Nvidia CES 2022 Recap: RTX 3050/4th Gen Max-Q/RTX 3090 Ti

Featured Image

Airing live side-by-side just right after AMD’s presentation at CES 2022 was Team Green’s slideshow of its supposed achievements into the gaming scene. Aside from trying really hard to still present itself positively to gamers, Nvidia also announces a few new GPU lineups.

Part 01: Something new for the miners – Geforce RTX 3050 discrete version

Spearheading the GPU announcement (after an entire boring segment about cloud gaming, development partners, smart TVs, new G-SYNC monitors, and some voodoo feature called Reflex) is the actual confirmation of the RTX 3050 discrete GPU version.

Much like the mind-boggling comparison of the RX 6500 XT to the RX 570 and GTX 1650, the RTX 3050 was also compared to the GTX 1050 Ti and 1650. Only this time, the performance tier and price tier as completely out of whack, with performance numbers that you would definitely roll your eyes over (see featured image). Even worse, DLSS benchmarks are thrown directly to the charts, making the comparisons even more nonsensical (“double” the FPS).

At the very least, we did get to confirm a lot of the spec details that were leaked over the last few months. Performance should be similar, if slightly better than the GTX 1660 Ti, but a bonafide RTX card, all fancy technologies related to the platform are supported. The 2560 CUDA core checks out, which is exactly the same as the number of CUDA cores on an RTX 3050 Ti mobile (laptop). The boost clock sits very near to 1800Mhz. TDP is a smidge higher than the GTX 1660 Ti (130 watts). Although, the 4GB version never materialized. The one introduced in the presentation was only the 8GB GDDR6 version, with a 128-bit memory bus.

8GB VRAM?! Yes, unfortunately, these specs kind of still make it an active potential target of crypto miners. If the GTX 1660 Super continues to be mining profitable-ish to this day, you can bet that the supply of these cards would still be significantly strained due to crypto mining’s infinite demand.

If you are interested in this GPU, your best hope is two things. One, that you can line up on the first day to catch the first wave. Two, that you can grab onto one on that release wave, in the hopes of getting it somewhat near its paper MSRP of 250 USD (350 to 400 USD).

Otherwise, be prepared for even more disgusting retail prices than the already detestable GTX 1660 Ti and 1660 Super. (hint: it’s over 500 USD)

Part 02: More ‘air’ in laptops – 4th Gen Max-Q, RTX 3070/3080 Ti goes mobile

Another mildly fascinating announcement related to GPUs (after another boring segment about the Nvidia Studio and its Omniverse feature), is the next shift to 4th Gen Max-Q technologies (Nvidia’s sexy form factor, performance, and cooling balancing act) as well as the unveiling of the new RTX laptops.

The main draw of the new 4th Gen Max-Q tech is smarter control with how or what work is being down with the power draw being required. The simply introduced CPU optimization feature, for instance, allows the unit to communicate with the latest mobile CPUs such as Intel Alder Lake and AMD Zen 3+ at a low-level framework. This is supposedly done to gain even more granular access (and therefore energy-to-performance control) to whatever the CPU is currently doing.

In tandem with the aforementioned feature is Rapid Core Scaling, which smartly allocates more power to the physical cores of the GPU chip, allowing higher clock boosts, while disabling other cores for efficient energy usage. This can be very important in graphical workloads that are more instruction-related, and therefore benefit significantly with few cores having higher clocks.

And of course, what Max-Q generational feature set would it be without battery-related upgrades? 4th Gen’s Battery Boost 2.0 is set to provide “70 percent more battery life”, at the cost of (smartly?) scaling down the performance, hopefully not a 1:1 factor this time around.

For the 3080 and 3070 Ti announcements… no actual specs or tangible performance measurements were actually shown. But, it was specifically promoted for high refresh rate 1440p gaming, so we can probably start speculating from there. In any case, the lowest paper MSRP for the models would seem to be 1500 USD for the 3070 Ti mobile models, and 2500 USD for the 3080 Ti mobile models.

Oh well, we suppose that these new laptops will also add to the new 160+ RTX laptops on the market, which is yes, also proudly specifically announced on this presentation.

Part 03: You’re still buying it anyway: Geforce RTX 3090 Ti

“The RTX 3090 TIE, our next BF GPU.”

Last but not the least, another one of the confirmed rumors by this official announcement was the Geforce RTX 3090 Ti. Yes, folks, that beast of power drainer is finally unmasked.

Aaaaand no new additional information will be unveiled until later this January. All that we know (that was already leaked prior anyway), was that:

  • It will use the same 8nm GA102 die, used by the rest of the high-end RTX 30-series GPUs.
  • VRAM will be 24GB GDDR6X with a memory speed of 21 Gbps.
  • Almost eight percent faster memory clock than the 3090 Ti
  • 40 TFLOPS for raw GPU performance (base RTX 3090 has 36 TFLOPS)
  • 78 TFLOPS for raytracing, 320 TFLOPS for tensor (AI) core tasks (base RTX 3090 has 69 TFLOPS raytracing, 285 TFLOPS for tensor core tasks)
  • All in all, prooooobably 10 percent faster than the RTX 3090 (at presumably much, much higher TDP, and of course, much higher astronomical prices as well)

After this very brief 30-second “big” reveal, there’s an additional section about Nvidia’s newest AI developments, but meh. We’ll skip that entirely.

If you’re interested in all the other boring parts we skipped, you can still watch the entire Nvidia CES 2022 announcement video. (WARNING: no video sections, comments are disabled, dislikes are also disabled)

NVIDIA CES 2022 Special Address
Leave a Reply

Your email address will not be published.

Related Posts