguments there. But the RTX 2060 @349 is not unreasonable (given the obviously higher manufacturing costs and added complexity vs the 1060, ie, bigger dies, more expensive GDDR6), unless you can give us a breakdown of actual cost of parts, dies, vram, etc, to demonstrate your case rather than just assuming.The price of the top-end has ballooned to over $1000, while the performance for $300 has stagnated. We've barely seen a ~40% advancement in 3 years, especially since the release of Polaris and Pascal. According to TPU, the 1660TI (released Feb 2019) at $270 is about 28% faster than the 1060 (which launched at $300, July 2016) or 40% faster than the RX 480 (launched at $240 July 2016). In the same time-frame (actually a bit shorter), the 2080ti (launched September 2018) is 85% faster than the 1080 (launched may 2016) at 4K.
I heard that it costs Nvidia less than $300 to make a 2080Ti so you can imagine the kind of profit margins Nvidia is enjoying.7 years ago, we had the 660ti ( 2gb DDR5, 192-bit bus) for for $300. Based on production costs (I presume someone KNOWS the production costs of todays RTX 2060 6gb DDR6 and 8gb models) if they are going to argue that these cards should maintain that $300 point forever. Well, lets hear it.
Source? Link? Anybody can hear anything.I heard that it costs Nvidia less than $300 to make a 2080Ti so you can imagine the kind of profit margins Nvidia is enjoying.
i just saw the most recent video posted by Moores Law is Dead and he posted source in the video description but its not opening for me.Source? Link? Anybody can hear anything.
The bigger the dies, the lower the yields. So that guys speculation is useless without actual yield numbers.i just saw the most recent video posted by Moores Law is Dead and he posted source in the video description but its not opening for me.
Nvidia from single 300mm silicon wafer is getting 70 GT102 dies.I heard that it costs Nvidia less than $300 to make a 2080Ti so you can imagine the kind of profit margins Nvidia is enjoying.
Yeah which is why is crazy to believe they get away with selling it for $1200.Nvidia from single 300mm silicon wafer is getting 70 GT102 dies.
It is completely excluding the yields and functionality of the dies. Each GPU, ALONE can cost Nvidia 100$. Add to that 11 GDDR6 dies, and you get 210$ for just GPU die, and GDDR6 memory. 300$ can be in the ball park of the manufacturing costs of each RTX 2080 Ti.
Critical point. And how do we know the actual yields and functionality of the dies? Is there any source that publishes such data? What we do know is that bigger dies result in worse yields, functionality. Cant arrive at the numbers without all parts of the equation.It is completely excluding the yields and functionality of the dies...
Not really. Nvidia could sell GTX 1080 Ti for 700$, because the die was almost the same size as the die of RTX 2070, and the GDDR5 memory was going at the time for 5-8$. So you could get GPU and memory subsystem for under 120$ for that GPU, with total manufacturing costs being in the ballpark of 150-200$, at best for that GPU.Yeah which is why is crazy to believe they get away with selling it for $1200.
Well, 70 dies is the best case scenario. Each 16/12 nm wafer costs 6000-6500$. You can calculate the GPU costs yourselves, guys.Critical point. And how do we know the actual yields and functionality of the dies? Is there any source that publishes such data? What we do know is that bigger dies result in worse yields, functionality. Cant arrive at the numbers without all parts of the equation.