NVIDIA’s upcoming GeForce RTX 4090 and GeForce RTX 4070 graphics cards based on the Ada Lovelace GPU architecture have been further detailed by Kopite7kimi in his latest tweet. These graphics cards will rock some seriously insane specs and performance compared to the current Ampere variants.
NVIDIA Rumored To Equip GeForce RTX 4090 With 24 GB / 21 Gbps Memory at 600W, RTX 4070 With 12 GB / 18 Gbps Memory at 300W
In his tweet, Kopite7kimi focuses on two NVIDIA ‘Ada Lovelace’ graphics cards, the GeForce RTX 4090 and the GeForce RTX 4070. The leaker has already revealed the alleged specs of other cards such as the RTX 4090 Ti and the RTX 4080 but it looks like he was able to dig further info regarding the two cards mentioned above. The main difference is that the GeForce RTX 4090 will be based on an optimized AD102 GPU core while the RTX 4070 will utilize an optimized AD104 GPU core.
RTX 4090, PG137/139-SKU330, AD102-300, 21Gbps 24G GDDR6X, 600W
RTX 4070, PG141-SKU341, AD104-400, 18Gbps 12G GDDR6, 300W
— kopite7kimi (@kopite7kimi) May 10, 2022
NVIDIA Ada Lovelace ‘GeForce RTX 40’ GPU Configurations
|GPU Name||GPCs / TPCs||SMS Per TPC / Total||CUDA Cores||L2 Cache||Memory Bus|
|AD102||12 / 6||2 / 144||18432||96MB||384-bit|
|AD103||7 / 6||2 / 84||10752||64MB||256-bit|
|AD104||5 / 6||2 / 60||7680||48MB||192 bit|
|AD106||3 / 6||2 / 36||4608||32MB||128 bit|
|AD107||3/4||2 / 24||3072||32MB||128 bit|
NVIDIA GeForce RTX 4090 / AD102 GPU But Optimized
The NVIDIA GeForce RTX 4090 will also rock the same AD102 GPU as the ‘Ti’ variant but in a cut-down configuration based on the rumor. the AD102 GPU will make use of 144 SMs, a 71% increase over the existing GA102 GPU and house a massive 18,432 CUDA core count. In addition to the SM counts, the Ada Lovelace GPUs will also feature increased L2 cache sizes. Starting with the AD102 GPU, the flagship would be outfitted with up to 96 MB of L2 cache, an insane 16x increase over the 6 MB L2 cache featured on GA102.
The graphics card will additionally rock 24 GB of GDDR6X memory clocking in at 21 Gbps across a 384-bit bus interface and delivering over 1.0 TB/s bandwidth. from the TGP figures we have seenthe card will have a TBP of 600W at reference and that’s going to go up with the non-reference designs.
As we saw with the GeForce RTX 3090 Ti, there are certain models that go beyond the 450W TBP (reference) and up to 516W (custom). The same is going to be the case with Ada Lovelace-based cards. And while this will require lots of cooling, it looks like AIBs are well and ready. We are likely going to see the use of more hybrid-styled GPU coolers rather than traditional air-coolers once we go above the 600W TBP limit. This will mark a 33% increase in TBP over the RTX 3090 Ti.
NVIDIA GeForce RTX 4070 / AD104 at 300W
Finally, there are the NVIDIA GeForce RTX 4070 class GPUs which will utilize the AD104 GPUs. the AD104 GPUs will feature 60 SM units for a total of 7680 CUDA cores & a total of 48 MB of L2 cache. The RTX 4070 series in specific will feature 12 GB of GDDR6 memory running across a 192-bit bus interface and feature a TGP of 300W. Considering that these are GDDR6 chips, we are likely going to get 16-18 Gbps speeds which should provide a bandwidth of around 400 GB/s. This card is said to utilize the PG141-SKU341 PCB.
Now considering the 300W TDP, the GeForce RTX 3070 Ti already features a reference TDP of 290W so you are looking at a measly 10W increase but at the same time, the performance is said to be on par with an RTX 3080 Ti-class graphics card which is rated at 350W. The 12 GB memory also sounds very appealing since it’s the same amount featured on the RTX 3080 Ti. It will all depend on pricing of this card as the 70 class is one of the most popular amongst the high-end graphics card and for users to get 3080-Ti class performance ($1199 US) at around $500-$600 would be a huge upgrade .
The NVIDIA Ada Lovelace GPU family is expected to bring a generational jump similar to Maxwell to Pascal. It is expected to launch in the second half of 2022 but expects supply and pricing to be similar to current cards despite NVIDIA spending billions of dollars to acquire those good good TSMC 5nm (4N) wafers†
NVIDIA GeForce RTX 40 Series Graphics Card Lineup (Rumored):
|Graphics Card||GPU||PCB Variant||SM Units / Cores||Memory / Bus||Memory Clock / Bandwidth||TGP||Power Connectors|
|NVIDIA GeForce RTX 4090 Ti||AD102-350?||TBD||144/18432?||48GB / 384-bit||24 Gbps / 1.15 TB/s||~900W||2x 16 pin|
|NVIDIA GeForce RTX 4090||AD102-300?||PG137/139 SKU330||>144 / 18432?||24GB / 384-bit||21 Gbps / 1.00 TB/s||~600W||1x 16 pin|
|NVIDIA GeForce RTX 4080||AD103-300?||TBD||>84 / 10752?||16GB / 256-bit||21 Gbps / 672 GB/s||~350W||1x 16 pin|
|NVIDIA GeForce RTX 4070||AD104-400?||PG141-310 SKU341||>60 / 7680||12 GB / 192 bit||18 Gbps / 432 GB/s||~300W||1x 16 pin|