How the QPI bus defines

Intel's QuickPath Interconnect technology is abbreviated as QPI, which translates to fast channel interconnection. In fact its official name is called the Csi,common System interface public interface, used to achieve direct interconnection between chips, rather than connecting to the North Bridge via FSB, directed at AMD's HT bus. Regardless of speed, bandwidth, the bandwidth of each stitch, power consumption and other specifications are beyond the HT bus.

QPI schematic

1, QPI---quickpath interconnect used in lieu of FSB

2, more bandwidth---data transmission speed is the FSB 1600 of twice times -25.6gb/s

3, more efficient---support more than one system bus connection

Technical characteristics of QPI

Greater bandwidth

Intel's QuickPath Interconnect technology is abbreviated as QPI, which translates to fast channel interconnection. In fact its official name is called the Csi,common System interface public interface, used to achieve direct interconnection between chips, rather than connecting to the North Bridge via FSB, directed at AMD's HT bus. Regardless of speed, bandwidth, the bandwidth of each stitch, power consumption and other specifications are beyond the HT bus.

QPI is a serial high-speed point-to-point connection protocol based on packet transmission, which uses differential signal and special clock to transmit. In terms of latency, QPI is almost the same as FSB, but can improve access bandwidth. A group of QPI has a clock signal with 20 data transmission lines, as well as send (TX) and receiver (RX).

A QPI packet contains 80 bits, requiring two clock cycles or four transmissions to complete the transmission of the entire packet (the QPI clock signal rate is half the transmission rate). Of the 20bit data transmitted each time, 16bit is the real effective data, the remaining four bits are used for cyclic redundancy check, in order to improve the reliability of the system. Since the QPI is bidirectional, it can also receive data transmitted at the other end while being sent, so that the total bandwidth of each QPI bus = The number of transmissions per second (that is, the QPI frequency) x bi-directional for each transmitted valid data (that is, 16bit/8=2byte). So the total bandwidth of the QPI frequency is 4.8gt/s =4.8gt/sx2bytex2=19.2gb/s,qpi frequency is 6.4gt/s total bandwidth =6.4gt/sx2bytex2=25.6gb/s. (bit-bit, byte-Byte, 1byte=8bit)

More efficient

In addition, qpi another bright spot is to support multiple system bus connections, which Intel calls MULTI-FSB. The system bus will be divided into multiple connections, and the frequency is no longer a single fixed, and there is no need to go through the FSB as before. According to the system each subsystem to the data throughput demand, each system bus connection speed also may be different, this characteristic certainly is more flexible than the AMD current HyperTransport bus.

The impact of QPI on AMD and Nvidia

As an industry leader, the progress of the Intel platform is someone who is happy and unhappy. AMD, for example, faces the issue of how to catch up with the speed of Intel's processor innovation, and if AMD cannot keep up with Intel in the future, its market share will surely be getting smaller. Of course, AMD has its perfect graphics technology support, which is what Intel lacks.

If the AMD CPU can integrate its GPU, the market influence is also huge.

Nvidia's situation, Intel's goal is the CPU integration of the GPU, and Nvidia's goal is the GPU consolidated CPU, although Nvidia itself is full of confidence, from the current competitive situation, one relies on graphics technology, chipset-sustaining nvidia, in the face of Intel's crackdown, SLI must be promoted on the Intel platform, which has a significant impact on Nvidia's low-end, low-end graphics market in the face of Intel's and AMD's CPU-integrated GPU solutions.

QPI = Quality Performance index, Quality index.