Home






Frequently Asked Questions

1. What is the HyperTransport Technology Consortium?
2. What are the benefits of joining the Consortium?
3. What does it cost to join the Consortium?
4. When and why was the Consortium founded?

HyperTransport Technology FAQs

5. What is HyperTransport Technology?
6. What are the key values of HyperTransport technology?
7. How does HyperTransport technology compare to other bus technologies?
8. What is the latest version of the HyperTransport specification?
9. Why is HyperTransport technology inexpensive to implement?
10. What are HyperTransport hosts, caves, tunnels and bridges?
11. At what clock speeds does HyperTransport technology operate?
12. What is the width of the HyperTransport I/O link bus?
13. With what buses and I/O technologies is HyperTransport technology compatible?
14. What is the HTX connector standard?
15. Is HyperTrasnport technology plug-n-play compatible?
16. Does Hyper Trasnport technology support hot plugging?
17. How can I get HyperTransport technology training?

HyperTransport-Enabled Products FAQs

18. What HyperTransport technology-enabled products are available?
19. What applications use HyperTransport technology?

 

Answers to Frequently Asked Questions

1. What is the HyperTransport Technology Consortium?

 The HyperTransport Technology Consortium (HTC) was founded to provide an open, standardized, high-performance interconnect technology with the industry’s lowest latency.  Formed in 2001 by leading technology innovators as Advanced Micro Devices (AMD), Alliance Semiconductors, Apple, Broadcom, Cisco, NVIDIA, PMC-Sierra and Sun Microsystems, the Consortium is a membership-based, non-profit organization responsible for licensing, managing and promoting HyperTransport technology industry- wide.

Membership is open to commercial organizations and educational institutions. Membership classes include Promoter, Contributor, Adopter and Academic (for non-commercial, academic institutions only). For more information on membership classes and benefits please refer to our membership benefits information page.

> Back to Top


2. What are the benefits of joining the Consortium?

Membership in the HyperTransport Consortium provides royalty-free commercial license to HyperTransport technology, trademarks and access to the HyperTransport Consortium’s patent pool. Other important benefits for Promoter and Contributor members include advanced visibility on and voting rights on HyperTransport technology developments and enhancements, as well as extensive business promotion and industry exposure through the Consortium’s marketing industry and media programs. For more information on membership classes and benefits please refer to our membership benefits information page.

> Back to Top


3. What does it cost to join the Consortium?

Each membership class carries distinct benefits and annual membership fees. For details on membership benefits and membership fees, please refer to our membership benefits information page.

> Back to Top


4. When and why was the Consortium founded?

In early 2001, a group of leading industry companies including Advanced Micro Devices, Alliance Semiconductor, Apple Computer, Broadcom Corporation, Cisco Systems, Nvidia, PMC-Sierra, Sun Microsystems and Transmeta, joined efforts to form the HyperTransport Consortium, with the mission to promote and proliferate the adoption of HyperTransport technology – a technology originally developed by AMD as a way to reduce latency and increase bandwidth in chip-to-chip interconnect design.

> Back to Top

 

HyperTransport Technology FAQs

5. What is HyperTransport Technology?

HyperTransport is a fully scalable, high-performance, low latency, low cost chip-to-chip, board-to-board and chassis-to-chassis interconnect technology, ideally suited for high-performance appliances, embedded systems, networking systems, PCs, workstations, gaming systems, servers, and supercomputers. HyperTransport is the only CPU-native interconnect in the industry.  The technology  enables efficient, high throughput with minimized latency for direct processor-to-processor, processor-to-I/O and processor-to-subsystem (networking adapters, co-processor and other high-performance peripheral boards via HTX connector standard). Since its market launch in 2001, HyperTransport has enjoyed widespread industry adoption and addresses the widest range of applications, ranging from video game consoles - like Microsoft’s Xbox - all the way to the world’s top performing server clusters and supercomputers.

> Back to Top


6. What are the key values of HyperTransport technology?

Key values of HyperTransport technology include:

  • Industry’s lowest latency and highest bandwidth for maximum performance
  • Royalty-free technology for lowest cost
  • Application flexibility as processor-to-processor, processor-to-I/O, processor-to-peripheral, board-to-board and/or chassis-to-chassis interconnect technology
  • Excellent scalability
  • Low power consumption
  • Transparent extensibility to other interconnect standards like PCI, PCI-X and PCI Express
  • Backward compatibility across all HyperTransport specification revision

For more information on the HyperTransport HTX connectivity standard, please refer to the HTX page.

> Back to Top


7. How does HyperTransport technology compare to other bus technologies?

As compared to older address/data/control multi-drop, shared buses such as PCI, PCI-X or SysAD, HyperTransport offers a flexible, packet-based, daisy-chain architecture that provides vast amount of bandwidth combined with the lowest achievable latency.. As compared to new-generation, I/O technologies such as PCI Express, HyperTransport is processor native – i.e. embedded in the processor chip and therefore not burdened by the latency penalties imposed by the extra processor-to-I/O interface control logic that PCI Express requires. HyperTransport is native on several processor families marketed by a diverse breed of processor manufacturers, including AMD, Bay Microsystems, Broadcom, NetLogic Microsystems, Nvidia, Raza Microelectronics, PMC-Sierra, Tarari and targeting a wide range of applications including PCs, servers, networking, communication, media processing and embedded designs. 

> Back to Top


8. What is the latest version of the HyperTransport specification?

The latest HyperTransport technology specification is HyperTransport 3.0, which was released on April 24, 2006. HyperTransport 3.0 boosts the maximum clock speed of HyperTransport links to 2.6 GHz, which thanks to HyperTransport’s double data rate (DDR) capability delivers up to 5.2 GigaTransfers/second per bit and yields the following aggregate performance (2 unidirectional HyperTransport links):
8-bit Links          16-Bit Links             32-bit Links
10.4 GB/s              20.8 GB/s              41.6 GB/s
In addition to increasing maximum performance by 86 percent compared to HyperTransport 2.0, HyperTransport 3.0 adds powerful architectural and operational features, such as:

  • AC mode of operation (capacitive coupling) for longer haul interconnects (full clock speed up to 1m), typical of backplane and chassis-to-chassis applications
  • Enhanced DC operational mode
  • DC/AC auto-sensing and auto-configuration
  • Link Splitting (converting a 16-bit link into two independent 8-bit links)
  • Hot plugging
  • Dynamic power management (real time, on-the-fly, system-transparent link clock/width optimization)
  • 100 percent backward compatibility with previous HyperTransport specifications (HyperTransport 1.x, 2.0 and 3.0 devices can coexist and interoperate within the same HyperTransport architecture due to system-transparent, auto-sensing/selection of common denominator specification setting at product power-up) .

For more information about HyperTransport 3.0 and previous specification releases, please refer to our press releases.

> Back to Top


9. Why is HyperTransport technology inexpensive to implement?

There are several factors that make implementing HyperTransport less expensive as a business investment than other interconnects.

  • First: HyperTransport is a royalty-free technology. This means that each system shipped using HyperTransport technology is free from any per-product royalty dues – a typically burdensome financial exposure of typical technology licensing schemes.
  • Second: The other factors have to do with the structure of HyperTransport’s electrical medium. Specifically, enhanced 1.2 volt LVD signaling minimizes signal noise; the use of much fewer, non-multiplexed interconnect lines reduces signal activity and dual-data rate data transfers lower clock rates while increasing data throughput. All of these factors contribute to minimizing PCB size and complexity as well as overall manufacturing costs.
  • Third: Finally and most importantly, all HyperTransport specifications are backward-compatible with previous specification generations; therefore any investment in HyperTransport technology is preserved over time.

> Back to Top


10. What are HyperTransport hosts, caves, tunnels and bridges?

HyperTransport links interconnect HyperTransport devices in a daisy-chain topology. Each HyperTransport link must have both a host that acts as a HyperTransport link(s) master and an endpoint device, or cave. Additional HyperTransport devices need to be of the tunnel type and they are inserted in the daisy-chain link between the host and the cave. If a HyperTransport link needs to be connected to another type of I/O interconnect standard, a bridge device is required to… bridge the two interconnects, as in the case of a bridge controller interfacing HyperTransport to PCI Express. Tunnels and bridges can act as link endpoints or cave devices also. Each HyperTransport link can support up to 32 HyperTransport controllers (inclusive of host, cave, tunnels and bridges). For more information on host, cave, tunnel and bridge functions, please refer to the HyperTransport white papers and the HyperTransport Bridge Functions overview.

