• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Question 'Ampere'/Next-gen gaming uarch speculation thread

Page 92 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
O

Ottonomous

Senior member
How much is the Samsung 7nm EUV process expected to provide in terms of gains?
How will the RTX components be scaled/developed?
Any major architectural enhancements expected?
Will VRAM be bumped to 16/12/12 for the top three?
Will there be further fragmentation in the lineup? (Keeping turing at cheaper prices, while offering 'beefed up RTX' options at the top?)
Will the top card be capable of >4K60, at least 90?
Would Nvidia ever consider an HBM implementation in the gaming lineup?
Will Nvidia introduce new proprietary technologies again?

Sorry if imprudent/uncalled for, just interested in the forum member's thoughts.
 
moonbogg said:
So we can expect 2-sided water blocks? It would have to sandwich the card between two blocks and have the inlet on one block and the outlet on the other block with a flow connector between the blocks, right? If that's the case, I'd imagine water blocks to be more expensive as well.
Click to expand...
Doubtful. Really, with a full coverage block on the front you probably wouldn't see a whole lot of benefit to another one on the back side for memory. Encapsulated BGA chips like that have a reasonably good path to the board as well. Keep the board cool, slap a backplate on and be done with it.
 
Hitman928 said:
Both are intakes based upon the fan blade shape. From what I can tell and based upon the assumption that under the shroud there is a path for air flow from the fan on the front of the PCB to the fins in the middle X section, here is the air flow that I think is happening.

The front of pcb fan takes are from the bottom side of the case (standard) and pushes are out the back of the case as well towards the back of the card (front of the case) with some air coming out towards the bottom of the case as well, but most pushed around the 2nd fan towards the front of the case.

The back of pcb fan takes air from the middle of the case and pushes it straight through towards the bottom of the case. This design would push most of the hot air towards the front and bottom of the case. In an open air setup this wouldn't be an issue at all but in an actual case it could become a problem with trying to get cool air to the front side fan while removing the hot air coming off the card.

Ideally you'd probably want fresh air flowing from the top or upper back of the case and have fans pushing air out in the front and bottom of the case with a side panel fan towards the back of the case feeding the front side GPU fan fresh air. Ventilated PCI slot covers on the case would probably work OK too.
Click to expand...
The raw cooler image helps with understanding I find. You can see the central fins are all blocked off from the front of the cards, while the shroud blocks airflow from exiting top and bottom. So, as you said the front fan pushes air down and out the back. The central fins don't seem to have airflow at all really. The back fins are connected to the vapour chamber with four big heat pipes, and probably provide the majority of the cooling since the airflow is so much less restricted.
1597954839172.png

I just don't understand yet why they didn't just put a slight s-bend in the heat pipes past the board to put them close the the level of the PCB, and mount both fans on the same side. Outside of aesthetics, it's nice in a standard case to have them down facing because it's a lot harder to drop something in the fan that way.
 
aigomorla said:
Since the 3090 is now officially spoiled, I am assuming there is going to be no 3080ti, as i hear the 3090 is now the official flag ship.
However is there any chance that you guys think Nvidia will pull a 3090Ti, like a few months after launch and rub salt on early adopters?
Click to expand...

I think if they release a 3080 and 3090 with no Ti variants, then Ti variants for both are all but guaranteed in my opinion. I'd expect prices of Ti variants to be the same or very similar to the 3080 and 3090, but with 3080Ti performance basically matching 3090 or coming close at basically half the price, lol. So yes, 3090 folks will get burned as usual, just like Titan buyers always do before the Ti drops. But these folks are fine with it and they know what to expect by now, so no harm no foul. I wouldn't be surprised if the 3080Ti comes out on TSMC 6 or 9 months after the initial release with better clocks and power efficiency, much like the GTX 580 after the GTX 480.
I do think Nvidia will drop something great for 1080Ti folks either now or with a 3080Ti later. It's just too big of an enthusiast segment to totally ignore forever IMO. I expect an impressive performer in the $7-800 range this time around. It makes no sense to go 2 entire generations without offering something to their traditional enthusiast crowd.
That being said, I could so easily be absolutely wrong of course. 1080Ti folks could easily be left with no options to upgrade again.
 
jpiniero said:
That's why they are using the double sided cooler.
Click to expand...
A double sided cooler would violate ATX specs and block off cardslots, so NV wouldn't build a reference design that way. Or well, they could, I suppose, but they never have so far so I consider that an unlikely alternative. 🙂

jpiniero said:
So you're saying the fan is on the side of the PCB? Doesn't make any sense, that's stupid.
Click to expand...
As you may be aware, the PCB only covers about 2/3 the length of the cooler, which means the fan is free to push air unrestricted straight through the heatsink, which of course improves airflow and efficiency quite dramatically.

