NVIDIA GeForce RTX 3070 Ti performance
The RTX 3070 Ti tested in Geekbench.
The upcoming Ti upgrade for RTX 3070 will be released next week. Today NVIDIA released a press driver that allows reviewers to begin their testing. We have already seen results from Ashes of the Singularity , which unfortunately are not very accurate. Thankfully, other reviewers have put RTX 3070 Ti through other benchmarks, such as Geekbench.
This software offers three GPU tests: Vulkan, CUDA, and OpenCL. The RTX 3070 Ti has been tested with CUDA and OpenCL APIs. The test system was equipped with Intel “Coffee Lake” Core i9-9900K clocked at 5.0 GHz. It was equipped with 16GB of DDR4-3000 memory.
Overall the RTX 3070 Ti is 7% to 13% faster than RTX 3070 non-Ti in these tests. It trades places with RTX 2080 Ti, which was the last-gen flagship.
| NVIDIA GeForce RTX 3070 Ti Performance in Geekbench | ||||
|---|---|---|---|---|
| OpenCL | CUDA | |||
| NVIDIA GeForce RTX 3090 | 205635 | 132% | 238424 | 150% |
| NVIDIA GeForce RTX 3080 | 183452 | 118% | 204325 | 129% |
| AMD Radeon RX 6900 XT | 165115 | 106% | – | – |
| AMD Radeon RX 6800 XT | 156957 | 101% | – | – |
| NVIDIA GeForce RTX 3070 Ti | 155763 | 100% | 158616 | 100% |
| NVIDIA GeForce RTX 2080 Ti | 149053 | 96% | 176497 | 111% |
| NVIDIA GeForce RTX 3070 | 135886 | 87% | 148009 | 93% |
| NVIDIA GeForce RTX 3080 Laptop GPU | 128854 | 83% | 138850 | 88% |
NVIDIA GeForce RTX 3070 Ti in Geekbench, Source: Geekbench
NVIDIA GeForce RTX 3070 Ti features GA104-400 GPU with 6144 CUDA cores. This model is equipped with 8GB of GDDR6X memory, clocked at 19 Gbps. This memory upgrade provides additional 160 GB/s of bandwidth over the existing non-Ti SKU, something that can probably be taken advantage of at higher resolutions. Whether that is the case we should know soon enough. NVIDIA RTX 3070 Ti will launch on June 10th with a starting price of 599 USD. Reviews should go live a day sooner on June 9th.
| NVIDIA GeForce RTX 30 Specifications | |||||
|---|---|---|---|---|---|
| VideoCardz | RTX 3090 | RTX 3080 Ti | RTX 3080 | RTX 3070 Ti | RTX 3070 |
| Picture |  |  |  |  |  |
| GPU | GA102-300 | GA102-225 | GA102-200 | GA104-400 | GA104-300 |
| GPU Clusters | 82 | 80 | 68 | 48 | 46 |
| CUDAs | 10496 | 10240 | 8704 | 6144 | 5888 |
| RTs | 82 | 80 | 68 | 48 | 46 |
| Tensors/TMUs | 328 | 320 | 272 | 192 | 184 |
| ROPs | 112 | 112 | 96 | 96 | 96 |
| Base Clock | 1395 MHz | 1365 MHz | 1440 MHz | 1575MHz | 1500 MHz |
| Boost Clock | 1695 MHz | 1665 MHz | 1710 MHz | 1770 MHz | 1725 MHz |
| Memory | 24 GB G6X | 12 GB G6X | 10 GB G6X | 8 GB G6X | 8 GB G6 |
| Bus | 384-bit | 384-bit | 320-bit | 256-bit | 256-bit |
| Memory Clock | 19.5 Gbps | 19 Gbps | 19 Gbps | 19 Gbps | 14 Gbps |
| Bandwidth | 936 GB/s | 912 GB/s | 760 GB/s | 608 GB/s | 448 GB/s |
| TDP | 350W | 350W | 320W | 290W | 220W |
| MSRP | $1,499 | $1,199 | $699 | $599 | $499 |
| Release Date | Sep 24th, 2020 | June 3rd, 2021 | Sep 17th, 2020 | June 10th, 2021 | Oct 29th, 2020 |
Source: Geekbench (CUDA) , Geekbench (OpenCL) via @Benchleaks