- Scan-Line Interleave
Scan-Line Interleave (SLI) from
3dfx is a method for linking two (or more)video card s or chips together to produce a single output. It is an application ofparallel processing forcomputer graphics , meant to increase the processing power available for graphics. SLI from 3dfx was introduced in1998 and used in theVoodoo2 line of graphics accelerators. However, the original Voodoo Graphics card and theVSA 100 were also SLI-capable.Nvidia Corporation reintroduced the SLI acronym in2004 (though it now stands forScalable Link Interface ) and intends for it to be used in modern computer systems based on thePCI Express bus.Function
3dfx's SLI design was the first attempt, in the consumer PC market, at combining the rendering power of two video cards. The two 3dfx cards were connected by a small
ribbon cable inside the PC. This cable shared graphics and synchronization information between the cards. Each 3dfx card rendered alternating horizontal lines of pixels composing a frame.Criticism
3dfx incorporated the notion of adding an extra video card into its already (relatively) extensive advertising campaigns. Advertisements in various video gaming magazines had contained such quotes as "double the hardware, double the power", or "Twice as fast". Although many believed that adding another expensive card would double frame-rates, this proved not to be the case.
Aside from this, successfully running the SLI configuration operational often proved difficult. At the time (as the case with the Voodoo2), some motherboards would not function with three video cards installed (i.e. 2d card, Voodoo2 #1 and Voodoo2 #2). In other cases, there were issues with conflicts as system resources in normal ranges were already accounted for, partly because many features that are today integrated into motherboards (such as the network interface and sound playback facilities) were at that time provided by add-on cards, each requiring additional system resources. Some systems lacked enough open PCI slots to support the setup and add-on cards were sometimes sacrificed, with the loss of hardware facilities.
If the technical issues were surmounted, it was further discovered that some applications had problems operating under SLI. Furthermore, those programs that did function often experienced tearing and other display artifacts resulting in a growing number of dissatisfied enthusiasts. Still, a small number of fans touted SLI as the future of gaming and as such, SLI became more of a status symbol than a viable system upgrade.
Towards the end of the phenomenon, even a perfectly functional SLI setup running an older game would rarely outperform an identical system with a single card. Some blamed this on the fact that the performance of 3d cards of the time were heavily reliant on the main CPU's speed for proper fill rates, and that there existed no viable CPU at the time to do SLI setup justice. Some felt that the PCI bus was too congested to allow for such things, while others blamed the way that the technology was implemented. Whichever was true, Dennis "Thresh" Fong, a long time advocate of SLI (and quite secretive about his system configuration), contributed to the growing dissent when he published his benchmarks in PC Gamer after the SLI fanfare had begun to decline. It showed an average of 2 frames per second faster than PC Gamer's own test computer using one card. Even this difference was attributed by some to the efficiency of Fong's systemic configuration as a whole, rather than the SLI arrangement.
Few actually experienced 3dfx SLI firsthand. By the time that the setup was cost effective, there were single card solutions that rivaled the theoretical capabilities of the SLI configuration, for lower costs. As a result, SLI received little widespread exposure, although many gamers still consider the 3dfx SLI setup to have been "the" setup to have for hardcore gaming despite the doubts of any significant performance advantage.
ee also
*
Scalable Link Interface
*ATI CrossFire External links
[http://helppage.3dfx.sk/#SLI The 3dfx Help Page]
Wikimedia Foundation. 2010.