So, there’s a new rumor coming from your regular websites. NVIDIA is launching its new GTX 1070 Ti graphics card, the rumor now is that it cannot be overclocked and that all cards will get a 1607/1683 MHz base/boost clock frequency.
This rumor as such is rather hard to believe. We do think that you’ll only see AIB cards as Nvidia will likely not make any reference (Founders) cards. So the rumor, and again, .. rumor .. is that the clock frequency cannot be changed by board partners (factory tweaks) by the AIB partners, e.g. all cards would be clocked the same 1607/1683 MHz base boost clock. If this rumor is true, it might be in place to prevent the 1070 Ti to surpass GTX 1080 performance.
Expreview (Google translation):
Xiao Bian get exclusive news, NVIDIA will be fully unified market GTX 1070 Ti frequency, that is, regardless of which you buy a GTX 1070 Ti their frequency will be fixed at 1607MHz-1683MHz , if you want to be stronger performance only Can be manual overclocking, for non-hardware enthusiasts are not friendly, or graphics card manufacturers have to surprise, such as get a software overclocking to avoid restrictions on NVIDIA.
Earlier on sources confirmed a release end of October. The GPU has 20 Shader clusters, just one would be disabled totalling towards 2432, which is just 128 shader procs lower than that of the GTX 1080. The clock speeds of the GTX 1070 Ti would be higher than the GTX 1070 as well at 1607 MHz and a 1683 MHz GPU Boost woth 8.00 GHz (GDDR5-effective) memory clock. The TDP would be 180W, similar to the GTX 1080. If I perform some reverse math on this, the product would offer just over 8 TFLOPS in perf.
 |
||||||
|---|---|---|---|---|---|---|
| Reference GeForce | Titan X | GTX 1080 | GTX 1070 Ti | GTX 1070 | GTX 1060 | GTX 980 Ti |
| (2016 edition) | ||||||
| GPU | GP102-400-A1 | GP104-400-A1 | GP104-300-A1 | GP104-200-A1 | GP106-400-A1 | GM200 |
| Architecture | Pascal | Pascal | Pascal | Pascal | Pascal | Maxwell |
| Transistor count | 12 Billion | 7.2 Billion | 7.2 Billion | 7.2 Billion | 4.4 Billion | 8 Billion |
| Fabrication Node | TSMC 16 nm | TSMC 16 nm | TSMC 16 nm | TSMC 16 nm | TSMC 16 nm | TSMC 28 nm |
| CUDA Cores | 3,584 | 2,560 | 2,432 | 1,920 | 1,280 | 2,816 |
| SMMs / SMXs | 28 | 20 | 19 | 15 | 10 | 22 |
| ROPs | 96 | 64 | 64 | 64 | 48 | 96 |
| GPU Clock Core | 1,417 MHz | 1,607 MHz | 1,607 MHz | 1,506 MHz | 1,506 MHz | 1,002 MHz |
| GPU Boost clock | 1,531 MHz | 1,733 MHz | 1,683 MHz | 1,683 MHz | 1,709 MHz | 1,076 MHz |
| Memory Clock | 2,500 MHz | 2,500 MHz | 2,000 MHz | 2,000 MHz | 2,000 MHz | 1,753 MHz |
| Memory Size | 12 GB | 8 GB | 8 GB | 8 GB | 3 GB / 6 GB | 6 GB |
| Memory Bus | 384-bit | 256-bit | 256-bit | 256-bit | 192-bit | 384-bit |
| Memory Bandwidth | 480 GB/s | 320 GB/s | 256 GB/s | 256 GB/s | 192 GB/s | 337 GB/s |
| FP Performance | 11.0 TFLOPS | 9.0 TFLOPS | 8.17 TFLOPS | 6.45 TFLOPS | 4.61 TFLOPS | 6.4 TFLOPS |
| GPU Thermal Threshold | 94 Degrees C | 94 Degrees C | 94 Degrees C | 94 Degrees C | 94 Degrees C | 91 Degrees C |
| TDP | 250 Watts | 180 Watts | 180 Watts | 150 Watts | 120 Watts | 250 Watts |
| Launch MSRP ref | $1200 | $499 | $399 | $379/$449 | $249/$299 | $699 |
