KitGuru KitGuru.net – Tech News | Hardware News | Hardware Reviews | IOS | Mobile | Gaming | Graphics Cards
Today we're back with another look at Nvidia's RTX 4070 Super, perhaps the most well-received of the three 40-series Super refreshes that launched at the start of the year. Specifically we are reviewing the Inno3D Twin X2 model, a compact graphics card that aims to deliver a ‘no frills' experience, targeting those who want a quiet and cool-running GPU without added bells and whistles to drive up the price.
You could say that the Inno3D RTX 4070 Super Twin X2 is most notable for the features it lacks, rather than what it offers – there's no RGB for instance, no dual-BIOS, no factory overclock or even the ability to raise the power limit beyond stock figures. Instead, it is very much a ‘throwback' graphics card, prioritising low noise and temperatures while retaining a compact footprint. Coming in at £589.99 here in the UK it is one of the cheapest 4070 Supers on offer, so let's see how it performs.
| RTX 4080 | RTX 4070 Ti Super | RTX 4070 Ti | RTX 4070 Super | RTX 4070 | |
| Process | TSMC N4 | TSMC N4 | TSMC N4 | TSMC N4 | TSMC N4 |
| SMs | 76 | 66 | 60 | 56 | 46 |
| CUDA Cores | 9728 | 8448 | 7680 | 7168 | 5888 |
| Tensor Cores | 304 | 264 | 240 | 224 | 184 |
| RT Cores | 76 | 66 | 60 | 56 | 46 |
| Texture Units | 304 | 264 | 240 | 224 | 184 |
| ROPs | 112 | 96 | 80 | 80 | 64 |
| GPU Boost Clock | 2505 MHz | 2610 MHz | 2610 MHz | 2475 MHz | 2475 MHz |
| Memory Data Rate | 22.4 Gbps | 21 Gbps | 21 Gbps | 21 Gbps | 21 Gbps |
| L2 Cache | 65536 KB | 49152 KB | 49152 KB | 36864 KB | 36864 KB |
| Total Video Memory | 16GB GDDR6X | 16GB GDDR6X | 12GB GDDR6X | 12GB GDDR6X | 12GB GDDR6X |
| Memory Interface | 256-bit | 256-bit | 192-bit | 192-bit | 192-bit |
| Memory Bandwidth | 716.8 GB/Sec | 672 GB/Sec | 504 GB/Sec | 504 GB/Sec | 504 GB/Sec |
| TGP | 320W | 285W | 285W | 220W | 200W |
First, a quick spec recap. Just like the RTX 4070 Ti and RTX 4070, the new 4070 Super uses a cut-down AD104 die, measuring 295mm2. The fundamental building blocks are still the same of course, with the RTX 4070 Super offering a total of 56 Streaming Multiprocessors (SMs), each housing 256 CUDA Cores, for a total of 7168. We also find 56 RT cores, 224 Tensor cores, 224 Texture Units, and 80 ROPs.
TSMC's N4 node has Nvidia cranking up the clock speed significantly this generation, with the RTX 4070 Super sporting the same 2475MHz rated boost clock as the original model. Inno3D's Twin X2 model we are reviewing does not come factory overclocked.
The memory configuration also remains the same as both the 4070 and 4070 Ti. That means a relatively narrow 192-bit memory interface, so even with 12GB GDDR6X running at 21Gbps, total memory bandwidth comes in at 504 GB/s, lower than the RTX 3070 Ti. That said, there has been a substantial upgrade to the L2 cache with the Ada architecture, with the RTX 4070 now offering 49.1MB, compared to just 6MB for GA102.
Considering the increased core-count, power draw is naturally a touch higher than the RTX 4070, with the 4070 Super boasting a 220W TGP. Inno3D has not increased this out of the box for the Twin X2, and the power limit cannot be manually increased beyond 220W.
