Soldato
Anyone looking rationally and logically at it would agree with you. Its about the position in the product stack. People have been easily indoctrinated by marketing,when companies mess about with product stacks and naming,and end up selling you less for more. So they go by naming and expect the next generation with the same generational naming is a one for one replacement. It isn't. It happened before with smartphones,when companies introduced intermediate models. Its usually the new intermediate model above it,which is the true replacement with a price increase,not the one they think(and then people wonder why the similarly priced "replacement" is a sidegrade). It's basically a form of shrinkflation to increase profit margins.
With dGPUs,it's not only position in the stack,its the size of the dGPU which indicates the transistor count and memory bus. The larger the dGPU used,the higher the relative percentage compared to the top end,and usually the larger the relative memory bus(which means more memory bandwidth). The closer this is to the higher end model,the more the midrange is performant.
To go back a decade ago,a 60 series dGPU was using the second largest chip in the stack. So a GTX460TI was the second tier dGPU in the range,and third overall in the range. The same as the GTX560TI,HD6850,etc. This was the case for a decade until then. This is why mainstream dGPUs had huge improvements generation to generation on average. They were using the second best dGPU chip or even a salvage of a top tier chip.
By the time you got to Pascal,the 60 series had fallen to the third level dGPU,ie,two levels of dGPU chips above it. It was number five in the range. AMD could only compete with the second tier Nvidia dGPU(GP104),so even though Polaris was technically second in line,it was really a much lower tier chip. But this is when we started to see a period of stagnation.
Things got a slight rebalance with Ampere. So the RTX3060TI used a second tier chip,instead of the third tier chip hence why it is such a great performing dGPU for the price by modern standards. The RX6700XT used the second tier AMD chip which would make it a mainstream dGPU by old standards.
Now,with Ada Lovelace,the RTX4060,is now the fourth level. So you have an AD102,AD103 and AD106 above it. It's one tier down the dGPU stack than the RTX3060 was.
So it's not really mainstream/midrange anymore but uses an entry level/low end chip. Its equivalent in dGPU tier to a GTX950,GTX1050TI or RTX3050.The RTX4060TI is one tier down from the RTX3060TI position.
This is the big problem now with mainstream dGPUs. They are getting lower and lower down the stack,so the performance jumps at mainstream are getting relative worse,compared to the high end. Also at the same time the price is being pushed up.
Even though AMD is a bit better than Nvidia in this regard,people have forgotten the RX6600 replaced the RX5600XT/RX5700 at a higher price,and the RX6600XT replaced the RX5700XT at a similar price. Even at under £300 the RX6600 is not faster than the RX5700 and barely faster than an RX5600XT. The reference RX5700 8GB could be has for as low as £250!
The RX6600XT is the same speed as the RX5700XT but the RX5700XT could be had for not much more. But when you look at the same size of the dGPU,it's no wonder as on the same 7NM process it uses a smaller dGPU and less memory chips. It was cheaper to make. But the worse kicker,is the RX5700XT was originally supposed to be an RX680/RX690 before AMD pushed it up a bit higher because Nvidia jacked up pricing:
AMD’s RX 5700 XT was called RX 690 in its E3 presentation slides
Evidence of the disregarded graphics card branding comes straight from AMD's E3 press deck
www.pcgamesn.com
It actually was an RX480/RX590 replacement looking at die size,etc. AMD did an RX7900XT with it!
Last edited:
