The GeForce 256, an arguably revolutionary chip that NVIDIA developed over two decades ago to achieve what the company had described as offloading from the CPU key tasks for rendering 3D images, making it the very first graphics processing unit ever, is now 25 years old, according to a corporate timeline from NVIDIA that can confirm the GPU was originally launched in August 1999, just months after the company unveiled the RIVA TNT2, the first 32-bit frame buffer. The GPU, which became available for purchase in October that year, counts 32 MB of SDR memory, a GPU clock of 120 MHz, and a VGA connector among its specifications.
Key specs of the GeForce 256 include:
- Process Size: 220 nm
- Transistors: 17 million
- GPU Clock: 120 MHz
- Memory Clock: 143 MHz
- Memory Size/Bandwidth: 32 MB, 1.144 GB/s
- Bus Interface: AGP 4x
- Suggested PSU: 200 W
- Outputs: 1x VGA
A look at one of the original VisionTek models:
From NVIDIA’s “Tonight I’m Gonna Game Like It’s 1999!” article:
…man, those graphics… I was getting only 20 FPS at 640×480, which is why I got a big advance on my allowance from my (admittedly sometimes ‘ok’) parents, so I could buy the NVIDIA GeForce 256, the world’s first “Graphics Processing Unit” (GPU). Maximum PC has been hyping it for months and it doesn’t disappoint!
NVIDIA’s TNT2 gaming accelerator went head to head with the Voodoo 3, but this all-in-one 220nm, 120MHz GPU smashes it, with up to 50% faster performance, Direct3D 7.0 support, and more.