Friends, how many of you dreamed of
getting money, but doing nothing at the same time? Naturally, everyone
would like that, but this does not happen – you say! So, Artur Knows says
that you can get as close as possible to this and earn
significant money on liabilities. To do this, you need to know where
and most importantly, when to invest money. Artur is
just talking about this on his channel. And if you have no money at all
, then it will show sites where you can earn it. I advise you to go, read
the channel and subscribe, the link is in the description. Greetings to all, dear
friends. In this 640 issue, as always, I have
collected the most interesting iron news, but
before I start, I want to express my deep gratitude to
all the sponsors of the channel for their financial support.
And only with your permission,
I will start with fresh information about Ada Lovelace. NVIDIA has answered a
number of popular questions from users about the
RTX 4000 graphics cards that were asked on Reddit. Among other things, the company
disclosed details of how it considered the performance gain of
new products compared to its predecessors, and also
explained some of the nuances. During the presentation of video cards,
NVIDIA stated that the new products are up to 2-4 times faster than their predecessors
in terms of performance. Users were
skeptical about this, because usually the increase from generation
to generation is from 30 to 50%.
NVIDIA on this occasion says
that the RTX 4090 provides a performance increase
through a combination of software and hardware improvements. That is, roughly speaking, we are
told about DLSS 3.0 technology, which allows us
to achieve such results. One of the users also
asked NVIDIA how much faster the new generation video cards
are faster than their predecessors without using DLSS 3. NVIDIA did not provide new data,
but offered to take a look at the chart with the performance of
video cards in Division 2 and Assassin's Creed Valhalla, which
do not support DLSS. As you can see, the younger RTX 4080
in one of the games turned out to be at the level of the previous flagship
3090 Ti, and in one it was only slightly behind. At the same time, the new flagship
RTX 4090 is at least one and a half times faster than its predecessor
without the use of DLSS. They also said that DLSS 3
can also be implemented using old
Ampere video cards, but NVIDIA is unlikely to do this. As follows from the explanations
of NVIDIA representatives, DLSS 3 is a combination of
three technologies – DLSS Frame Generation, DLSS Super Resolution and NVIDIA
Reflex.
DLSS Super Resolution or, more simply,
DLSS 2 and NVIDIA Reflex are already well-known technologies that
work excellently on the current generation. The difference in DLSS 3 is the
Frame Generation component, and it is this that provides the key difference for the new
upscaling algorithm. And since this novelty
requires a hardware optical flow unit , the implementation on
Amperes will be very costly. At the same time,
NVIDIA representatives admitted the possibility of implementing Frame Generation on
tensor cores without using an optical
flow accelerator. But they explained that in this
case the technology is unlikely to show the required
performance and its use would not be
appropriate.
For this reason, the company has
no plans to make DLSS 3 available on
Turing and Ampere architecture cards and their owners will continue to use
DLSS 2. And since DLSS 2 is
part of DLSS 3, legacy technology will continue to be
supported. NVIDIA will not stop training
its AI model and will regularly release updates to the
DLSS 2 libraries. In addition, game developers
will be able to add
support for DLSS 2 separately and separately for DLSS 3 in their future releases, depending on their
desire. The very integration of DLSS 3 into games
is no more difficult than DLSS 2, and NVIDIA
has the necessary plug-ins for this.
In addition, native
support for the new upscaling algorithm will soon be
added to Unity, Unreal Engine and Frostbite Engine. Also of interest
is the demo version of Cyberpunk 2077 DLSS 3 on the RTX 4090 video card
. Here, for the first time
, the clock speed of the GPU is shown on the indicators,
which ranged from 2800 to 2850 MHz. NVIDIA's CEO
hinted that the RTX 4090 was already clocked at 3GHz
during internal overclocking tests, but we've yet to
find out if that's true. DLSS 3 itself not only boosts the
frame rate from 60 to 171 FPS, but also reduces
GPU power by more than 100W. It's definitely cool, but
it's too early to say it's a real breakthrough
since we haven't seen independent tests of the new DLSS yet.
The exact characteristics of
all three chips from the new line also appeared, and the area of the flagship
chip was 608 mm2. In turn, the chips from the RTX
4080 turned out to be more modest both in size and
area. And in order to receive news
even more quickly, do not forget to subscribe to
my telegram or group in VK using the links in the description.
