Benchmarks of the Geforce GTX TITAN-X have surfaced over at Compubench giving us an insight into the OpenCL performance of the GM200 GPU. The flagship GPU leads the charts in OpenCL performance taking its spot at the No.1 aggregate spot. The benchmarks can be found over at the Compubench database (via Videocardz.com).
OpenCL performance of the GM200 with 3072 CUDA Cores does not disappoint
The Geforce GTX TITAN-X will have 3072 CUDA Cores divided into 24 SMMs and a die size of somewhere around 600mm^2. It has adopted a more interesting looking black tone over the silver standard. Tthe actual ‘X’ is not present on the shroud rather, and only the “TITAN” branding is visible. Many enthusiasts who have owned a GeForce GTX Titan GPU before the Titan X have been asking for a Titan Black revision, they did get one but it wasn’t a complete black design as compared to the latest beauty. Anyways, here are the benchmarks in question.
As you can see the Titan manages to land the top spot in every chart - which is expected though still impressive. Keep in mind however that the benchmark only tests cards at vanilla settings or stock clocks so the overclocked cards under the TITAN-X could still theoretically outperform the GPU. OpenCL performance is a very good indicator of the professional worth of a core and this seems to be no exception. Keep in mind though that there doesnt seem to be a clear FP64 test given although one particular extension in the OpenCL info dump seems to suggest it is present. The card was clocked at 1076 Mhz which should be the approximate launching clock of the Geforce GTX TITAN-X.
NVIDIA GeForce GTX Titan X “Initial” Specifications:
NVIDIA GeForce GTX Titan X | NVIDIA GeForce GTX Titan Black | NVIDIA GeForce GTX 980 | NVIDIA GeForce GTX 970 | NVIDIA GeForce GTX 960 | |
GPU Architecture | Maxwell | Kepler | Maxwell | Maxwell | Maxwell |
GPU Name | GM200 | GK110 | GM204 | GM204 | GM206 |
Die Size | ~600mm2 | 561mm2 | 398mm2 | 398mm2 | 228mm2 |
Process | 28nm | 28nm | 28nm | 28nm | 28nm |
CUDA Cores | 3072? | 2880 | 2048 | 1664 | 1024 |
Texture Units | TBA | 240 | 128 | 104 | 64 |
Raster Devices | TBA | 48 | 64 | 64 | 32 |
Clock Speed | 1076 MHz (Preliminary) | 889 MHz | 1126 MHz | 1051 MHz | 1127 MHz |
Boost Clock | TBA | 980 MHz | 1216 MHz | 1178 MHz | 1178 MHz |
VRAM | 12 GB GDDR5 | 6 GB GDDR5 | 4 GB GDDR5 | 4 GB GDDR5 | 2 GB GDDR5 |
Memory Bus | 384-bit | 384-bit | 256-bit | 256-bit | 128-bit |
Memory Clock | 7.0 GHz | 7.0 GHz | 7.0 GHz | 7.0 GHz | 7.0 GHz |
Memory Bandwidth | 336.0 GB/s | 336.0 GB/s | 224.0 GB/s | 224.0 GB/s | 112.0 GB/s |
TDP | 225-250W | 250W | 165W | 145W | 120W |
Power Connectors | 8+6 Pin | 8+6 Pin | Two 6-Pin | Two 6-Pin | One 6-Pin |
Price | $999 US? | $999 US | $549 US | $329 US | $199 US |