Never short of releases, NVIDIA recently announced the arrival of the new Pascal-powered NVIDIA Quadro P5000 & P6000 at SIGGRAPH 2016. The latest iteration of NVIDIA’s Quadro family and the direct successors of the Maxwell-based Quadro M5000/M6000 cards, the new P5000/P6000 cards are claimed to be fastest professional graphics cards available on the market. Based on the new GP104 and GP102 GPU cores, the P5000 and P6000 enable next level development for VR and visualization content.

Specifications:

Quadro M5000

prev_pny-professional-graphics-cards-quadro-m5000-fr

Quadro M6000

prev_pny-professional-graphics-cards-quadro-m6000-fr

Quadro P5000

prev_pny-professional-graphics-cards-quadro-p5000-fr

Quadro P6000

prev_pny-professional-graphics-cards-quadro-p6000-fr

CUDA Cores 2048 3072 2560 3840
GPU Memory 8 GB 12/24 GB 16 GB 24 GB
Memory Interface 256-bit 384-bit 256-bit 384-bit
Memory Bandwidth 211 GB/s 317 GB/s Up to 288 GB/s Up to 432 GB/s
System Interface PCI Express 3.0 x16 PCI Express 3.0 x16 PCI Express 3.0 x16 PCI Express 3.0 x16
Max Power Consumption 150W 250W 180W 250W
Connectors DP 1.2 (4) + DVI-D DL (1) + Stereo DP 1.2 (4) + DVI-D DL (1) + Stereo DP 1.4 (4) + DVI-D DL (1) + Stereo DP 1.4 (4) + DVI-D DL (1) + Stereo
Max Resolution 4096 x 2160 4096 x 2160 7680 x 4320 7680 x 4320

Quadro P6000 Review

In terms of performance, NVIDIA claims the new flagship P6000 is rated for 12 TFLOPS FP32 (compared to the M6000’s 7.0 TFLOPS) and a clock-speed of ~1560MHz, theoretically giving the card 71%  more shading performance and 37% more ROP than the previous gen M6000, while around a 9% performance improvement over the recently released NVIDIA Titan X.  Though the P6000 is paired with the same 24 GB amount of memory as the “refreshed” 24 GB M6000 released this spring, as most 12 GB M6000 owners probably did not upgrade to this reboot, owners will essentially double up on memory capacity relative to past cards.  Additionally, the P6000 will be the first and only Pascal card with GDDR5X supporting ECC memory.  TDP for the P6000 is at 250W, making it an easy drop-in replacement for systems upgrading from the Maxwell M6000, which also utilizes 250W.  The card is cooled not only by a chamber heatsink, but also a blower fan which helps circulate the air to the openings.

Quadro P5000 Review

The P5000 is the P6000’s slightly smaller, cheaper, and lower power sibling, delivering 8.9 TFLOPS compared to the M5000’s 4.3 TFLOPS.  2560 cuda cores results in a performance boost of ~2.6x over its predecessor, while ROP throughput has increased 65%, giving a larger upgrade for the P5000 over the M5000 than the P6000 over its M6000.  The P5000 card comes with 16 GB of GDDR5X memory, essentially doubling the M5000’s 8 GB memory with a 36% increase in bandwidth.  Though the Pascal Card does require a larger TDP of 180W compared to its predecessor, the difference is not too big for most to still make a drop-in replacement of the M5000.

Features

A notable upgrade for the Pascal cards is the new display controller.  Both cards offer (4) 1.4 DisplayPorts, enabling higher resolutions and refresh rates.  The cards now support up to (4) 5K monitors @60Hz, or even (4) 4K monitors @120Hz, a strong upgrade from the Maxwell Cards’ (4) 4K’s @60Hz.  NVIDIA is specifically targeting the compute problem resolution, scientific visualization, and VR design fields with these cards.  With what they are offering, the P6000 and P5000 certainly fit the bill as the current best cards on the market.

Pricing and Availability

The NVIDIA Quadro P6000 and P5000 will be available in October 2016. MSRP for the P6000 is set at $6,999 and $2,499 for the P5000. For best price and availability, please contact our sales here.