• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

RDNA 3 rumours Q3/4 2022

Status
Not open for further replies.
ADA not even out for a whole month and Ampere already abandoned?

amp.png
 
The 4090 is not the full Ada die and they have quite a bit of headroom there so I expect when AMD launches the 7950XTX, they will respond with a 4090 Ti and a 600W TGP.
I’m not convinced they do. They’ve already cranked up the power and have a die which is 608mm2. I believe the max die size for 5nm is 800(ish)mm2.

And given nVidia have to cram everything onto one die, they can’t just fill that extra space with more shaders as they’ll need to add all the other components of a chip.

AMDs chiplet design means their main chip is 350mm2 and it is filled with shaders.
 
Last edited:
I didn't think you would be so upset by my comment about you not watching the video, that you would lash out by calling me foolish and make baseless accusations.


Oh so you did watch the video but decided to ask a question that was answered in the video because reasons.


So the one question I did ask you, you don't answer. But you did decide to run your mouth on some pointless tangential rambling.
I am sorry if I offended you, I wont be following up this discussion as there is no point. I am free to believe that Adored claims are BS, you are free to believe them to be correct. Peace.
 
I am sorry if I offended you, I wont be following up this discussion as there is no point. I am free to believe that Adored claims are BS, you are free to believe them to be correct. Peace.
Writes an entire post full of baseless accusations. Signs it off with peace. :rolleyes:
I guess you tried.
 
I’m not convinced they do. They’ve already cranked up the power and have a die which is 608mm2. I believe the max die size for 5nm is 800(ish)mm2.

And given nVidia have to cram everything onto one die, they can’t just fill that extra space with more shaders as they’ll need to add all the other components of a chip.

AMDs chiplet design means their main chip is 350mm2 and it is filled with shaders.

What? The 4090 is already the full size die, the die wouldn't be any bigger than the 608mm2 it already is.

The only difference is some of the cores on the 4090 are DISABLED and that's the difference. To be exact, 12% of the cores on the RTX4090 are disabled/turned off and would be turned on for a RTX4090 ti
 
Last edited:
What’s giving me pause is this: I can’t work out why, if AMD have a part that is within spitting distance of the 4090 on raw performance, they have let NVidia run rampant in the halo segment for 6 weeks and hoover up the early adopters without any significant competition. Which leads to the idea that they don’t have such a part.

But I suppose we find out in 36 hours or less either way…
 
It's a disgrace what Nvidia has done to the 90 class and not the first time with 90 class, the original 90 class were dual gpu cards not a single gpu and then they made them single gpu pro cards with NVLINK and then even removed that now. It's a P***take really.

They also destroyed the 80 class as instead of it being a £700 segment card, it now takes £1300. If you want to spend your £700 in the Ada lineup then you are only getting a 4060Ti which will be the same as a high end Ampere card that you could have had for two years now. Yeah well done nvidia.
 
The latest MLID video is out with new info.

Confirms that rdna3 is launching as an efficiency and cost focused architecture not max performance. Price will definitely be lower than Nvidia.

* 7700XT is not faster than the 6950xt

* Reference 7900XTX is 350w TDP and 2x8 pin, but AIB models are 450w TDP and 3x8 pin and many will be using the same 3/4 slot coolers used on the RTX4090

* 7900XTX does not match RTX4090 in rasterisation or ray tracing. However with the extra 100w TDP on AIB models they might get some really nice performance gains with overclocking so it's up in the air to how fast it can get.

* 7900XT and XTX will be in good supply, AMD is confident in its decision to front load supply to premium models after seeing how well the RTX4090 has sold.

He doesnt say that at all, he says he expects the 7900xtx to trade blows with the 4090 depending on the game. And that is such a safe position to take

Heres something to consider, amd have already stated that rdna3 has >50% better performance per watt. If we extrapolate out exactly 50% from the 6900xt 300w performance to a 350w reference card, based on techpowerup numbers the 7900xt is going to curb stomp the 4090. To only match the 4090 performance, based on 50% greater performance per watt the 7900xt would only need 250watts

As for the 7700xt, greymon on twitter has already stated that we can expect 6950xt performance in a latop using navi32, which would almost certainly be limited to ~150w power. a 200-230w 7700xt using N32 chip will absolutely beat the 6950xt if that is true
 
What’s giving me pause is this: I can’t work out why, if AMD have a part that is within spitting distance of the 4090 on raw performance, they have let NVidia run rampant in the halo segment for 6 weeks and hoover up the early adopters without any significant competition. Which leads to the idea that they don’t have such a part.

But I suppose we find out in 36 hours or less either way…
A company's release timeline isn't too flexible. You can't just bring things forward without causing problems.

Also 6 weeks isn't that long considering these cards will be around for the next 2 years.
 
He doesnt say that at all, he says he expects the 7900xtx to trade blows with the 4090 depending on the game. And that is such a safe position to take

Heres something to consider, amd have already stated that rdna3 has >50% better performance per watt. If we extrapolate out exactly 50% from the 6900xt 300w performance to a 350w reference card, based on techpowerup numbers the 7900xt is going to curb stomp the 4090. To only match the 4090 performance, based on 50% greater performance per watt the 7900xt would only need 250watts

As for the 7700xt, greymon on twitter has already stated that we can expect 6950xt performance in a latop using navi32, which would almost certainly be limited to ~150w power. a 200-230w 7700xt using N32 chip will absolutely beat the 6950xt if that is true

The 4090 is 75% faster than the 6900XT at 4K. https://www.techpowerup.com/review/nvidia-geforce-rtx-4090-founders-edition/32.html

Add 50% to PPW the 6900XT takes it from 57% to 85.5%, the extra 50 watts is 17%, add that to the 85.5% you get exactly 100% of the RTX 4090 at 350 watts.

Its almost too perfect, i think trading blows is accurate.
 
Status
Not open for further replies.
Back
Top Bottom