As we get closer and closer to the avalanche of next-generation products from AMD, Nvidiaand Intel, the frequency of leaks is increasing day by day. Of all the hardware coming to market later this year, perhaps none is as hot as NVIDIA’s. RTX 40 Series.

The successor to the current generation Turing GPUs, the RTX 40 series should deliver a noticeable performance boost across the board. So far, we’ve seen plenty of reports, leaks, and rumors about the insane power limits of these cards, as well as most of the lineup’s specs.

Today we add to the long list of rumors for Ada Lovelace and this latest comes from a familiar source. notorious leak, @koptie7kimiwho regularly provided leaks on the initial launch lineup of the RTX 40 series a posted yet another update on Twitter, this time going over the clock speeds of the flagship SKU, along with a few other tidbits.

Keep updating.
RTX 4090, AD102-300-A1, 16384FP32, 384bit 21Gbps 24G GDDR6X, 450W, base 2235 boost 2520 real max >2750;
RTX 4080, AD103-300-A1, 10240FP32, 256bit 21Gbps 16G GDDR6X, 420W,
RTX 4070, AD104-275-Kx(x is a number)-A1, 7168FP32, 160bit 18Gbps 10G GDDR6, 300W.

– kopite7kimi (@kopite7kimi) July 4, 2022

About a week ago, Kopite provided a major update for RTX 4090, RTX4080and RTX 4070. Today’s tweet is sort of a follow-up to this one with most of the information from the previous leak(s) with a few more details added. More importantly, we have a first look at what to expect from the top end in terms of speeds.

Latest RTX 4090 Updates

Starting with the RTX 4090, the leaker mentions that NVIDIA’s next BFGPU would have a base clock speed of 2235MHz with a clock boost of 2520MHz. That’s impressive on its own, but the card takes it a step further and boosts up to 2.8GHz in-game, which represents the actual maximum GPU clock speeds. With some aftermarket variants, we might see that number go up to 3GHz.

Those kind of numbers are the highest we’ve ever seen from NVIDIA. In fact, the base clock speed is around 60% higher than the RTX 3090, while the boost clock is around 50% higher over the RTX 3090. This comparison really puts the improvements into perspective and shows how much AMD, which has generally been the winner when it comes to clock speed, needs to be concerned this time around.

NVIDIA uses TSMCit is N4 (5nm) process node, which is more advanced than the N5 (5nm) wafers that AMD uses for RDNA3. That alone puts the red team at a disadvantage, but efficiency should always be AMD’s bread and butter, especially with Ada Lovelace’s outrageous TGPs. The RTX 4090 should have a 450W TGP, a figure that will be pushed even further with custom models.

Additionally, Kopite reiterates the 4090 specs from its previous leaks. As we already know, the RTX 4090 uses an AD102-300 GPU, but this time the leaker mentioned it as “AD102-300-A1“. This GPU has 16,384 CUDAs Coresa.k.a 128 Feeds processors (SMS) . The full AD102 GPU has 148 SM, which means the 4090 is using a scaled-down version, and NVIDIA is saving the full die for a possible 4090Ti/TITAN card all the way.

Anyway, the 16,384 CUDA Core spec combined with the card’s maximum boost clock (2750 MHz), the RTX 4090 has about 90 TFLOPs computing power in single-precision workloads. It’s just a little less than the 100 TFLOPs figure we heard about a few weeks ago for the TITAN-class Ada Lovelace card, and almost double what the RTX 3090 Ti is capable of. Rest assured, the RTX 4090 will be a powerful beast.

This feeling is further restored by the memory specs of the card as it sports 24 GB of GDDR6X memory running on a 384 bit memory bus, clocked at 21 Gbps. All that comes to give more 1 TB/s memory bandwidth. This is the same as the current BFGPU, the RTX 3090 Ti. Finally, it is important to note that the RTX 4090 will use a single 16 pins (PCIe Gen5) power connector, which is capable of supplying up to 600W power.

Latest RTX 4080 Updates

Moving on to the RTX 4080 and RTX 4070 SKUs, the two graphics cards are said to feature different GPUs. The RTX 4080 uses a scaled down version of the AD103-300 GPUs with 80 broadcast processors activated out of a total of 84, which makes 10,240 CUDA Cores. Again, a 4080 Ti model in the future will likely use the full AD103 die.

Moreover, Kopite tells us to expect 16 GB of GDDR6X memory in the RTX 3080 distributed across a 256 bit wide bus interface, also operating at 21 Gbps like the 4090. However, due to the reduced VRAM capacity and a slower memory bus, the bandwidth is reduced to only 672 GB/s. Previously, the memory speed would have been 18 Gbps according to the same leaker.

As for power, we are considering a 420W TBP, which is now confirmed from its previously uncertain status in Kopite’s latest leak round-up, and is barely below the benchmark power requirement for the RTX 4090. There is no word on clock speeds for the thr 4080. Finally we see the same”A1” being appended to the end of the GPU name for the RTX 4080, meaning it uses the “AD103-300-A1“GPUs.

Latest RTX 4070 Updates

Finally, coming to the RTX 4070, this card would use a scaled down version of the AD104 GPU, called “AD104-275-Kx-A1” (the ‘x’ is a placeholder for a number). We will see the AD104 GPU fully utilized in the release of the RTX 4070 Ti next year. As for this particular GPU, it rocks 56 stream processorsWhere 7,168 CUDA Coresabout 1000 more than the RTX-3070Tiwho has 6,144 cores, so this is a nice generational upgrade.

The RTX 4070 comes with 10 GB of GDDR6 Memory, not GDRR6X – NVIDIA’s special flavor of the G6 developed in tandem with Micron exclusively for RTX cards – which is a downgrade from the current-gen RTX 3070 which features GDDR6X memory.

This 10 GB VRAM operates at an unusual frequency 160 bit memory bus with the speed clocked at 18 Gbpsgiving us 360 GB/s memory bandwidth, lower than the RTX 3070 448 GB/s bandwidth. This is because the RTX 3070 has a 256 bit wide bus interface, despite 2 GB less memory. Additionally, the RTX 4070 is going to feature a 300W TBP (Total Board Power), so expect higher power requirements in third-party variants.

Kopite goes out of its way to remind us that the MSRP of the xx70 class will not be lower than that of the RTX 3070 or 3070 Ti, so a price increase may be in order. Let’s just hope this doesn’t give scalpers the chance to exploit buyers like last time, and that enough stock is available for everyone.

---

That covers all of our bases for the RTX 40 series, or at least the family’s launch lineup. NVIDIA is expected to release one card per month, starting with the lowest option, the RTX 4070 in either September Where October. From there, expect the RTX 4080 next month and the 4090 the following month. The RTX 4060on the other hand, is supposed to throw around CES 2023 in January.

AMD is expected to launch its RDNA3 GPUs in November, which means that the red and green team will compete to win the crown. Whether NVIDIA will be able to defend its throne or AMD will reign supreme in a shake-up the company has likely been contemplating for a long time remains to be seen.