Nvidia says its new RTX 30 series mobile GPUs are twice as power efficient as previous generations thanks to its new AI-powered Max-Q 3.0 with Dynamic Boost 2.0 technology. Nvidia also confirmed that its implementation of resizable BAR technology, branded Smart Access Memory by graphics rival AMD, is on the way.
In its CES livestream, Nvidia's senior VP of its GeForce graphics division, Jeff Fisher, said Dynamic Boost 2.0 uses AI technology to dynamically shift power between CPU, GPU and GPU memory on a per-frame basis.
The net result? Twice the power efficiency of the outgoing RTX 20 series mobile GPUs.
To put the new RTX 30 series and the achievements of Max-Q 3.0 in context, Fisher drew comparisons with the GTX 10 series.
"Four years ago, laptops powered by GTX 1080 averaged 40mm thick and weighed over nine pounds. After four years of innovation, Max-Q gen 3.0 laptops are half the size, half the weight and up to 10 times the performance," Fisher said.
Fisher didn't go into any detail regarding the character of the claimed 'AI' intervention in Dynamic Boost 2.0 and whether it used pre-baked profiles generated courtesy of large AI networks or whether the AI algorithms are all crunched in real time locally by the GPU.
However, he did confirm Nvidia's support for resizable BAR within the PCI Express interconnect.
"Games use GPU memory for textures, shades and geometry, constantly updating, as the player moves through the world. Today, only part of the GPU’s memory can be accessed at any one time by the CPU, requiring many memory updates. With resizable BAR, the game can access the entire GPU memory, allowing for multiple updates at the same time, improving performance," Fisher explained, conspicuously failing to note that AMD got there first with its Smart Access Memory feature.
He also didn't provide any specifics re the performance boost on offer thanks to resizable BAR support.
Anyway, however you slice it, the new RTX 30 series Mobility GPUs look like a big step forward. Nvidia reckons even the $999 RTX 3060 mobile is faster than the outgoing RTX 2080 Super for laptops. So, the 3070 and 3080 promise to be absolute beasts.