NVIDIA launches fastest GPU in the planet

It’s a bold claim but it has already been proven that NVIDIA’s GeForce GTX 480 featuring the next-gen CUDA architecture, Fermi, is the fastest GPU in the world right now.

The NVIDIA GeFroce GTX 480 which is the first DX11 is at least 1.5 times and up to 3 times faster than their previous version, the GTX 285. It is also gives much better performance than its competition, the Radeon HD 5870. Check out this Unigine Heaven Benchmark test to see how it fared against the Radeon GPUs.


What makes it fast? This huge chunk of a GPU (9.5 inches in length) has a whopping 1536MB / 384-bit GDDR5 memory. This makes it suitable for games requiring complex Tessellation, Ray Tracing, physics processing (PhysX) and GPU computing (Open CL, DirectCompute, CUDA C and CUDA C++). Check out this Supersonic Sled demo.

It can also support 3D Vision Surroudn which allows you to do Full HD 3D gaming at 5760 x 1080 utilizing three LCD monitors. Video demo here.


Running a machine configured with the GTX 480 is not for the faint-hearted though. First you have to shell out about Php25000 to Php26000 for the unit alone. Then you have to consider bumping your power supply to accommodate the almost 300W that it needs to run. With that power requirement, you also have to consider outfitting your right with proper cooling solution. So unless you’re a hardcore gaming enthusiast or somebody who needs a powerful machine to do 3D rendering, animation and all those video production stuff, better stick with the entry to mid-level GPUs.

There’s a lower, but still powerful version 400 series by the way which is the NVIDIA GTX470 that goes for around P16,000 to P17,000. These GeForce GTX 400 GPUs are already available in the market.

For more technology news and gadget reviews, follow us on Facebook, Twitter and Instagram
  • I don’t know anyone who still owns a desktop. Wish they would concentrate on fast cards for laptops that don’t get crazy hot.

  • Jercouzen

    I own a desktop. As do a lot of people like me who require workstation class performance on their personal computers.

    Noticed something on the article though. It’s not the memory that makes this thing fast, it’s the architecture. Nvidia made sure it doesn’t choke on the memory end, but all the magic happens in the CUDA cores.

    And though it’s fast, it’s still fairly selective to tessalation heavy DX11 titles. Mostly because it’s what Nvidia designed it to be. A hell of a lot more triangles, rendered a hell of a lot faster.

    What’s more interesting here is that this card is actually gimped.

    The GTX 480 is limited by its TDP due to manufacturing process constraints. It doesn’t even have all 512 CUDA cores active. Maybe when TSMC moves on to the next silicon process, Nvidia can roll out something like a GTX485 or 495 with all 512 cores operational, higher clock speeds, and all within the same thermal/power envelope.