摘要:
In an example, a system and method for centering in a high-performance interconnect (HPI) are disclosed. When an interconnect is powered up from a dormant state, it may be necessary to “center” the clock signal to ensure that data are read at the correct time. A multi-phase method may be used, in which a first phase comprises a reference voltage sweep to identify an optimal reference voltage. A second phase comprises a phase sweep to identify an optimal phase. A third sweep comprises a two-dimensional “eye” phase, in which a plurality of values within a two-dimensional eye derived from the first two sweeps are tested. In each case, the optimal value is the value that results in the fewest bit error across multiple lanes. In one example, the second and third phases are performed in software, and may include testing a “victim” lane, with adjacent “aggressor” lanes having a complementary bit pattern.
摘要:
In an example, a high-performance interconnect (HPI) is provisioned without a separate stream lane. To provide equivalent functionality, stream lane data are provided within data lines during idle periods. Because one stream lane may be provided per 20 data lanes, elimination of the stream lane saves approximately 5% of area. In a pre-data time, the 20 data lanes may be brought high from midrail to represent one species of data (for example, Intel® in-die interconnect (IDI)), and brought low to represent a second species of data (for example, Intel® on-chip system fabric (IOSF)). To represent additional species of data, such as link control packets (LCPs) for example, lanes can be divided into two or more groups, and a single bit can be encoded into each group. LCP can also be encoded into a post-data time, for example by ceasing flit traffic and manipulating a “VALID” lane from midrail to 0 or 1.
摘要:
A supersequence is sent to another device to indicate a transition from a partial width link state to another active link state. The supersequence is to be sent over one or more lanes of a link and is to include at least a portion of a start of data sequence (SDS) to include a predefined sequence and a byte number value. The byte number value is to indicate a number of bytes measured from a preceding control interval.
摘要:
An apparatus that includes a physical interface for a serial interconnect is provided. The physical interface includes a buffer that is selectable to function as a drift buffer or an elastic buffer by a voltage level on a buffer control line. The physical interface also includes encoding logic that can be enabled or disabled by a voltage level on a logic control line. Further, the physical interface also includes and an ordered set generator that can be enabled or disabled by a voltage level on a communications control line.
摘要:
A periodic control window is embedded in a link layer data stream to be sent over a serial data link, where the control window is configured to provide physical layer information including information for use in initiating state transitions on the data link. The link layer data can be sent during a link transmitting state of the data link and the control window can interrupt the sending of flits. In one aspect, the information includes link width transition data indicating an attempt to change the number of active lanes on the link.
摘要:
Methods and apparatus relating to fast deskew when exiting a low-power partial-width high speed link state are described. In one embodiment, an exit flit on active lanes and/or a wake signal/sequence on idle lanes may be transmitted at a first point in time to cause one or more idle lanes of a link to enter an active state. At a second point in time (following or otherwise subsequent to the first point in time), training sequences are transmitted over the one or more idle lanes of the link. And, the one or more idle lanes are deskewed in response to the training sequences and prior to a third point in time (following or otherwise subsequent to the second point in time). Other embodiments are also disclosed and claimed.
摘要:
A set of training sequences is generated, each training sequence to include a respective training sequence header, and the training sequence header is to be DC-balanced over the set of training sequences. The set of training sequences can be combined with electric ordered sets to form supersequences for use in such tasks as link adaptation, link state transitions, byte lock, deskew, and other tasks.
摘要:
A periodic control window is embedded in a link layer data stream to be sent over a serial data link, where the control window is configured to provide physical layer information including information for use in initiating state transitions on the data link. The link layer data can be sent during a link transmitting state of the data link and the control window can interrupt the sending of flits. In one aspect, the information includes link width transition data indicating an attempt to change the number of active lanes on the link.
摘要:
A serial data link is to be adapted during initialization of the link. Adaptation of the link is to include receiving a pseudorandom binary sequence (PRBS) from a remote agent, analyzing the PRBS to identify characteristics of the data link, and generating metric data describing the characteristics.
摘要:
Systems, methods, and apparatuses including a Physical layer (PHY) block coupled to a Media Access Control layer (MAC) block via a PHY/MAC interface. Each of the PHY and MAC blocks include a plurality of Physical Interface for PCI Express (PIPE) registers. The PHY/MAC interface includes a low pin count PIPE interface comprising a small set of wires coupled between the PHY block and the MAC block. The MAC block is configured to multiplex command, address, and data over the low pin count PIPE interface to access the plurality of PHY PIPE registers, and the PHY block is configured to multiplex command, address, and data over the low pin count PIPE interface to access the plurality of MAC PIPE registers. The PHY block may also be selectively configurable to implement a PIPE architecture to operate in a PIPE mode and a serialization and deserialization (SERDES) architecture to operate in a SERDES mode.