• 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.

Rumour: AMD's 4096 Shader Big Navi in Q4 2019?

Last edited:
s6545 said:
All that's been suggested so far by AMD is RDNA 2 (7nm+) in 2020:

https://www.pcgamesn.com/wp-content/uploads/2019/06/ammd-rdna2-2020-580x326.jpg

An unofficial roadmap suggests Q1/Q2 for Navi 20 (7nm):

https://www.pcgameshardware.de/screenshots/1020x/2019/06/GPU-Roadmap-Juni-2019-pcgh.png

It's possible the Navi 20 in 2020 release is for desktops and that RDNA 2 is for the next Xbox console.

Sorry if thats disappointing, but you won't have to wait for too long.
Click to expand...

Could just be AMD not showing their hand though.
 
shankly1985 said:
Could just be AMD not showing their hand though.
Click to expand...

I hope so. It maybe the smart thing to do for them, to catch Nvidia off guard. But you gotta wonder why, if they have something, why they haven't already released it, especially since the RTX 2070 Super beats everything AMD is offering.
 
Last edited:
s6545 said:
I hope so. It maybe the smart thing to do for them, to catch Nvidia off guard. But you gotta wonder why, if they have something, they haven't already released it, especially since the RTX 2070 Super beats everything AMD is offering.
Click to expand...

Because they are holding the good stuff back for the people who make them most of their GPU revenue: Microsoft and Sony.
 
Panos said:
From hints buildzoid & Jay2cents have said that CES2020 will be the date the bigger GPUs will be out. That means well into March next year to see them.

And I hope AMD & Nvidia come with a big indium soldered IHS for the next 7nm chips, because 251mm^2 is too small surface to cool 200W effectively on air.
Click to expand...

You realise the contact area to the heatspreader is the same size, but it adds two thermal interfaces before hitting the heatsink itself.

Heatspreaders are for pressure and protection of dies, they are worse than direct die cooling. There is never a reason to thermal interfaces and a thin pieces of metal will spread heat better than going direct to the heatsink and one thermal interface between them.

A heatsink IS a heatspreader. Direct die is superior, always, but depending on die/pressure for mounting/etc, it's often deemed less safe. CPUs generally have heatspreaders much more often because of the installation/mounting need for cpus while gpus rarely have the cooler removed.
 
Panos said:
From hints buildzoid & Jay2cents have said that CES2020 will be the date the bigger GPUs will be out. That means well into March next year to see them.

And I hope AMD & Nvidia come with a big indium soldered IHS for the next 7nm chips, because 251mm^2 is too small surface to cool 200W effectively on air.
Click to expand...

These 200W will be down by a fair amount if the GPU is kept cooler. The higher the temperature, the higher the TDP is.
If you keep the GPU around 50°C, the TDP will be down at least to 175W.

drunkenmaster said:
You realise the contact area to the heatspreader is the same size, but it adds two thermal interfaces before hitting the heatsink itself
Click to expand...

251 sq.mm is not that small. In fact, they have always been using GPUs with similar sizes but prior to Navi's blower, we had never witnessed such a poor thermal behaviour.

It must be a thermal pad and thermal paste, plus very small heatsink plus inefficient turbine, which make all the heat trapped in the GPU which reaches mind-blowing 110°C.
Then, the heat slowly escapes to the other metal parts on the card, and you basically get a toaster to cook some food with it.
 
4K8KW10 said:
251 sq.mm is not that small. In fact, they have always been using GPUs with similar sizes but prior to Navi's blower, we had never witnessed such a poor thermal behaviour.
Click to expand...

You continue the misinformation. The last blower Nvidia used was on the GTX1080Ti FE and had the same thermal and acoustic issues with the chip burning at 84C core.
Seems everyone is ignoring that GPU somehow. And both the Pascal Titans but some might argue XYZ for them.
These are in comparable transistors count to the 5700XT.

As for the "never witnessed" please never ever use the word never.
At high power >200W GPU between 250mm^2 and 260mm^2. The closest Nvidia has come was the 260mm^2 9800XT+, with 0.754billion transistors. And was horrible to cool.
AMD 270X/370X are close in size but didn't burn more than 90W while heavily overclocked.
GTX 760 reference blower, run again very hot hitting 90C+, and that at sub 190W package at 294mm^2.

So tell me which GPU with similar size at same power consumption we had before. Last time checked there was none, but I would truly like to know also.
And here is the list to help you with the Nvidia top GPUs that burn that amount of power

https://www.reddit.com/r/nvidia/comments/9bwihc/major_nvidia_die_size_comparison/
 
4K8KW10 said:
Better don't deal with things which you don't understand :D
Click to expand...

You seem to be unable to grasp that die size alone means very little, thermal density does. Essentially the metric you want is watts/mm^2 being dissipated. In that comparison 251mm^2 is VERY small and has extreme thermal density, that is why it's hard to cool. Regardless the point here is that a heatspreader is not really a 'heat spreader', and never has been. It's about protecting the die, it's a piece of metal, nothing more or less, so is the heatsink. Adding extra poorly (comparably) conducting interfaces reduces cooling, always. Heatspreaders as used on a cpu, don't improve cooling.
 
Poneros said:
The new scoop is:
Arcturus = HPC = not desktop
Click to expand...

Yes, so far there's been no evidence that Arcturus is for gaming, in fact one of the guys at AMD hinted that it had no graphics capability. I think it's purely for 'compute' related tasks, e.g. not aimed at consumers.

One possibility is that it could be used in the PS5 for Ray Tracing (working alongside the GPU), as the GPU itself apparently does not support ray tracing. I doubt this though, it would just increase the cost.
 
Last edited:
Poneros said:
The new scoop is:
Arcturus = HPC = not desktop
Click to expand...

AMD doesn't use separate architectures for HPC and desktop.

AMD-Arcturus.jpg


AMD Next Generation Arcturus GPU To Succeed Navi GPU – New Radeon Family Naming Scheme Also To Be Introduced After Navi
https://wccftech.com/amd-arcturus-7nm-next-gen-gpu-revealed/
 
You must log in or register to reply here.
Back
Top Bottom