ByAdan Flannigan 2019-07-05 1062
It appears that NVIDIA is not Intel, and does not arbitrarily allow AMD to capture its own market share. It is obviously not the best strategy to put the RTX 2080 in a lower performance than the RX 5700 XT. The NVIDIA decided to develop a new GPU to regain its competitive advantage.
Like the RTX 2070/2060, the RTX 2060 Super uses the TU106 core, but the RTX 2060 has a significantly improved specification: The stream processor was upgraded from 1920 to 2176 and the ROPs from 48 to 64. The GDDR6 memory was increased from 192-bit 6 GB to 256-bit 8 GB and the frequency kept at 14 GHz. According to NVIDIA, performance is improved by at least 15 to 20%. It is no problem to surpass GTX 1080.
Since the number of complete TU106 core stream processors is 2304, the RTX 2070 Super can only use a higher order TU104 core and castrate it accordingly. Compared to the RTX 2070, the number of stream processors has increased from 2034 to 2560. The original 144 texture units were also increased to 160. The memory specifications remain unchanged at 14 Gbit / s and 8 GB. According to NVIDIA, the performance of the RTX 2070 Super can be seen in the GTX 1080 Ti.
As for the RTX 2080 Super, NVIDIA didn’t use the complete TU104 core when release RTX 2080. The RTX 2080 Super is the complete of the TU104, which with 3072 stream processors and 192 texture units.
Since the RTX 2060/2070 has added Super 256 stream processors and the RTX 2080 Super has only added 128 more to further improve performance, NVIDIA has decided to increase the memory frequency from the original 14 Gbps to 15.5 Gbps, increasing the memory bandwidth as well. It reached 496 GB / s.
The parameter specifications of RTX 2060/2070/2080 Super are as follows:
Graphic | RTX 2060 Super | RTX 2060 | RTX2070 Super | RTX 2070 | RTX 2080 Super | RTX 2080 |
Architecture | Turing | Turing | Turing | Turing | Turing | Turing |
GPU | TU106-410 | TU106-200 | TU104-410 | TU106-400 | TU104-450 | TU104-400A |
Number of transistors | 108 billion | 108 billion | 13.6 billion | 108 billion | 136 billion | 136 billion |
Core area | 445 ㎜² | 445 ㎜² | 545 ㎜² | 445 ㎜² | 545 ㎜² | 545 ㎜² |
Manufacturing process | 12 nm | 12 nm | 12 nm | 12 nm | 12 nm | 12 nm |
CUDA | 2176 | 1920 | 2560 | 2304 | 3072 | 2944 |
ROPs | 64 | 48 | 64 | 64 | 64 | 64 |
Tensor Cores | 272 | 240 | 320 | 288 | 384 | 368 |
RT Cores | 34 | 30 | 40 | 36 | 48 | 64 |
VRAM | 8GB | 6GB | 8GB | 8GB | 8GB | 8GB |
Memory Type | GDDR6 | GDDR6 | GDDR6 | GDDR6 | GDDR6 | GDDR6 |
TDP | 175W | 160W | 215W | 175W | 250W | 225W |
Price | $399 | $349 | $499 | $499 | $699 | $699 |
In the above form, the RTX 2060 Super shows the largest increase: When the RTX 2060 is removed, a large vacuum is created between the RTX 2060 Super and the GTX 1660 Ti. Will coexist with the RTX 2060. The RTX2060 Super with 399 US dollars is 50 US dollars more expensive than the RTX 2060.
The RTX 2070/2080 Super is completely identical to the original RTX 2070/2080 after its official launch, i.e. the RTX 2070/2080 is completely withdrawn from the market.
You may also want to read: |
Huami Amazfit Bip 2 VS Xiaomi Mi band 4 |
MIUI 11: List of Xiaomi devices getting MIUI 11 upgrade leaks |
Huawei P40 and P40 Pro are on the way |
Extensive Product Selection● Over 300,000 products ● 20 different categories ● 15 local warehosues ● Multiple top brands | Convenient Payment● Global payment options: Visa, MasterCard, American Express ● PayPal, Western Union and bank transfer are accepted ● Boleto Bancario via Ebanx (for Brazil) | ||
Prompt Shipping● Unregistered air mail ● Registered air mail ● Priority line ● Expedited shipping | Dedicated After-sales Service● 45 day money back guarantee ● 365 day free repair warranty ● 7 day Dead on Arrival guarantee (DOA) |