NVIDIA GeForce RTX 4070 Graphics Card Now Reportedly Features 7680 Cores, 12 GB GDDR6X Memory, 300W TBP, Matches RTX 3090 Ti in 3DMark

Aug 4, 2022 02:01 EDT

A specifications update for the NVIDIA GeForce RTX 4070 graphics card has been posted by leaker Kopite7kimi according to which the gaming product may feature more cores, memory, & graphics performance matching an RTX 3090 Ti.

NVIDIA GeForce RTX 4070 Allegedly Gets Spec Bump: 7680 Cores, 12 GB GDDR6X Memory, 300W TBP, Matching An RTX 3090 Ti

The same leaker had previously posted the same configuration for the RTX 4070 Ti but it looks like the NVIDIA GeForce RTX 4070 will be taking its place after the alleged spec bump. The card now has the same amount of cores and memory albeit at a much lower TGP than what was reported earlier.

Related StoryHassan Mujtaba
NVIDIA Is Giving Away A Special Cyberpunk 2077-Themed GeForce RTX 4090 Graphics Card

NVIDIA GeForce RTX 4070 'Expected' Specifications

The NVIDIA GeForce RTX 4070 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 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 Ti was reported to get a 400W TBP, the RTX 4070 will be getting a 100W lower TBP of 300W. Now, this could be the power limit for custom designs with the Founders Edition roaming around the 255-285 power limit. The graphics card is expected to be based on the PG141-SKU331 PCB design. The graphics card is expected to deliver 11,000 points in 3DMark Time Spy which will easily match a stock RTX 3090 Ti graphics card.

Related StoryNathan Birch
Forza Horizon 5 is Finally Getting DLSS Support, but DLSS 3 isn’t Quite a Lock
3DMark Time Spy Extreme Graphics
Score
0
4000
8000
12000
16000
20000
24000
0
4000
8000
12000
16000
20000
24000
RTX 4090
19k
RTX 4080
15k
RTX 3090 Ti
11.3k
RTX 4070
11k
RX 6950 XT
10.7k
RTX 3090
10.1k
RTX 3080 Ti
9.8k
RTX 3080
8.9k
RTX 3070 Ti
7.3k
RTX Titan
7.1k
RTX 3070
6.7k
RTX 2080Ti
6.6k
RTX 2080
5.1k

As for its feature set, the NVIDIA GeForce RTX 4070  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 $499-549 US graphics card.

NVIDIA GeForce RTX 4070 Series Preliminary Specs:

Graphics Card NameNVIDIA GeForce RTX 4070 TiNVIDIA GeForce RTX 4070NVIDIA GeForce RTX 3070 TiNVIDIA GeForce RTX 3070
GPU NameAD104-400?AD104-400?Ampere GA104-400Ampere GA104-300
Process NodeTSMC 4NTSMC 4NSamsung 8nmSamsung 8nm
Die Size~300mm2~300mm2395.2mm2395.2mm2
TransistorsTBDTBD17.4 Billion17.4 Billion
PCBNVIDIA PG141-SKU331NVIDIA PG141-310 SKU341NVIDIA PG141NVIDIA PG142
CUDA Cores~7680~768061445888
TMUs / ROPsTBD / 160TBD / 160192/ 96184 / 96
Tensor / RT CoresTBD / TBDTBD / TBD192/ 48184 / 46
Base ClockTBDTBD1575 MHz1500 MHz
Boost Clock~2.6 GHz~2.5 GHz1770 MHz1730 MHz
FP32 Compute~40 TFLOPs~38 TFLOPs22 TFLOPs20 TFLOPs
RT TFLOPsTBDTBD42 TFLOPs40 TFLOPs
Tensor-TOPsTBDTBD174 TOPs163 TOPs
Memory Capacity12 GB GDDR6X?12 GB GDDR6X?8 GB GDDR6X8 GB GDDR6
Memory Bus192-bit192-bit256-bit256-bit
Memory Speed21 Gbps21 Gbps19 Gbps14 Gbps
Bandwidth504 GB/s504 GB/s608 Gbps448 Gbps
TBP~400W285W?290W220W
Price (MSRP / FE)$599 US?$499 US?$599 US$499 US
Launch (Availability)2022202210th June 202129th October 2020

Comments

Load Comments