Wednesday, September 2nd 2020

NVIDIA Announces GeForce Ampere RTX 3000 Series Graphics Cards: Over 10000 CUDA Cores

NVIDIA just announced its new generation GeForce "Ampere" graphics card series. The company is taking a top-to-down approach with this generation, much like "Turing," by launching its two top-end products, the GeForce RTX 3090 24 GB, and the GeForce RTX 3080 10 GB graphics cards. Both cards are based on the 8 nm "GA102" silicon. Join us as we live blog the pre-recorded stream by NVIDIA, hosted by CEO Jen-Hsun Huang.

Update 16:04 UTC: Fortnite gets RTX support. NVIDIA demoed an upcoming update to Fortnite that adds DLSS 2.0, ambient occlusion, and ray-traced shadows and reflections. Coming soon.
Update 16:06 UTC: NVIDIA Reflex technology works to reduce e-sports game latency. Without elaborating, NVIDIA spoke of a feature that works to reduce input and display latencies "by up to 50%". The first supported games will be Valorant, Apex Legends, Call of Duty Warzone, Destiny 2 and Fortnite—in September.
Update 16:07 UTC: Announcing NVIDIA G-SYNC eSports Displays—a 360 Hz IPS dual-driver panel that launches through various monitor partners in this fall. The display has a built-in NVIDIA Reflex precision latency analyzer.
Update 16:07 UTC: NVIDIA Broadcast is a brand new app available in September that is a turnkey solution to enhance video and audio streaming taking advantage of the AI capabilities of GeForce RTX. It makes it easy to filter and improve your video, add AI-based backgrounds (static or animated), and builds on RTX Voice to filter out background noise from audio.
Update 16:10 UTC: Ansel evolves into Omniverse Machinima, an asset exchange that helps independent content creators to use game assets to create movies. Think fan-fiction Star Trek episodes using Star Trek Online assets. Beta in October.
Update 16:15 UTC: Updates to the AI tensor cores and RT cores. In addition to more numbers of RT- and tensor cores, the 2nd generation RT cores and 3rd generation tensor cores offer higher IPC. Making ray-tracing have as little performance impact as possible appears to be an engineering goal with Ampere.
Update 16:18 UTC: Ampere 2nd Gen RTX technology. Traditional shaders are up by 270%, raytracing units are 1.7x faster and the tensor cores bring a 2.7x speedup.
Update 16:19 UTC: Here it is! Samsung 8 nm and Micron GDDR6X memory. The announcement of Samsung and 8 nm came out of nowhere, as we were widely expecting TSMC 7 nm. Apparently NVIDIA will use Samsung for its Ampere client-graphics silicon, and TSMC for lower volume A100 professional-level scalar processors.
Update 16:20 UTC: Ampere has almost twice the performance per Watt compared to Turing!
Update 16:21 UTC: Marbles 2nd Gen demo is jaw-dropping! NVIDIA demonstrated it at 1440p 30 Hz, or 4x the workload of first-gen Marbles (720p 30 Hz).
Update 16:23 UTC: Cyberpunk 2077 is playing big on the next generation. NVIDIA is banking extensively on the game to highlight the advantages of Ampere. The 200 GB game could absorb gamers for weeks or months on end.
Update 16:24 UTC: New RTX IO technology accelerates the storage sub-system for gaming. This works in tandem with the new Microsoft DirectStorage technology, which is the Windows API version of the Xbox Velocity Architecture, that's able to directly pull resources from disk into the GPU. It requires for game engines to support the technology. The tech promises a 100x throughput increase, and significant reductions in CPU utilization. It's timely as PCIe gen 4 SSDs are on the anvil.

Update 16:26 UTC: Here it is, the GeForce RTX 3080, 10 GB GDDR6X, running at 19 Gbps, 238 tensor TFLOPs, 58 RT TFLOPs, 18 power phases.
Update 16:29 UTC: Airflow design. 90 W more cooling performance than Turing FE cooler.
Update 16:30 UTC: Performance leap, $700. 2x as fast as RTX 2080, available September 17. Up to 2x faster than the original RTX 2070.
Update 17:05 UTC: GDDR6X was purpose-developed by NVIDIA and Micron Technology, which could be an exclusive vendor of these chips to NVIDIA. These chips use the new PAM4 encoding scheme to significantly increase data-rates over GDDR6. On the RTX 3090, the chips tick at 19.5 Gbps (data rates), with memory bandwidths approaching 940 GB/s.
Update 16:31 UTC: RTX 3070, $500, faster than RTX 2080 Ti, 60% faster than RTX 2070, available in October. 20 shader TFLOPs, 40 RT TFLOPs, 163 tensor cores, 8 GB GDDR6
Update 16:33 UTC: Call of Duty: Black Ops Cold War is RTX-on.