Personally I believe the fan is on the reverse side of the cooler because of the very thick heatpipes going through much of the heatsink. It would be difficult or maybe impossible to fit the fan on the front side of the sink with those heatpipes there in the straight and level configuration we see them in, and if the pipes are kinked to make more room the heatsink would be greatly more difficult or impossible to build as it has been depicted.

MrTeal said:
The central fins don't seem to have airflow at all really.
Click to expand...
It's possible that in the final production version the front side fan will be able to push air into the central fin section. It'd be stupid quite frankly if that wasn't the case, as the central fins would essentially be useless decorations without active airflow... 😛
 
Did anyone ever consider that maybe NVidia leaked that image themselves, then conducted a "fake" investigation into who may have leaked it to really throw off the tech websites? As a form of misdirection of course.
 
A/// said:
Did anyone ever consider that maybe NVidia leaked that image themselves, then conducted a "fake" investigation into who may have leaked it to really throw off the tech websites? As a form of misdirection of course.
Click to expand...

I dunno, doing a unique looking card does sound very nVidia and it's going to be drawing a ton, soo...
 
FaaR said:
It's possible that in the final production version the front side fan will be able to push air into the central fin section. It'd be stupid quite frankly if that wasn't the case, as the central fins would essentially be useless decorations without active airflow... 😛
Click to expand...
Yeah, but...
1597973404933.png
 
MrTeal said:
Yeah, but...
Click to expand...
A DIMM doesn't have a (rumored) power dissipation of upwards of 350W though. 🙂

When you need to cool off that high a wattage I would think you'd want to use every opportunity to do so. Without airflow in the midsection there would be a large portion of the cooler doing basically nothing, which would be a really, really strange design choice I'd say.
 
FaaR said:
A DIMM doesn't have a (rumored) power dissipation of upwards of 350W though. 🙂

When you need to cool off that high a wattage I would think you'd want to use every opportunity to do so. Without airflow in the midsection there would be a large portion of the cooler doing basically nothing, which would be a really, really strange design choice I'd say.
Click to expand...
350W?

Its 50W less than 10th gen Intel CPUs use.

Im sorry. 😛
 
aigomorla said:
Since the 3090 is now officially spoiled, I am assuming there is going to be no 3080ti, as i hear the 3090 is now the official flag ship.
However is there any chance that you guys think Nvidia will pull a 3090Ti, like a few months after launch and rub salt on early adopters?
Click to expand...

What's in a name? Before the 2080 Ti was the top card. Now the top card is the 3090. The stack hasn't changed, just the names.

If we do see refreshes later on I expect they'd be the 3080/3090 SUPER instead of bringing out the Ti badge.

If you want a more cynical take it would be that they aren't using Ti this time around because they don't expect it to be enough of an upgrade to justify the moniker.
 
Glo. said:
https://twitter.com/x/status/1296891307999338496

So yeah, its triple slot...

P.S. It appears that rumors that next gen Nvidia GPUs are a complete power hogs, were complete bollocks, right?
Click to expand...

That thing is massive. Also clearly shows no air flow for the fins in the middle section which I still don't understand. Seems to me (again, not mechanical engineer here so maybe I'm just mistaken) that they could have sacrificed some aesthetics for improved cooler functionality and space efficiency but we'll see how well it does in the reviews. That thing looks like a beast. Hopefully it has the performance to match, lol.
 
That's extra thicc (and tall, since heatsink extends above the I/O bracket).

Space hog optimized for 1-GPU (maybe 2) systems: all client (not HEDT) ATX motherboards have 3 slot clearance between first and second PCIE x16 slots. More than 2 requires liquid cooling.

CastleBravo said:
Wow, are those 120mm fans? They look high-rpm too.
Click to expand...
Taking the height of the I/O bracket to be 120 mm, fan diameter is about ~105-110 mm.
 
RTX 2080 Ti was able to be fit in most mITX cases.

No more High-End small form factor builds, I guess... 🙁
 
So, looking back at the 3080 leaked heatsink, that one is definitely two slot and the fan appears to be smaller, more like a 92mm. This seems to be a different heatsink in the same vein, but wow it's massive. It should be quite quiet though, hopefully.

Edit: What's the chance this is just a good photoshop from the previous leaks? A triple slot card that large is just a massive jump.
 
Glo. said:
RTX 2080 Ti was able to be fit in most mITX cases.

No more High-End small form factor builds, I guess... 🙁
Click to expand...
If I'm measuring it right compared to the 267mm 2080 FE, this card is ~326mm long. I had a triple slot dual 290X Devil 13 that weighed 5 lbs and came with a little stick that you used wedged between the card and the bottom of the case for extra support. That card was 305mm long. There's a lot of ATX cases that something like this wouldn't fit in.

This almost has to be fake.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top