Nvidia Debuts GeForce RTX Chip Series With Lower Latency

Nvidia debuted its 28-billion transistor Ampere-based 30 Series graphics chips for PC gamers, ideal for Microsoft and Sony’s next-generation consoles to unveil by the holidays. The 30 Series GeForce RTX chips (available September 17) are comprised of the RTX 3070 ($500), 3080 ($700), and 3090 ($1,500), with second generation RTX (real-time ray tracing graphics). According to chief executive Jensen Huang, there are “hundreds of RTX games” in development, joining “Minecraft,” “Control” and “Wolfenstein: Youngblood,” which already feature RTX.

VentureBeat reports that, “Nvidia said the Ampere-based chips will deliver 3D graphics performance that is two times better in performance than the previous-generation Turing architecture and 1.9 times more power-efficient.”

Huang, who said that “Fortnite” will “turn on” its real-time ray tracing features, demonstrated Nvidia Reflex, which “reduces latency, or interaction delays, by up to 50 percent” as well as Omniverse Machinima, “which enables players to use games to tell stories with real-time animations.” Huang also introduced the free Nvidia Broadcast, which enables more reliable game streaming.

Reflex will help gamers in such scenarios as one in Riot Games’ “Valorant,” where an opponent “traveling at 1,500 pixels per second, is only visible for 180 milliseconds.” A typical gamer has machines and displays with latencies up to 100 milliseconds, and Reflex will shave off 50 of them.

The new GPUs can also “can handle 30 teraflops of shader processing, 58 teraflops of RT core processing, and 238 teraflops of Tensor core processing,” which is roughly “twice the performance of the previous generation, with the ability to handle features like DLSS (deep learning super sampling), which uses AI to fill out details and handle processing more efficiently.”

Nvidia also improved the input-output system, in partnership with Microsoft’s new DirectStorage for Windows application programming interface.

Nvidia and Micron worked together “to create GDDR6X, which Nvidia said is the world’s fastest discrete graphics memory for the RTX 30 Series with speeds up to 1 TB per second memory bandwidth.” The company also worked with Samsung to “come up with the 8N Nvidia custom manufacturing process.”

Elsewhere, VB reports that, when Nvidia focused on real-time ray tracing — the “holy grail” for computer graphics — it did so at the expense of improving traditional performance found in the Pascal architecture. With the new 30 Series chips, Huang stated, “to all my Pascal friends, it’s now safe to upgrade.” With a claim of double the power of equivalent Turing cards and 1.9-time more efficiency, “those are exactly the kinds of gains that PC gamers are waiting for.”

“Ampere is more exciting than the next-gen consoles in part because the PS5 and Xbox still feel somewhat mysterious,” it says. “And in contrast to that, Nvidia shared everything about its cards.” For example, it touted the RTX 3080 as offering 10GB of GDDR6X memory that runs at 19Gbps and uses deep learning DLSS that will run ray-traced games at 4K and 60 fps.

Nvidia Broadcast Brings RTX-Powered AI to Livestreaming Video, VentureBeat, 9/1/20
Nvidia’s GeForce RTX 3090 Is a $1,499 GPU for 8K Gaming, Engadget, 9/1/20
The Nvidia/Arm Deal Could Create the Dominant Ecosystem for the Next Computer Era, VentureBeat, 9/1/20