> Back to Top


11. At what clock speeds does HyperTransport technology operate?

HyperTransport technology devices are designed to operate at clock speeds from 200MHz up to 2.6 GHz in increments of 200 MHz, and utilize double data rate (DDR) technology transferring two bits of data per clock cycle, for an effective transfer rate of up to 5.2 Gigatransfers/sec in each direction. Since transfers can occur in both directions simultaneously (each HyperTransport link is formed by two unidirectional links), aggregate transfer rates of 20.8 Gigabytes/second and 41.6 Gigabytes/second can be achieved with 16-bit wide and 32-bit wide link configurations respectively. To allow for system design optimization, the respective clock speed of the receive and transmit links may be set at different rates.

> Back to Top


12. What is the width of the HyperTransport I/O link bus?

HyperTransport links are fully scalable from 2-bit to 32-bit wide.  The two unidirectional links can have different widths so as to best tailor system architecture to specific performance, integration, applications and cost targets. HyperTransport devices transparently negotiate the bus width during system initialization and operate accordingly thereafter. To allow for further system optimization, the clock speed of the receiving and transmitting links may be set at different rates.

> Back to Top


13. With what buses and I/O technologies is HyperTransport technology compatible?

HyperTransport technology is fully compatible with PCI, PCI-X and PCI Express by means of embedded software mapping. In addition, because of its bandwidth and packetized data/command protocol, HyperTransport is easily extensible via bridge devices to any advanced I/O technologies, such as RapidIO, AGP 8x, Firewire, USB, InfiniBand, PL-3, SPI-4.2, SPI-5.0, and gigabit Ethernet. For more information on bridge devices, please refer to the HyperTransport Bridge Functions overview.

> Back to Top


14. What is the HTX connector standard?

The HyperTransport HTX connector is a HyperTransport Consortium standard that defines the mechanical and electrical characteristics of a slot-type, board-level connector enabling the direct connection of latest-generations high-performance, board-level subsystems to the system CPU, or CPUs in case o multi-processor systems.  The HTX connector allows compute-intensive subsystems, like co-processors, high performance computing (HPC) server cluster network interface adapters (NIC), communication controllers, storage controllers, etc. to interact with the system processor(s) and memory at high bandwidth and minimum latency via a dedicated HyperTransport-based link.  The current HTX specification supports 8- or 16-bit HyperTransport links and clock speeds up to 800 MHz, yielding 6.4 Gigabytes/second of aggregate bandwidth. HTX is based on a standard PCI Express mechanical connector (installed in reverse to prevent board misplacements) carrying HyperTransport-specific signals. By leveraging the economy of scale of the multi-sourced PCI Express connector, HTX is extremely cost-effective to implement. For more information on the HTX interface connector standard, please visit the HTX pages.

> Back to Top


15. Is HyperTrasnport technology plug-n-play compatible?

Yes. HyperTransport I/O devices are designed to use the standard plug-n-play methodology, and are boot, run, and driver compatible with any standard PCI compliant operating system.

> Back to Top


16. Does Hyper Trasnport technology support hot plugging?

The HyperTransport 3.0 specification introduced full hot plugging support in the I/O links as well as in the processor-to-processor, data-coherent links. For more information about hot plugging support, please refer to the HyperTransport 3.0 page.

> Back to Top


17. How can I get HyperTransport technology training?

MindShare, one of the leading technical training companies in the hardware industry, offers a comprehensive HyperTransport training course. The course details all facets of HyperTransport technology including protocol, electrical environment, error detection and reporting, and configuration. The HyperTransport course also provides the necessary background for understanding performance considerations when designing HyperTransport-based systems. MindShare has also published a detailed HyperTransport system architecture book. For more information on MindShare please click here.

> Back to Top

 

HyperTransport Technology-Enabled Products FAQs

18. What HyperTransport technology-enabled products are available?

There are a number of products in production today. For a complete list, please go to the HyperTransport Product page.

> Back to Top


19. What applications use HyperTransport technology?

HyperTransport is used in a wide range of performance-oriented applications, ranging from game consoles, embedded systems, networking systems, PCs, workstations, gaming systems, servers, server clusters, data center platforms and supercomputers.

> Back to Top