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

#51
Arkz
Chomiq
They just wiped the floor with that pricing. 3080 for $699, 3070 for $499.

Really surprised they went so low.

3090 $1499.
Yeah but lets see what shops actually sell em at.
Posted on Reply
#52
Raendor
kayjay010101
Haha I feel so bad I sold my 2080 Ti to this dude a month ago for $1300 and he thought he got such a good deal because I included a $150 waterblock and the card was $1350 new.. Now it's worth barely $500 lmao
I’d still buy 2080ti for 300 :D
Posted on Reply
#53
mrthanhnguyen
Any info 3080ti coming? 3090 is the Titan replacement not 2080ti. I dont want to buy 3080 then they release 3080ti a month after that.
Posted on Reply
#54
mouacyk
mrthanhnguyen
Any info 3080ti coming? 3090 is the Titan replacement not 2080ti. I dont want to buy 3080 then they release 3080ti a month after that.
They always do, up to 6 months down the line. May be they call it RTX 3085 this time.
Posted on Reply
#55
Fluffmeister
Very impressive indeed, and all those cores! Core blimey!

Damn the 3070 is tempting at that price, but let's be honest, I'd frickin love a 3090.
Posted on Reply
#56
Raendor
Dimi
RTX 3080 it is for me. Should be a great upgrade from my 1070.
I think they’ll do the same thing as with pascal and release 3080ti with something like 14gb ram. I bought 1080 at launch in 2016 and since then decided never go fo x80 card despite that it was good. Makes more sense to get x70 and upgrade to x80ti fir the price of x80.
Posted on Reply
#57
voodooFX
mouacyk
Any sane air cooling enthusiast can see that NVidia left a dumpster fire for Intel and AMD to deal with, by dumping the GPU VRM heat right into the CPU.
this is the first thing I thought
but maybe the think that everybody is into watercooling nowadays...
Posted on Reply
#58
nguyen
A 2080 Ti at 2000mhz core clock is only about 17.5TFlops of shaders performance and the 3080 is 30TFlops at 1700mhz ?
But yeah the 3080 Ti if there is one will be the one to get, 1000-1200usd with 11GBs of VRAM that will almost match the 3090 (which basically is the Titan).
I bet Nvidia was able to get a sweet deal with Samsung 8N to get the prices so low.
Posted on Reply
#59
Raendor
voodooFX
this is the first thing I thought
but maybe the think that everybody is into watercooling nowadays...
that design would work perfect in some itx cases actually
Posted on Reply
#60
xkm1948
3090 for me. Guess I will just retire the 2080Ti to my work PC.
Posted on Reply
#61
mouacyk
Raendor
I think they’ll do the same thing as with pascal and release 3080ti with something like 14gb ram. I bought 1080 at launch in 2016 and since then decided never go fo x80 card despite that it was good. Makes more sense to get x70 and upgrade to x80ti fir the price of x80.
It doesn't work that way. Memory works in 32-bit incrementals and chips are available in 1GB or 2GB sizes only. The next step up without stepping on 3090 is 352-bit, which is 11GB. Putting 22GB on is also way too close to 3090.
Posted on Reply
#62
R0H1T
Chomiq
They just wiped the floor with that pricing. 3080 for $699, 3070 for $499.

Really surprised they went so low.

3090 $1499.
How, do you know something about RDNA2 performance that we don't? Remember "Ampere being 50% faster than Turing & sipping 50% less power" claims? How'd that turn out o_O

1.9x perf/W last I checked. Nvidia definitely knows more about RDNA2 perf than you or me ~ so just like during the Zen launch you can be sure that, much like Intel, Nvidia is taking AMD seriously because they have to!
Posted on Reply
#63
Nater
He called the 3090 the "new" Titan...but no mention if it'll have Pro CAD support...

Was hoping to avoid getting a gaming card AND still wanting a Quadro card for SolidWorks.
Posted on Reply
#64
Pumper
Finally, some price cuts. Since there won't be this bullshit with "standard" and "founder's edition" at different prices, that means that both the 3070 and 3080 are launching at $100 lower than the 2000 series.

3090 is the new Titan, so that's a $1000 price cut.

Now to wait for proper benchmarks and then RDNA2 just to see if AMD will be able to force nvidia to reduce their prices even more.

The disappointing thing is that 3070 should have been at least a 10GB card and the 3080 12GB.
Posted on Reply
#65
boomheadshot8
Just wait for release and they will tested, I think 3060 is also a good deal if it's around 350$
Posted on Reply
#66
midnightoil
Chomiq
They just wiped the floor with that pricing. 3080 for $699, 3070 for $499.

Really surprised they went so low.

3090 $1499.
They're expecting RDNA2 to be faster and more efficient than Ampere ... whilst hoping the huge, brute force 3090 will retain them the absolute crown. Given how inefficient this is looking as an architecture, and that they're on a slightly shrunk 10nm process (that's what Samsung 8nm is), that will be touch and go too ... depends how big the biggest Navi 2x is.

In light of that, they can't maintain their current pricing model.
Posted on Reply
#67
Franzen4Real
xkm1948
3090 for me. Guess I will just retire the 2080Ti to my work PC.
same, 2080ti going out into the living room for big screen and VR gaming.
Posted on Reply
#68
Calmmo
I'm onboard the 3080 train, but not feeling confident on a 10gb card being enough for the next 2 years.
Posted on Reply
#69
chris.london
The RTX 3090 is only 20% faster than the 3080 at twice the price. The 3080 will likely clock higher too. I think this settles it for me.
Posted on Reply
#70
lynx29
Chomiq
How mature of you.
awww . well i wasn't the only one that thought it was in good fun, sorry, you want a hug?
Posted on Reply
#71
kapone32
boomheadshot8
Just wait for release and they will tested, I think 3060 is also a good deal if it's around 350$
It all remains to be seen.
Posted on Reply
#72
Chomiq
midnightoil
They're expecting RDNA2 to be faster and more efficient than Ampere ...
Based on what? A single RT demo or next gen console games running at 4K30 with checkerboard rendering?
Posted on Reply
#73
Legacy-ZA
boomheadshot8
Just wait for release and they will tested, I think 3060 is also a good deal if it's around 350$
I am hoping for around $300 for the RTX3060. Perhaps we might even see a $175 price tag down the line for the RTX3050. Leaving a lot of room for the Ti's as AMD launches their new hardware.
Posted on Reply
#74
mouacyk
3080 did step up an SKU tier though, from 256-bit (8GB) to 320-bit (10GB). That's a change backwards toward the Fermi refresh tiering. That has to be a sign of something to expect from RDNA2.
Posted on Reply
#75
Vya Domus
I am a bit baffled by that "two floating point operations per clock" thing. Whenever Nvidia/AMD/Intel/etc talk about TFLOPS, what they really mean is fused multiply add and those are already "two floating point operations per clock". So ... what's this then ?

You need either ~8000 shaders or rather FP units (kind of impossible) or ~4 Ghz (also obviously impossible) in order to get 30 TFLOPS. Something's off.
Posted on Reply
Add your own comment