What is AI acceleration?

Isn't AI game-specific? How can a “dedicated hardware paradigm” help with such a problem?

What AI routines are accelerated by AIseek’s Intia™ processor?

How is AIseek’s Intia processor uniquely suited to AI?

What is AI acceleration?
Graphic chips revolutionized the quality of computer games by off-loading CPU-intensive graphical computations to dedicated hardware. AI acceleration does the same for the CPU-intensive computations of artificial intelligence. AIseek’s Intia processor is the world’s first AI accelerator – a dedicated processor for game AI. 

Isn’t AI game-specific? How can a “dedicated hardware paradigm” help with such a problem?
Game AI is best thought off as a hierarchy or pyramid. The top layers of the pyramid encompass “high-level” AI (such as game-level or character-level decisions). The bottom layers represent “low-level” AI (including tasks such as pathfinding, line-of-sight, and terrain analysis). These low-level tasks are characterized by three features: they are based on common, well-known algorithms; they are CPU-intensive; and they constitute the necessary “building blocks” that underlie high-level AI. These three features make low-level AI well-suited for off-loading to dedicated hardware. 

What AI routines are accelerated by AIseek’s Intia processor?
AIseek’s Intia processor accelerates low-level AI tasks. The routines accelerated include movement (in particular, pathfinding and group movement); sensory simulation (in particular, line-of-sight) and terrain analysis. 

How does acceleration of low-level AI tasks help the development of higher-level AI?
In general, sophisticated character-level or game-level decisions depend heavily on lower-level AI functions. For example, consider an NPC seeing a city at a distance, deciding to go to the city to sell an item, and then traveling to the city through the woods. The NPC’s sophisticated decision (“I want to sell this item”) depends on lower-level functions (perception, movement). A similar analysis would apply to the case of an NPC seeing an enemy, deciding to attack that enemy, and finally moving in the right direction to attack. By eliminating the CPU bottleneck associated with low-level AI, the Intia processor frees developers to create more intelligent characters and more believable game worlds. 

How fast is AIseek’s Intia processor?
One of best benchmarks for the speed of the Intia processor is its performance on several low-level AI tasks. For pathfinding, the processing time for each 100 nodes of path depth is only 10µs, making the Intia processor about 100-200 times faster than a comparable software solution (specifically, the A* algorithm). For line-of-sight calculation, the Intia’s checks are several orders of magnitude faster than current software methods: 512 agents can be checked against 512 agents (“who sees whom”) in a mere 0.02s. 

Does the pathfinding algorithm of AIseek’s Intia processor use heuristics?
No. Unlike today’s software-based approaches (e.g., A*), the Intia’s pathfinding uses no heuristics, thereby guaranteeing that the optimal path will always be found. This optimality also means that the Intia processor avoids the common pitfalls of A*, including failures to find a path when one exists, and the generation of “artifacts” (e.g., weird, unrealistic paths). If a path exists, the Intia will always find it. Furthermore, the optimality of the Intia’s pathfinding includes excellent support for tactical considerations. Thus, in finding the required path, the pathfinding algorithm can take full account of any tactical information, such as the need to find a path that passes through certain locations (e.g., hiding points, enemy positions). 

Does AIseek’s Intia processor support highly dynamic worlds? Is pre-processing of the game map required?
The Intia processor offers excellent support for highly dynamic worlds. Technically speaking, the Intia processor does not require pre-processing of the game map. Thus the game world can change continuously (through explosions, the construction of new buildings etc.), with the Intia processor rapidly re-calculating the required AI computations (e.g., pathfinding, line-of-sight, terrain analysis) in real-time. In fact, the processor’s strong support for constantly changing maps is a major contribution towards the creation of more believable, life-like game worlds. 

What game graphs formats are supported by AIseek’s Intia processor?
To facilitate the integration of the Intia processor with existing studio AI modules, all common game graph formats are supported. These formats including grids, navigation meshes and waypoint graphs. 

What games genres will benefit most from accelerated AI?
Obviously, AI-intense genres will benefit most from accelerated AI. These include RTS, RPG and Action/Shooters. Here are a few examples. With accelerated AI, RTS worlds can be large, complex and highly dynamic, with thousands of AI-controlled units moving and behaving intelligently at all times. In RPG games, cities and landscapes can come alive with thousands of fully-simulated characters. With full simulation, all characters remain alive and intelligent both on-screen and off-screen (i.e., no loss of AI level-of-detail), resulting in much more believable, life-like worlds. For Action games and Shooters, accelerated AI means no more dumb enemies. Computer-controlled combatants can now team-up cleverly to give even the most experienced player a worthy fight. 

How is AIseek’s Intia processor uniquely suited to AI?
Computationally speaking, AI problems are typically search-intensive, translating into large graphs with a high degree of branching. The Intia processor tackles these specific processing challenges with a technological breakthrough – the Graph Processing Core (GPC). The GPC lies at the heart of the Intia processor, delivering fast and optimal solutions to graph-based computations.