NVIDIA’s New RTX 3090, the foremost Powerful GPU Yet?

What we’ve Learned

The Nvidia GeForce RTX 3090 is now confirmed because the next halo graphics card from Team Green and Jensen has spilled the beans (most of them, anyway) on specs and performance. If you liked the simplest performance from Nvidia’s Ampere architecture, prepare to require out a little loan because the king of the GPU hierarchy and therefore, the best graphics card (‘best’ as in ‘fastest’) won’t come cheap. The RTX 3090 sets a replacement high-bar for single-GPU pricing at $1,499, not counting Nvidia’s Titan series that it’s apparently meant to exchange. Here’s everything we all know about the GeForce RTX 3090.

The GeForce RTX 3090, is the first 90-series suffix we have seen from Nvidia since the GTX 690 back in 2012. That was a dual-GPU variant of the GTX 680, back when multi-GPU was a thing. Which, technically it still is, but support has been seriously lacking lately. Regardless, the RTX 3090, is the only Ampere GeForce GPU that has NVLink support this round, just in case you’ve got $3,000 sitting around. (Don’t do it!) But let’s hit the specs.

We’ve covered the high-level view of Nvidia’s new Ampere GPUs elsewhere, and you’ll examine the GeForce RTX 3080 and GeForce RTX 3070 in their dedicated articles. The main target here is on the RTX 3090. After months of speculation and waiting, we finally have the hard details. It’s big, quite literally. Nvidia’s RTX 3090 reference model (we’re unsure if Nvidia remains using Founders Edition branding) sports a triple-slot cooler and features a 350W TDP. You would possibly need a PSU, case, and CPU upgrade to form the foremost of this bad boy.

A Quick Glance

● 24GB GDDR6X at 19.5Gbps

● 10496 CUDA cores and 35.7 TFLOPS of FP32 compute

● Samsung 8N manufacturing process

GPU Specifications

We heard various that Nvidia might double the number of shader cores per SM. What we didn’t expect was doubled the FP32 shaders while still packing 82 SM’s. And Nvidia could have theoretically gone even bigger (the GA100 is an 826 mm square chip, where GA102 is seemingly only around 627 mm square). Still, the resulting 36 TFLOPS of computing is going to be a huge boost to performance … provided the remainder of your PC can continue.

Raw compute power is 150% quite the RTX 2080 Ti, for both the CUDA cores and therefore, the Tensor cores. As in, on paper, the RTX 3090 is 2.5 times as fast because of the previous king. Actually, maybe that’s not fair it should be compared with the Titan RTX, right? Then it’s only 2.2 times as fast, plus it costs $1,000 less.

I’m a touch sad that the GDDR6X memory, ‘only’ clocks in at 19.5Gbps, and I am in need of a GPU hat to eat (chocolate, please!), but we’re still watching 24 GB of memory and 936 GBps of bandwidth. That’s a 52% increase relative to the RTX 2080 Ti, and Nvidia likely has some architectural improvements thereto it makes better effective use of that bandwidth.

The Most Powerful And Expensive

The GeForce RTX 3090 is not just the foremost expensive GeForce card to date; it is the largest graphics card Nvidia has sold. We have seen various third party designs push the bounds of excellent sense (in an honest way, provided you’ve got an outsized PC), but Nvidia has previously limited its designs to 2-slot solutions. No more! The RTX 3090 may be a triple-slot card, measuring 12.3-inches long and 5.4-inches tall.

It’s a monster! And that I like it. Be still, my heart! Forgive techno-lust, but this is often definitely an exciting GPU. Soon, it’ll be here.

The RTX 3090 also sports a 350W TDP (or TGP if you favor, which is the power to the whole GPU, not counting any extra power employed by the things like VirtualLink). To assist deal with the added thermal output, Nvidia has significantly altered the cooling design compared to previous-generation GPUs.

It’s Features

Nvidia’s GeForce RTX 3090 will make use of Micron’s GDDR6X memory. This GPU and therefore, the RTX 3080 are the sole ones slated to use GDDR6X for now, and Nvidia had to further improve the signal delivery to spice up speeds.

HDMI 2.1 makes its debut during a graphics card, but the three DisplayPort connectors remain stuck at 1.4a. Both standards can drive an 8K display, but where HDMI 2.1 can do 8K120 via DSC, DisplayPort 1.4a requires DSC just to urge to 8K60.

Nvidia has also added PCI-E Gen4 support to its Ampere GPU. It’s worth remarking that this probably won’t matter much for gaming performance because of the large 24 GB of VRAM means there should be fewer data going back and forth over the PCI-E bus. The opposite problem in fact is the fastest gaming CPUs still come from Intel, and Intel doesn’t have a desktop PCI-E Gen4 solution yet. Which will accompany next year’s Rocket Lake processors, which can once more use Intel’s 14 nm++(++) process?

That’s a 52% increase relative to the RTX 2080 Ti, and Nvidia likely has some architectural improvements thereto it makes better effective use of that bandwidth.

Scroll to Top

This website or the third party tools used by this website use cookies that are necessary for the purposes set out in the cookie policy. If you would like to know more or deny consent to all or some cookies, please read the cookie policy.