Alleged specifications of NVIDIA’s GeForce RTX 4070 Ti graphics card have been revealed and it may just match the RTX 3090 Ti in performance.
NVIDIA GeForce RTX 4070 Ti Allegedly Features 7680 Cores, 12 GB GDDR6X Memory & RTX 3090 Ti Performance
According to leaker Kopite7kimi, the NVIDIA GeForce RTX 4070 Ti might end up being a beast of a card, offering performance that matches the GeForce RTX 4090 Ti, the existing flagship. The leaker reported on some new specifications and it looks like the 4070 series will be benefiting from much higher power limits than their predecessors.
As I have mentioned before, there is an AD104 SKU with a 400W limit.
a full-fat AD104 with 7680FP32
21Gbps 12G GDDR6X
It can easily match RTX 3090 Ti.
— kopite7kimi (@kopite7kimi) August 1, 2022
NVIDIA GeForce RTX 4070 ‘Expected’ Specifications
The NVIDIA GeForce RTX 4070 Ti is expected to utilize the full-fat AD104 GPU core with 7680 cores or 60 SM units. From previous leaks, we also know that the AD104 GPU will come packed with 48 MB of L2 cache and up to 160 ROPs which is simply insane. That’s a 25% increase in core count and a 12x increase in cache versus the GeForce RTX 3070 Ti which rocks the GA104 GPU core.
The clock speeds are not confirmed yet but considering that the TSMC 4N process is being used, we are expecting clocks between the 2.0-3.0 GHz range. The higher than usual clock speed bump comes from the fact that NVIDIA is making a two-node jump considering the Ampere GPUs with Samsung 8nm node was in reality a 10nm process node with some optimizations. NVIDIA is skipping 7nm and going straight for a 5nm node and not even the vanilla variant but an optimized version of it. With Pascal on the TSMC 16nm node, NVIDIA delivered a huge frequency leap and we can expect a similar jump this time around too.
As for memory specs, the GeForce RTX 4070 Ti is expected to rock 12 GB GDDR6X capacities that are said to be clocked at 21 Gbps speeds across a 192-bit bus interface for 504 GB/s bandwidth. While the RTX 4070 is expected to rock a TBP of 300W the RTX 4070 Ti can end up with a power limit as high as 400W. Now, this could be the power limit for custom designs with the Founders Edition roaming around the 325-350W power limit. The graphics card is expected to be based on the PG141-SKU331 PCB design.
As for its feature set, the NVIDIA GeForce RTX 4070 Ti graphics cards would rock all the modern NV feature sets such as the latest 4th Gen Tensor Cores, 3rd gen RT cores, and the latest NVENC Encoder, and NVCDEC Decoder, and support for the latest APIs. They will pack all the modern RTX features such as DLSS, Reflex, Broadcast, Resizable-BAR, Freestyle, Ansel, Highlights, Shadowplay, and G-SYNC support too. Performance is said to easily match an RTX 3090 Ti which is a $1999 US card (MSRP). So we may be looking at that kind of performance on a $599-$649 US graphics card.
NVIDIA GeForce RTX 4070 Series Preliminary Specs:
|Graphics Card Name||NVIDIA GeForce RTX 4070 Ti||NVIDIA GeForce RTX 4070||NVIDIA GeForce RTX 3070 Ti||NVIDIA GeForce RTX 3070|
|GPU Name||AD104-400?||AD104-300?||Ampere GA104-400||Ampere GA104-300|
|Process Node||TSMC 4N||TSMC 4N||Samsung 8nm||Samsung 8nm|
|Transistors||TBD||TBD||17.4 Billion||17.4 Billion|
|PCB||NVIDIA PG141-SKU331||NVIDIA PG141-310 SKU341||NVIDIA PG141||NVIDIA PG142|
|TMUs / ROPs||TBD / 160||TBD / 144||192/ 96||184 / 96|
|Tensor / RT Cores||TBD / TBD||TBD / TBD||192/ 48||184 / 46|
|Base Clock||TBD||TBD||1575 MHz||1500 MHz|
|Boost Clock||TBD||TBD||1770 MHz||1730 MHz|
|FP32 Compute||~38 TFLOPs||~36 TFLOPs||22 TFLOPs||20 TFLOPs|
|RT TFLOPs||TBD||TBD||42 TFLOPs||40 TFLOPs|
|Tensor-TOPs||TBD||TBD||174 TOPs||163 TOPs|
|Memory Capacity||12 GB GDDR6X?||12 GB GDDR6||8 GB GDDR6X||8 GB GDDR6|
|Memory Speed||21 Gbps||18 Gbps||19 Gbps||14 Gbps|
|Bandwidth||504 GB/s||432 GB/s||608 Gbps||448 Gbps|
|Price (MSRP / FE)||$599 US?||$499 US?||$599 US||$499 US|
|Launch (Availability)||2022||2022||10th June 2021||29th October 2020|
Products mentioned in this post
#NVIDIA #GeForce #RTX #Alleged #Specifications #Leaked #Cores #GDDR6X #Memory #Matching #RTX