Nvidia Ampere Discussion [2020-05-14]

[CarstenS]

Ampere sales launch DDOS:en newegg web page exactly at sales launch time. I suppose there is couple of people trying to buy 3080

Bestbuy's search feature broke temporarily but is now working. I bet it was also due to 3080 purchasers.

edit. Scrap that, best buy search is again broken. I guess people try searching again and again to see which model is in stock

Probably not in stock anyway. Nvidia webpage in germany lists 3080 as not available.

 

[manux]

Probably not in stock anyway. Nvidia webpage in germany lists 3080 as not available.

I know some folks that got orders in though no one who I know bought from nvidia store.

edit. Probably this supply in, supply gone will keep happening just the same as with turing launch. Though my gut feeling is ampere is more in demand, turing wasn't the best proposition with no games using the new features and dlss1.0 being really bad.

 

[NightAntilli]

I know some folks that got orders in though no one who I know bought from nvidia store.

edit. Probably this supply in, supply gone will keep happening just the same as with turing launch. Though my gut feeling is ampere is more in demand, turing wasn't the best proposition with no games using the new features and dlss1.0 being really bad.

This looks more like a disguised paper launch.

 

[DegustatoR]

Well. Bye.

 

[manux]

This looks more like a disguised paper launch.

Go check what's happening with ps5 preorders. It's a real launch with products out there but due to real world issues in how manufacturing chips and products work the initial supply tends to be low while initial demand is high. Things will settle out over time, demand lowers and product availability gets higher.

 

[LordEC911]

Turing was 16nm relabelled as 12nm, versus Ampere 10nm relabelled as 8nm.

28nm was about the last time that it was possible to look at the node's "gate length number" and determine meaningful scaling comparisons against older nodes.

You will probably have to go back about a decade to find a time when transistor density for AMD and NVidia GPUs on the same node were actually meaningfully comparable. So you can't bring AMD into this.

Minimum metal pitch is probably the closest we have to a rough density metric between different nodes.

I would agree about not being able to compare anymore.
AMD likely was unable to use the full benefit of TSMC 7nm with RDNA1 while Nvidia has the ability to customize the Samsung 8nm process to a degree.

 

[DegustatoR]

With the emergence of low level graphics APIs such as DirectX 12 and Vulkan, game developers are able to implement SLI support natively within the game itself instead of relying upon a SLI driver profile. The expertise of the game developer within their own code allows them to achieve the best possible performance from multiple GPUs. As a result, NVIDIA will no longer be adding new SLI driver profiles on RTX 20 Series and earlier GPUs starting on January 1st, 2021. Instead, we will focus efforts on supporting developers to implement SLI natively inside the games. We believe this will provide the best performance for SLI users.

Existing SLI driver profiles will continue to be tested and maintained for SLI-ready RTX 20 Series and earlier GPUs.

For GeForce RTX 3090 and future SLI-capable GPUs, SLI will only be supported when implemented natively within the game.

https://videocardz.com/press-release/nvidia-sli-support-transitioning-to-native-game-integrations

It's dead, Jim.

 

[trinibwoy]

Well. Bye.

Not surprising. End of an era.

 

[Babel-17]

I was just at my local Best Buy store, and unlike the last two releases, they had received no stock. It was an online sale apparently, and they sold out immediately. I had applied to be notified about availability, and I checked after midnight and the Founders Edition was still "Notify Me" only.
Never notified, never any in store stock, this was not the best way to handle limited inventory. Great way to drum up desirability, and fend off interest in AMD's upcoming challenger, and I suppose all will be forgiven if nVidia can supply the demand without people having to hover over their keyboards while waiting for the click of the hour.

 

[manux]

I was just at my local Best Buy store, and unlike the last two releases, they had received no stock. It was an online sale apparently, and they sold out immediately. I had applied to be notified about availability, and I checked after midnight and the Founders Edition was still "Notify Me" only.
Never notified, never any in store stock, this was not the best way to handle limited inventory. Great way to drum up desirability, and fend off interest in AMD's upcoming challenger, and I suppose all will be forgiven if nVidia can supply the demand without people having to hover over their keyboards while waiting for the click of the hour.

I tried to get one from bestbuy web pages. Search broke so I wasn't able to find the cards. By the time search worked everything was gone.

 

[NightAntilli]

Let's check the numbers. ga102 die size is 628mm2 and has 28billion transistors. This leads to 44.6M transistors per mm2. tu102 was 754mm2 and 18.6B transistor leading to 24.7M transistor per mm2(tsmc 12nm aka enhanced 16nm?).

GA100 on tsmc 7nm is 826mm2 54B leading to 65.4M transistor per mm2.

navi10 on tsmc7 nm is 251mm2 and 10.3B leading to 41M transistors per mm2.

Samsung 8nm doesn't look that bad density wise. The usual rules of manufacturers not telling how they count transistors, cache versus logic having different density etc. apply. Above numbers should give good ballpark.

Density indeed seems ok. But I don't think you can necessarily discern the quality of the node based on the resulting product. The architecture itself also influences how dense you can be on a node. Additionally, density itself does not really give you any indication regarding other important factors like heat generation or yields. Samsung's 8nm is supposedly cheaper for a reason, I presume.

As a side-note, The Radeon VII is around 40MT/mm2, the 5700XT is 41MT/mm2 as you mentioned, and the XSX is around 42.5 MT/mm2. There is constant improvement here, on the same node. Granted, the XSX chip is not directly comparable to a standalone GPU, but still...

 

[manux]

Probably not in stock anyway. Nvidia webpage in germany lists 3080 as not available.

3080 became available in finnish nvidia webpage about 10 minutes ago. You might get lucky if you try german store now. Or perhaps cards are already gone.

 

[CarstenS]

3080 became available in finnish nvidia webpage about 10 minutes ago. You might get lucky if you try german store now. Or perhaps cards are already gone.

3090 or bust (or Big Navi)

 

[Man from Atlantis]

Gigabyte confirms GeForce RTX 3060 8GB, RTX 3070 16GB and RTX 3080 20G

Gigabyte’s Watch Dogs Legion code redeeming website lists many graphics cards that have not been announced by the manufacturer. In fact, these cards have not even been announced by NVIDIA yet.

The product list features GeForce RTX 3060 (S – SUPER?) 8GB model, GeForce RTX 3070 16GB (S- SUPER?) and Geforce RTX 3080 20GB (also labelled S). It remains unclear if the S means a SUPER series, or as our sources tend to believe, the Ti models.

 

[Jawed]

So, $600 for a 10GB 3080 at the end of October?

 

[trinibwoy]

Another 10GB is going to add non trivial power consumption. The 10GB 3080 is pushing the envelope as is.

 

[DegustatoR]

So, $600 for a 10GB 3080 at the end of October?

More likely to be $900 for 3080 20GB.

 

[Deleted member 13524]

So... AMD will be late to the party by announcing their cards only on October 28th and probably making them available in November. OTOH, looking at the availability numbers it doesn't look like nVidia will be selling a whole lot of RTX 3080 cards in the meantime.

 

[Deleted member 2197]

So... AMD will be late to the party by announcing their cards only on October 28th and probably making them available in November. OTOH, looking at the availability numbers it doesn't look like nVidia will be selling a whole lot of RTX 3080 cards in the meantime.

Sure, for a few days. Next week and beyond is a different matter.

 

[DegustatoR]

So... AMD will be late to the party by announcing their cards only on October 28th and probably making them available in November. OTOH, looking at the availability numbers it doesn't look like nVidia will be selling a whole lot of RTX 3080 cards in the meantime.

Do you expect AMD availability to be significantly better post launch?