NVIDIA offers two options for its GeForce Now service: one free, with one-hour playback times, and the other payment option in addition to removing this limit, saying that it enables RTX (Track Ray), because it obviously uses its own GPUs. Now, until now it was unknown what GPUs were, too which processor and how much RAM they would assign users (remember that as a streaming platform, they have powerful servers for it they allocate a portion of their resources for users connected to the service).
4 cores of Intel CC150 and 7.1 GB of RAM
A few days ago we were talking about Intel's mysterious processor, the CC150 of which nothing was known but benches and which was thought to have been the manager of a particular platform. In our case, we are in a hurry since we think this Intel CC150 it will be via GeForce Now so now that's all for sure.
However, the Intel CC150 is processing 8 cores and 16 threads and according to data collected and that we will show you through GeForce Now, at the beginning NVIDIA offers users 4 cores and 8 threads in a two-stage system. This means that for each server, there are two Intel CC150's and each one of them Shared between four users, allocating half the resources of each processor to each user.
You can also see that on the program there are 14,320 MB of RAM available. Since RAM is allocated to each processor in Windows Server 2019 environments (which by scanning are software they use to see which one), with the same amount as previously we have servers with 32 GB of RAM, 16 GB shared by the processor, and with it 7.1 GB per user (the rest will be used by the operating system, obviously).
The GeForce RTX T10-8 is a GPU utility
As for the GPU used in GeForce Now, we're also dealing with a puzzle because it's simply called «GeForce RTX T10-8»And that doesn't tell us much. At first it was thought that GeForce's servers would now use the RTX 2080 Ti but this was eventually ruled out. Now, with filtered benches it is possible to see that this graph has it TU102 GPU, which NVIDIA has not introduced in the market (done internally) to offer directly to GeForce Now servers.
The only thing that can be confirmed with this graph is that it has 8 GB VRAM shared, and based on the functionality we see in the Ashes of the Singularity benchmark, it seems to be better for the game in Full HD resolution than anything else, because of course those FPS when we go up to 4K resolution will be undesirable.
Enough to enjoy GeForce Now?
GeForce Now, unlike Google Stadia, is more thanhardware rental»There is more. It basically allows us to install our games visually on NVIDIA servers so that we can play streaming without having enough hardware.
As we have seen, the platform consists of a server with two Intel C150 processors with 8 cores and 16 threads each, with 16 GB of RAM per processor and a GeForce RTX T10-8 with TU102 GPU and 8 GB of Shared VRAM. Considering that each server is shared among 4 users, we have a Quad-corem processor, 7.1 GB of RAM and an excellent GPU with 8 GB of VRAM shared, which means that with insurance we will have 2 GB.
This is already enough games in full HD resolution, but we can't go much further unless with a paid version of the NVIDIA service you have robust allocation of resources it depends on the game we play, something that's working since we will be using virtual machines with HyperV from Microsoft to NVIDIA servers. And logically it isn't the same to use this service to play Fortnite than The Witcher 3.
