 |
| source L engadget.com |
Back in May, NVIDIA reported an incredibly ground-breaking GPU called the A100. The card was intended for server farm frameworks, notwithstanding, for example, the organization's own DGX A100, as opposed to whatever the normal purchaser would run at home. Today, the organization has reported the DGX Station A100 which, as the name infers, has the structure factor of a work area bound workstation. It accompanies four A100 GPUs — either the 40GB model that the first DGX A100 framework accompanied, or another 80GB rendition.
On the off chance that clients go with the last mentioned, they'll have 320GB of GPU memory to play with. (We're not desirous at all...)
Baca Juga
- Realme TV Stick: A Budget-Friendly Streaming Solution
- Remove Shutdown and Restart Button From Start Menu windows 7/8/10
- Realme C67: Smartphone Terbaru yang Membawa Inovasi ke Genggaman Anda
- Perkembangan Teknologi 5G: Masa Depan Terhubung
- Exploring Exciting New Features in Android 13: A Step-by-Step Tutorial
- Aplikasi Virtual Reality untuk Pendidikan: Masa Depan Pembelajaran Interaktif
- Whatsapp, instagram dan facebook down? Apa penyebabnya?
Like the DGX A100, the DGX Station A100 is intended for the science and business world. The sort of examination and designing outfits that are crunching huge datasets and putting intensely in AI and man-made reasoning. As my partner, Chris Schodt clarified not long ago (watch his explainer, it's genuinely acceptable), you won't game on an A100 card at any point in the near future. It uses engineering called Ampere, notwithstanding, which is presently being utilized in customer cards, for example, the GeForce RTX 3080. So while you can't have an 80GB GPU in your gaming PC, it is conceivable to use a portion of similar advances... if you actually have some genuine money to consume.
As indicated by NVIDIA, the DGX Station A100 offers "server farm execution without a server farm." That implies it connects to a standard divider source and doesn't need a server farm grade cooling framework. Inside the gold-ish case, you'll locate a 64-center AMD CPU with 512GB of memory and a 7.68TB NVME SSD. It's the main workgroup worker that bolsters NVIDIA's Multi-Instance GPU (MIG) innovation, the organization clarified in an official statement. That implies a solitary DGX Station A100 can give 28 separate GPU cases to resemble occupations or different clients to get to. BMW, Lockheed Martin, NTT Docomo, and the Pacific Northwest Laboratory have just received DGX Stations for a portion of their AI and exploration centered work.
The DGX Station A100 doesn't make its server farm kin out of date, however. NVIDIA declared today that the standard DGX A100 will be sold with its new 80GB GPU, multiplying memory ability to 640GB per framework. An update choice will likewise be accessible for clients who have just bought the DGX A100 with four 40GB A100 cards (320GB). NVIDIA hasn't unveiled any evaluation for its new endeavor grade equipment, yet for setting, the first DGX A100 dispatched with a beginning retail cost of $199,000 back in May.
