the only reason i have tendency to believe MS went with vega is that leaks for PS5 said its clocked at 1850 MHZ(which at the time sounded insane) and Dante dev kit at 1410 MHZ(already pushed to the max apparently based on the leak) . this to me means one is vega and one is navi given what we know today after amd computex.
It's clearly a troll at this point. I wouldn't bother.
According to the wccftech leak that said navi will be post GCN, yes the 64 CU limit is broken.
Wait did she actually say that? I woke up mid conference and missed that lol.
you said no numbers were given for thermals when there AMD literally gave us the 1.5x performance per watt number.
You're responding (I assume) to a person who is one of the top posters in this thread XD lol. The irony..
Agree. Best AMD presser in a long long time.
Nah, the only thing what I really have to change is to get rid of CUs as this is a thing of the past now. For my prediction I will only use stream processors for future releases of my prediction. I think the ECC vote is spot on, so I will go from there. So I really does not have to change a lot. For me the numbers in efficiency gains only confirms what I was already thinking. The updated version looks like this:
I don't think this dude is verified. Just using an avatar he made on mspaint.
Seems like our joke of navi being much more efficient ended up being true haha, but that is for the better as everything they talked about sounds good for consoles.Good morning everyone!
Just woke up and had to read up for the NAVI part of the AMD keynote. Zen 2 still a todo.
RNDA is indeed a welcome surprise. I am glad that the GCN rumors were false. As we don't know yet how the GPU architecture is organized is too early to say more but the numbers look promising. If one of their middle segment GPUs can beat a RTX 2070 is a good sign. How to judge about it though is a matter of the information that we will get at E3. It will be a long two weeks LOL
Agree. Best AMD presser in a long long time.
Nah, the only thing what I really have to change is to get rid of CUs as this is a thing of the past now. For my prediction I will only use stream processors for future releases of my prediction. I think the ECC vote is spot on, so I will go from there. So I really does not have to change a lot. For me the numbers in efficiency gains only confirms what I was already thinking. The updated version looks like this:
The opposite is the case imo - a general message to those going crazy with high computational power numbers!
Looking at those numbers in efficiency gains I see that Sony and MS doesn't need to pull out those TF numbers as they a 25% performance gain per clock which is even above what I had expected (15%). So I think I even have to reduce my prediction on the upper bound. I still hear the words of Lisa Su: "We like small". This also allows to better control power consumption and TDP for the consoles, so they can save on the cooling front and have a more "green" solution for the living room with a small form factor.
We will find out at E3 how much how I think about will materialize as true.
Good morning everyone!
Just woke up and had to read up for the NAVI part of the AMD keynote. Zen 2 still a todo.
RNDA is indeed a welcome surprise. I am glad that the GCN rumors were false. As we don't know yet how the GPU architecture is organized is too early to say more but the numbers look promising. If one of their middle segment GPUs can beat a RTX 2070 is a good sign. How to judge about it though is a matter of the information that we will get at E3. It will be a long two weeks LOL
Agree. Best AMD presser in a long long time.
Nah, the only thing what I really have to change is to get rid of CUs as this is a thing of the past now. For my prediction I will only use stream processors for future releases of my prediction. I think the ECC vote is spot on, so I will go from there. So I really does not have to change a lot. For me the numbers in efficiency gains only confirms what I was already thinking. The updated version looks like this:
The opposite is the case imo - a general message to those going crazy with high computational power numbers!
Looking at those numbers in efficiency gains I see that Sony and MS doesn't need to pull out those TF numbers as they a 25% performance gain per clock which is even above what I had expected (15%). So I think I even have to reduce my prediction on the upper bound. I still hear the words of Lisa Su: "We like small". This also allows to better control power consumption and TDP for the consoles, so they can save on the cooling front and have a more "green" solution for the living room with a small form factor. Not an unimportant feat in the times we live in.
We will find out at E3 how much of how I think about will materialize as true. As mentioned above 2 more weeks of waiting ...
to me, the biggest holy shit moment of the conference was the benchmark demo they ran comparing a Zen 2 CPU with a Navi GPU to an intel 9700k and a RTX 2080ti. the 2080ti could only manage 14 fps while the Navi paired with Zen 2 and whatever the hell PCIE4 is was able to do 24 fps.
I was not joking about that. Actually it was one of my main arguments for not seeing numbers around and above 12TF.
Okay, sorry for misunderstanding your comment LOL. Still not fully awake.
It doesnt make much of a difference because games are designed around a slow HDD. Next gen the baseline of storage bandwidth is much higher and so will it dramatically effect game performance.
Be happy that your biological clock did not wake you up in 5:40 am exactly when the navi section started lol.Okay, sorry for misunderstanding your comment LOL. Still not fully awake.
I am still like this ...
The navi chip releasing in july are the mid range navi 12 cards, PS5 is based off navi 10, which is larger than the navi 12
The gonzalo leak connected ps5s gpu to the navi 10 chip. It is a deriviative of the navi 10.