Update 16:35 UTC: RTX 3090 is the new TITAN. Twice as fast as RTX 2080 Ti, 24 GB GDDR6X. The Giant Ampere. A BFGPU, $1500 available from September 24. It is designed to power 60 fps at 8K resolution, up to 50% faster than Titan RTX.

Update 16:43 UTC: Wow, I want one. On paper, the RTX 3090 is the kind of card I want to upgrade my monitor for. Not sure if a GPU ever had that impact.
Update 16:59 UTC: Insane CUDA core counts, 2-3x increase generation-over-generation. You won't believe these.
Update 17:01 UTC: GeForce RTX 3090 in the details. Over Ten Thousand CUDA cores!
Update 17:02 UTC: GeForce RTX 3080 details. More insane specs.

Update 17:03 UTC: The GeForce RTX 3070 has more CUDA cores than a TITAN RTX. And it's $500. Really wish these cards came out in March. 2020 would've been a lot better.
Here's a list of the top 10 Ampere features.

Update 19:22 UTC: For a limited time, gamers who purchase a new GeForce RTX 30 Series GPU or system will receive a PC digital download of Watch Dogs: Legion and a one-year subscription to the NVIDIA GeForce NOW cloud gaming service.

Update 19:47 UTC: All Turing cards support HDMI 2.1. The increased bandwidth provided by HDMI 2.1 allows, for the first time, a single cable connection to 8K HDR TVs for ultra-high-resolution gaming. Also supported is AV1 video decode.

Update 20:06 UTC: Added the complete NVIDIA presentation slide deck at the end of this post.

Update Sep 2nd: We received following info from NVIDIA regarding international pricing:
  • UK: RTX 3070: GBP 469, RTX 3080: GBP 649, RTX 3090: GBP 1399
  • Europe: RTX 3070: EUR 499, RTX 3080: EUR 699, RTX 3090: EUR 1499 (this might vary a bit depending on local VAT)
  • Australia: RTX 3070: AUD 809, RTX 3080: AUD 1139, RTX 3090: AUD 2429
Add your own comment

502 Comments on NVIDIA Announces GeForce Ampere RTX 3000 Series Graphics Cards: Over 10000 CUDA Cores

#201
efikkan
medi01
How NV/AMD "know" what is being cooked in the other kitchen always was a mystery to me (and I wondered how much of it could be blocked on driver level).
The fact is they don't know, not until it's too late anyway.
Nvidia themselves don't even know the precise performance until the final stepping is ready, and that's usually 1-2 months before launch.
medi01
Nevertheless, 3080 and below pricing is... not usual, is it?
2080/3080 and 2070/3070 pricing seems the same to me.
Posted on Reply
#202
Zubasa
medi01
On one hand there is that 2+ times CU cores, on other, lower clocks and lower number of transistors per CU (???).

If increasing number of CUs 2.5 times only brings 1.5 times better perf, I'd call it disappointing.


