Nvidia Pascal GeForce X80, X80Ti & Titan Specs Leaked; GP104 GPU Supports Only GDDR5 Memory?
With debut of the next-gen Nvidia Pascal GPUs just around the corner, rumor mill is in full swing these days. Recently, a chart has leaked online which allegedly reveals specs of the new GeForce X family which would include GP104 based X80, and GP100 based X80Ti and X80 Titan graphics cards.
As shown in the chart below, the next-gen GeForce X80 Titan will be the most powerful card in the GeForce X lineup based on GP100 GPU which features 6144 CUDA cores, 384 texture mapping units, 192 render output units, and a 1025MHz base clock for a total of 12.595 teraflops.
In comparison, the X80Ti will allegedly feature a cut down GP100 GPU, which packs 5120 CUDA cores, 320 texture mapping units, and 160 render output units. The GPU is clocked at 1025MHz base for a total of 10.496 teraflops.
Next up in line is the GeForce X80, which is based on the most powerful GP104 configuration with 4096 CUDA cores, 256 texture mapping units, 128 render output units, along with a base clock of 1000MHz for a total of 8.192 teraflops.
To our surprise however; the report claims that GP104 would support 6GB of GDDR5 memory, rather than GDDR5X. Considering X80 is a high end graphics card, it seems unlikely that Nvidia will opt for GDDR5 over next gen GDDR5X – which offers twice the bandwidth and capacity.
Moreover, this specific part also contradicts with a previous report from Benchlife, a reputable source for hardware leaks, which states that the upcoming X80 graphics card will be configured with 8GB of memory rather than six.
Both these factors challenge the credibility of this “Nvidia Confidential” table, so we would advise you take it with grain of a salt. Meanwhile, here is what we know about Nvidia Pascal architecture so far.
Nvidia Pascal Specifications:
- Pascal graphics architecture
- 2x performance per watt estimated improvement over Maxwell
- To launch in 2016, purportedly the second half of the year
- DirectX 12 feature level 12_1 or higher
- Successor to the GM200 GPU found in the GTX Titan X and GTX 980 Ti
- Built on the 16nm FinFET manufacturing process from TSMC
- Allegedly has a total of 17 billion transistors, more than twice that of GM200
- Will feature four 4-Hi HBM2 stacks, for a total of 16GB of VRAM and 8-Hi stacks for up to 32GB for the professional compute SKUs
- Features a 4096-bit memory bus interface, same as AMD’s Fiji GPU power the Fury series
- Features NVLink (only compatible with next generation IBM PowerPC server processors)
- Supports half precision FP16 compute at twice the rate of full precision FP32
As for the pricing, the Nvidia Pascal cards are rumored to cost between $600 to $1,100. The chip maker is expected to unveil the new architecture during the GPU Technology Conference (GTC) 2016 which kickstarts in the first week of April.