- NVIDIA’s N1X chip has been seen as a result of a gack bench
- Springs shows that integrated GPUs have 6,144 CUDA core
- It is equal to RTX 5070 for pure basic counting, but when it comes to performance, there is a lot more
Remember NVIDIA rumors CPU that caused a lot of echoes about grapes last year? We have apparently seen this user chip in the benchmark leak, where the specific details of spraying are key aspects.
Tom’s hardware reports that the N1X chip, which is based on the arm (such as Qualcomm Snapdragon X CPUS) has been seen as a result of a gack bench, especially for the Open CL (graphics) test, where he scored 46,361.
The score at this point is meaningless. This is an early engineering sample of the N1X (in theory), and yet, if you want to assess the performance of the graphics, the gack bench is far from the artificial benchmark’s first choice.
You can like
As noted, though, it gives us a glimpse of this speculation, which shows that (now add salt) N1X will have 20 cores, apparently divided into a pair of 10 core clusters. It is a processor itself, but we also see integrated GPUs here, which shows 48 streaming multi -processors – which is equivalent to 6,144 Kada cover.
It seems a lot, okay? Well, that is, and in fact, those familiar with NVIDIA’s graphics cards will realize that it is in the balpark for the current General GPU at the middle distance-according to the RTX 5070, in reality, its basic count.
Analysis: Hope to carefully
(Image Credit: Nvidia)
So, is we getting a compact user chip that can go to a budget laptop or handhelds to supply a powerful RTX 5070 to the same frame rate? In one word, no, but the N1X is still shaping the Silicon’s intelligent piece, and the rivals will be sitting and taking notice.
For these reasons that the number of cores visible here on the GPU cannot be easily drawn – this benchmark has no patch on RTX 5070, of course – there are other important factors except the basic basic counting.
It includes clock speed and electricity provided to GPU, which is a very different scenario with such integrated graphics in the desktop PC compared to the full on graphics card. In addition to considering the power envelope, throw the bandwidth boundaries – on system memory in terms of piping tasks, without VRAM on board – and this is a good deal of upshot head winds.
This will not refrain from being a sterling actor for the N1X All -In Chip, but trying to estimate the correct level of graphics performance is not trying to provide it at this stage. (Certainly not from the benchmark here, as already noted).
Tom makes an interesting observation, which is because the leaked spears meet the Nvidia GB10 ‘Superchip’ made for powerful AI performance and has been launched in the era of small AI supercomputer. There is no reason why NVIDIA cannot put another spin of the target of targeting devices on it, which includes mini -PC and laptops, and in fact, being particularly excited about the potential use in game handhelds.
For now, though, this is still high in the rumor area. If the previous speculation is believed, we can see NVIDIA users revealing the CPU later this year, before launching in the early 2026.