ASP price bump started when AMD Raja times, which is Pascal.
And before that nVidia started shifting the x80 class GPUs down the stack and went from original based on the biggest x100 dies to now the x104-class dies.
As for the performance per CU, you never get 100% scaling off increasing the CU count given it is not practical to increase the cache and bandwidth etc by 2.5x as well.
Also you might lose some clock speed to keep a monster GPU within a certain power envelope.
Posted on Reply
#203
Gameslove
Geforce RTX 3070 is amazing, so expected, plus RTX IO.
Posted on Reply
#204
randompeep
RedelZaVedno
Just look at Microsoft FS 2020. 2080TI manages only 31FPS (with deeps to 21) over NY City at 4K/ULTRA and wants to use 12.7GB of VRAM. 3080 will hopefully get us to 55 fps (with deeps to 40ies). Having 16GB would probably smoothen these fps deeps further. Needless to say 3080 is still a godsent for flight simmers :)
Yeah, but newer architecture means more efficiency in texture processing. I'm a fanboy of my wallet and honestly Nvidia always killed it on this side of things. Just looking at RTX vs RX 5xxx series VRAM consumption tells you thetruth!
MS FS is a weird flex much needed for the hardware industry just to justify upgrading to the curent gen!! It's like a football match, whoever wins doesn't matter anyways - both parties are gonna be paid well enough (AMD & Nvidia). But the gaming market have been rising anyways since COVID 'era' started, so I wouldn't be surprised if AMD jumps on a smaller semiconductor tech as soon as next year tho!
Posted on Reply
#205
Icon Charlie
medi01
How NV/AMD "know" what is being cooked in the other kitchen always was a mystery to me (and I wondered how much of it could be blocked on driver level).

Nevertheless, 3080 and below pricing is... not usual, is it?
Going back to MSRP a few years.

The Standard 1080 was $599 in 2016
The Standard 2080 was $699 in 2018

What I think is that Nvidia is really afraid of NAVI and has priced accordingly.
Posted on Reply
#206
Xaled
jesdals
But does RTX 3090 support SLI
Why does it matter?

I don't believe the prices of 3070 and 3080s
Nvidia is just lying about launch prices and no one would be able to buy these cards for these prices.
Posted on Reply
#207
randompeep
PowerPC
With the VRAM discussion, you also have to mention that the 30x0 cards all have 60% faster VRAM than the previous generation (it makes the process of freeing up and writing to Memory much faster) and therefore will probably lead to a much smaller VRAM bottleneck than Turing had if you even encounter this as a problem, which most games so far don't even on 4K. 8-10 GB is still pretty good. But they'll probably release Ti variants of all these cards with increased VRAM, but also a much higher prices.
Hell yeah. Like someone make the crowd chill!
Can you guys imagine (it's unrealistically, but still) if they still used to make 100-150$ GPUs with those memories ? RTX 3010 & RTX 3030 lol. Probably they would kill the APU game of AMD. But I guess there is a deal behind the scenes.
'As long as you guys don't touch our area, we're leaving yours'
Posted on Reply
#208
medi01
efikkan
2080/3080 and 2070/3070 pricing seems the same to me.
Icon Charlie
The Standard 2080
Yeah, "the standard".

Major perf leap(mkay, everyone understands that claimed figures are well over real, but still a major leap) coming together with price modesty, hm.
Posted on Reply
#209
Xaled
Yeah the price of the never existing standard edition yeah
Posted on Reply
#210
Zubasa
Xaled
I don't believe the prices of 3070 and 3080s
Nvidia is just lying about launch prices and no one would be able to buy these cards for these prices.
An interesting tibit about pricing.
Jensen did finally called the 2080ti a $1200 GPU in this presentation.
I take it as he finally admits that the $999 MSRP was BS right from the get go.
EVGA sold a couple of cards at that price in the first week or so and it never came back in-stock until that SKU was replaced by something more expensive.
Posted on Reply
#211
Icon Charlie
Xaled
Why does it matter?

I don't believe the prices of 3070 and 3080s
Nvidia is just lying about launch prices and no one would be able to buy these cards for these prices.

I don't believe
Oh I am expecting the price to jack up because of Gerbal mentality to buy shiny new things.

It happened during the 2080 series of launches.
It will happen with this launch as well.
Posted on Reply
#212
DuxCro
Why the F only 8GB and 10GB of VRAM?
Posted on Reply
#213
Hellfire
Not sure if mentioned ('ve not read ANY of the previous posts) but surely the internal through fan, for anyone with air coolers on their CPU's could cause problems with their CPU coolers.

I see a lot of hot air being blown onto the CPU. I know a lot use AIO's or watercooling but highend air coolers are still used by many.

DuxCro
Why the F only 8GB and 10GB of VRAM?
I imagine the cost of GDDR6X would play a big factor, driving the price of the GPU up some? I'd imagine 12-16Gb cards to come eventually, maybe Ti's?
Posted on Reply
#214
Xaled
Icon Charlie
Oh I am expecting the price to jack up because of Gerbal mentality to buy shiny new things.

