NVIDIA GeForce RTX 4070 leak confirms 5888 CUDA cores, 12GB & 21Gbps memory

Published: Apr 4th 2023, 08:00 GMT   Comments

NVIDIA RTX 4070 is just around the corner

It appears we might have the very first benchmark of the upcoming RTX 4070 GPU. 

The RTX 4070 non-Ti is the upcoming AD104-based mid-range card. It is replacing the Ampere RTX 3070 model that was released more than two years ago (October ’20) but later updated to LHR (Lite Hash Rate) model in May ’21. It is about time NVIDIA has a successor, a successor that launches after RTX 4070 Ti.

According to a leaked benchmark, the RTX 4070 GPU has been tested on a Core i9-13900K system with an ASRock Z790 Taichi Carrara motherboard. Given how niche this motherboard is, we might assume that this is a reviewer preparing data for April 12th.  The card is confirmed to feature 12GB and 10.5 GHz (21 Gbps) memory. This is the same configuration as the RTX 4070 Ti.

Furthermore, the leak confirms 46 Streaming Multiprocessor count, which translates into 5888 CUDA cores. This is quite literally the last part of the RTX 4070 specs that still had to be confirmed. Now it seems we know everything.

GeForce RTX 4070 Specs, Source: Geekbench

As for the result itself, the card scores 202437 points in Geekbench CUDA test and 177594 in OpenCL. Since these synthetic tests in relatively short and not gaming-related APIs, there is really no point in comparing this data. Furthermore, both tests were performed with the older version of Geekbench (5 not 6). However, if one was still determined to compare this card to RTX 3070, then it would end up with 25-30% higher performance versus last-gen.

The RTX 4070 launches on April 13th at $599 price. The card will cost $200 (25%) less than RTX 4070 Ti while performing around 20% slower in first benchmarks.

NVIDIA RTX X070 CLASS GPUs
VideoCardz.comRTX 4070 TiRTX 4070RTX 3070 TiRTX 3070
Picture
Board-SKUPG141-SKU331PG141-SKU34XPG143-SKU15PG142-SKU10
ArchitectureAda (TSMC 4N)Ada (TSMC 4N)Ampere (SAMSUNG 8N)Ampere (SAMSUNG 8N)
GPUAD104-400AD104-250GA104-400GA104-30X
CUDA Cores
 
7680
 
5888
 
6144
 
5888
Base Clock
 
2310 MHz
 
1920 MHz
 
1575 MHz
 
1500 MHz
Boost Clock
 
2610 MHz
 
2475 MHz
 
1770 MHz
 
1725 MHz
Max FP32 Compute
 
40 TFLOPS
 
29 TFLOPS
 
22 TFLOPS
 
20 TFLOPS
Memory
 
12 GB G6X
 
12 GB G6X
 
8 GB G6X
 
8 GB G6
Memory Bus
 
192-bit
 
192-bit
 
256-bit
 
256-bit
Memory Speed
 
21 Gbps
 
21 Gbps
 
19 Gbps
 
14 Gbps
Memory Bandwidth
 
504 GB/s
 
504 GB/s
 
608 GB/s
 
448 GB/s
TDP
 
285W
 
200W
 
290W
 
220W
InterfacePCIe Gen4 x16PCIe Gen4 x16PCIe Gen4 x16PCIe Gen4 x16
MSRP
 
799 USD
 
599 USD
 
599 USD
 
499 USD
Release DateJanuary 5, 2023April 13, 2023June 10, 2021May 8, 2021

Source: (CUDA) Geekbench, (OpenCL) via @BenchLeaks




Comment Policy
  1. Comments must be written in English and should not exceed 1000 characters.
  2. Comments deemed to be spam or solely promotional in nature will be deleted. Including a link to relevant content is permitted, but comments should be relevant to the post topic. Discussions about politics are not allowed on this website.
  3. Comments and usernames containing language or concepts that could be deemed offensive will be deleted.
  4. Comments complaining about the post subject or its source will be removed.
  5. A failure to comply with these rules will result in a warning and, in extreme cases, a ban. In addition, please note that comments that attack or harass an individual directly will result in a ban without warning.
  6. VideoCardz has never been sponsored by AMD, Intel, or NVIDIA. Users claiming otherwise will be banned.
  7. VideoCardz Moderating Team reserves the right to edit or delete any comments submitted to the site without notice.
  8. If you have any questions about the commenting policy, please let us know through the Contact Page.
Hide Comment Policy
Comments