Teslac2070 mining

1. NVIDIA Tesla c2070 is not a graphics card like gtx590. The correct name is & quot; Operation GPU & quot;, It already has the function of CPU, so the platform it builds & quot; Can no longer & quot; It needs traditional CPU such as Intel and AMD as CPU (as shown in the figure)

2. Compared with the traditional super graphics workstation (server), its price is much cheaper. As shown in the figure, if we use the platform composed of four nvidiateslac270 (realizing the super graphics platform of four cores + 4gpu + 4x6ggddr5 memory + video memory), it only costs about RMB 100000, If it's a common platform of the same level (a 4-core SNB architecture of the most powerful CPU, two NVIDIA quadro6000 professional graphics cards constitute SLI, 24g memory), the price is much more expensive.

3. Tesla c2070 can also be used as a common graphics card in the pci-e16x slot of a common computer, but it's a design based on a professional graphics card, The game performance is much slower than gtx590, but the graphics rendering ability (autocad3d, 3dsmax, Maya) is more than 10 times faster than gtx590

Size specification 9.75 inch pciex16 Tesla GPU quantity 1 CUDA core quantity 448 CUDA core frequency 1.15 GHz al precision floating point performance (peak value) 515 gflops single precision floating point performance (peak value) 1.03 tflops dedicated memory total capacity *
Tesla c2050
Tesla c2070 3gb gddr5
6GB gddr5 memory frequency 1.5 GHz memory interface 384 bits Memory bandwidth 144 GB / s power consumption
Tesla c2050 238w thermal design power consumption system interface PCI x16 Gen2 cooling solution active fan radiator display support
al link DVI-I
display maximum resolution @ 60Hz 1
2560x1600 software development tools CUDA C / C + + / Fortran, OpenCL and directcompute toolkit
NVIDIA & amp; for visual studio; reg; NVIDIA Parallel Nsight™ * Note: when the ECC function is turned on, part of the dedicated memory will be used for ECC data, so the user's available memory capacity will be reced by 12.5% For example, a total storage capacity of 3 GB will only provide 2.68 GB of usable capacity for users.)

Size specification 9.75 inch pciex16 Tesla GPU quantity 1 CUDA core quantity 448 CUDA core frequency 1.15 GHz al precision floating point performance (peak value) 515 gflops single precision floating point performance (peak value) 1.03 tflops dedicated memory total capacity *
Tesla c2050
Tesla c2070 3gb gddr5
6GB gddr5 memory frequency 1.5 GHz memory interface 384 bits The memory bandwidth is 144 GB / s, the power consumption is 247w, the thermal design power consumption system interface is pciex16 Gen2, the cooling solution is active fan radiator, the software development tools are CUDA C / C + + / Fortran, OpenCL and directcompute toolkit
NVIDIA & amp; for visual studio; amp; reg NVIDIA Parallel Nsight?

Size specification 9.75 inch pciex16 Tesla GPU quantity 1 CUDA core quantity 448 CUDA core frequency 1.15 GHz al precision floating point performance (peak value) 515 gflops single precision floating point performance (peak value) 1.03 tflops dedicated memory total capacity *
Tesla c2050
Tesla c2070 3gb gddr5
6GB gddr5 memory frequency 1.5 GHz memory interface 384 bits Memory bandwidth 144 GB / s power consumption
Tesla c2050 238w thermal design power consumption system interface PCI x16 Gen2 cooling solution active fan radiator display support
al link DVI-I
display maximum resolution @ 60Hz 1
2560x1600 software development tools CUDA C / C + + / Fortran, OpenCL and directcompute toolkit
NVIDIA & amp; for visual studio; reg; NVIDIA Parallel Nsight™ * Note: when the ECC function is turned on, part of the dedicated memory will be used for ECC data, so the user's available memory capacity will be reced by 12.5% For example, a total storage capacity of 3 GB will only provide 2.68 GB of usable capacity for users.)

Difference:

1. The biggest difference is that Tesla is a computing card and has no graphics output function

3. Tesla computing card is suitable for complex computing with large data, such as weather forecast and other large computing data, and its budget performance is even much better than x-core CPU. Its essential function is similar to that of CPU. CPU = controller + arithmetic unit

expand knowledge:

Tesla GPU's 20 series proct family is based on the next generation CUDA architecture code named Fermi, which supports many "essential" features of technology and enterprise computing, including C + + support, ECC memory with extremely high precision and scalability, and double precision performance seven times higher than Tesla 10 series GPU

compared with the latest four core CPU, Tesla c2050 and c2070 can achieve the same supercomputing performance at one tenth of the cost and one twentieth of the power consumption

video card (graphics card), also known as display adapter, is the most basic configuration and one of the most important accessories of computer. As an important part of the computer host, the graphics card is the equipment of digital to analog signal conversion, and it undertakes the task of output display graphics. The graphics card is connected to the main board of the computer, which converts the digital signal of the computer into an analog signal for the display. At the same time, the graphics card still has the image processing ability, which can help the CPU work and improve the overall running speed

for people engaged in professional graphic design, graphics card is very important. Civil and military graphics chip suppliers mainly include amd (ultra micro semiconctor) and NVIDIA (NVIDIA). Today's TOP500 computer, including graphics card computing core. In scientific computing, graphics card is called display accelerator

the two cards only have different cores, but a better core is better, and the optimization will be better