It happened during the 2080 series of launches.
It will happen with this launch as well.
No, this is different from launch price thing as never even got close except 20 cards from evga in Black Friday or something
Zubasa
An interesting tibit about pricing.
Jensen did finally called the 2080ti a $1200 GPU in this presentation.
I take it as he finally admits that the $999 MSRP was BS right from the get go.
EVGA sold a couple of cards at that price in the first week or so and it never came back in-stock until that SKU was replaced by something more expensive.
Exactly, And I believe they did it once upon a year just to get publicity and most importantly if they get sued for such fraud they have a proof that they sold it at that price.
Posted on Reply
#215
randompeep
Zubasa
An interesting tibit about pricing.
Jensen did finally called the 2080ti a $1200 GPU in this presentation.
I take it as he finally admits that the $999 MSRP was BS right from the get go.
EVGA sold a couple of cards at that price in the first week or so and it never came back in-stock until that SKU was replaced by something more expensive.
thought EVGA was caught in a black hole even since their 2012ish wave of expensive cards haha
they were caught off guard when RGB wave started and I couldn't name a goofier modern GPU AIB. Apparently, their cards easily go out of stock, so their aviability started to suck big time in the last 2-3 years
Posted on Reply
#216
Quicks
WOW people get so excited about getting screwed over the previous generation, this is the trend to catch suckers. Always buy the best you can afford every 3rd generation. 1080TI is still a good card to have now only the 30XX will be a worthy successor. The 40XX will suck again and be overpriced, not that I am stating the 30XX is justified in pricing.
Posted on Reply
#217
ppn
DuxCro
Why the F only 8GB and 10GB of VRAM?
New memory type. Supply must be satisfied.
5888 Cuda cards with 8GB is like 2944 with 4GB, so we wait the SUPERS 16/20GB versions. Better wait for 5nm anyway with another 2x density, and 1/2 power.
Posted on Reply
#218
Fluffmeister
Xaled
Why does it matter?

I don't believe the prices of 3070 and 3080s
Nvidia is just lying about launch prices and no one would be able to buy these cards for these prices.
Bait and switch like Vega's "launch only" prices? o_O
Posted on Reply
#219
DuxCro
ppn
New memory type. Supply must be satisfied.
5888 Cuda cards with 8GB is like 2944 with 4GB, so we wait the SUPERS 16/20GB versions. Better wait for 5nm anyway with another 2x density, and 1/2 power.
RTX 3070 will use GDDR6, not GDDR6X
Posted on Reply
#221
Zubasa
randompeep
thought EVGA was caught in a black hole even since their 2012ish wave of expensive cards haha
they were caught off guard when RGB wave started and I couldn't name a goofier modern GPU AIB. Apparently, their cards easily go out of stock, so their aviability started to suck big time in the last 2-3 years
Maybe, or as I suspected and Xaled said above. It might have been a PR stunt and potentially to get the lawyers off their asses.
Posted on Reply
#222
DuxCro
Not gonna buy a founders edition. Would love to get Asus strix or Gigabyte Aorus and then OC until it starts to squeal.
Posted on Reply
#223
dicktracy
Consoles went from midrange specs to lowend now. No need to use overly gimped checkerboard ray tracing with RDNA2 either!
Posted on Reply
#224
randompeep
Zubasa
Maybe, or as I suspected and Xaled said above. It might have been a PR stunt and potentially to get the lawyers off their asses.
I don't know how are EVGA prices outside of Europe, but out there I can find similarities with the Vertu phones.
Not really tryna be a clown, I do respect Asus, Gigabyte, MSI (maybe the best), even Palit and others who tryna be competitve. No wonder why they've always been 'green'
Posted on Reply
#225
ppn
dicktracy
Consoles went from midrange specs to lowend now. No need to use overly gimped checkerboard ray tracing with RDNA2 either!
most definitely 3060 with 4608 Cuda if everything doubled and moved one tier above is making the unreleased consoles look like RTX 3050.
2080 x2 = 3070, 2080Ti x2=3080. except the memory bandwith, that remains the same and probably bottlenecking.
Posted on Reply
Add your own comment