Kjetil Raaen

Can gamers detect cloud delay

In many games, a win or a loss is not only contingent on the speedy reaction of the players, but also on how fast the game can react to them. From our ongoing project, we aim to establish perceptual thresholds for visual delays that follow user actions. In this first user study, we eliminated the complexities of a real game and asked participants to adjust the delay between the push of a button and a simple visual presentation. At the most sensitive, our findings reveal that some perceive delays below 40 ms. However, the median threshold suggests that motor-visual delays are more likely than not to go undetected below 51-90 ms. These results will in future investigations be compared to thresholds for more complex visual stimuli, and to thresholds established from different experimental approaches.

How much delay is there really in current games

All computer games present some delay between human input and results being displayed on the screen, even when no networking is involved. A well-balanced discussion of delay-tolerance levels in computer games requires an understanding of how much delay is added locally, as well as in the network. This demonstration uses a typical gaming setup wired to an oscilloscope to show how long the total, local delay is. Participants may also bring their own computers and games so they can measure delays in the games or other software. Results show that local delays constitute such a large share of the total delay that that it should be considered when studying the effects of delay in games, often far exceeding the network delay evaluated.

LEARS: A Lockless, Relaxed-Atomicity State Model for Parallel Execution of a Game Server Partition

Supporting thousands of interacting players in a virtual world poses huge challenges with respect to processing. Existing work that addresses the challenge utilizes a variety of spatial partitioning algorithms to distribute the load. If, however, a large number of players needs to interact tightly across an area of the game world, spatial partitioning cannot subdivide this area without incurring massive communication costs, latency or inconsistency. It is a major challenge of game engines to scale such areas to the largest number of players possible, in a deviation from earlier thinking, parallelism on multi-core architectures is applied to increase scalability. In this paper, we evaluate the design and implementation of our game server architecture, called LEARS, which allows for lock-free parallel processing of a single spatial partition by considering every game cycle an atomic tick. Our prototype is evaluated using traces from live game sessions where we measure the server response time for all objects that need timely updates. We also measure how the response time for the multi-threaded implementation varies with the number of threads used. Our results show that the challenge of scaling up a game-server can be an embarrassingly parallel problem.

The Effects of Delay on Game Actions: Moving Target Selection with a Mouse

In modern computer systems, user input, particularly for computer games, is affected by delay from local systems, networks and servers. While general awareness of the degradation effects of delay on player performance and quality of experience are well known, an understanding quantifying how specific player actions are impacted by delay is missing. This work presents a user study that gathers data on player actions for a range of delay and game conditions for the fundamental game action of selecting a moving target with a mouse. Analysis shows player sensitivity to delays in all conditions, with particular sensitivity when targets are fast. From the data, we derive a simple analytic model that is a promising step towards a broadly applicable tool to better understand and compensate for delay in games and interactive applications.

Instantaneous Human-Computer Interactions: Button Causes and Screen Effects

Many human-computer interactions are highly time-dependent, which means that an effect should follow a cause without delay. In this work, we explore how much time can pass between a cause and its effect without jeopardising the subjective perception of instantaneity. We ran two experiments that involve the same simple interaction: A click of a button causes a spinning disc to change its direction of rotation, following a variable delay. In our adjustment experiment, we asked participants to adjust the delay directly, but without numerical references, using repeated attempts to achieve a value as close to zero as possible. In the discrimination task, participants made judgements on whether the single rotation change happened immediately following the button-click, or after a delay. The derived thresholds revealed a marked difference between the two experimental approaches, participants could adjust delays down to a median of 40 ms, whereas the discrimination mid-point corresponded to 148 ms. This difference could possibly be an artefact of separate strategies adapted by participants for the two tasks. Alternatively, repeated presentations may make people more sensitive to delays, or provide them with additional information to base their judgements on. In either case, we have found that humans are capable of perceiving very short temporal delays, and these empirical results provide useful guidelines for future designs of time-critical interactions.

Games for Research: A Comparative Study of Open Source Game Projects

Video games have proved to be an interesting platform for computer scientists as games demand the latest technology, fast response times and effective utilization of hardware. Finding the right games to perform experiments are however difficult. Some researchers create their own smaller prototype games to test their ideas, without performing tests in larger scale productions, which decreases the practical applicability of the conclusion. An important reason is the lack of suitable games for research. This paper proposes a list of qualities and features required by researchers for a video game to be suitable for computer science research. Further, it evaluates four games with open source code and discuss their usefulness. We also consider the current state of open source games and possibilities for enhanced cooperation between the professional and research communities.

A demonstration of a lockless, relaxed atomicity state parallel game server (LEARS)

Games where thousands of players can interact concurrently pose many challenges with regards to the massive parallelism. Earlier work within the field suggests that this is difficult due to synchronization issues. In this paper, we present an implementation of a game server architecture based on a model that allows for massive parallelism. The system is evaluated using traces from live game sessions that has been scaled up to generate massive workloads. We measure the differences in server response time for all objects that need timely updates. We also measure how the response time for the multithreaded implementation varies with the number of threads used. Our results show that the case of implementing a game-server can actually be highly parallel problem.

Modeling User Performance for Moving Target Selection with a Delayed Mouse

The growth in networking and cloud services provides opportunities to host multimedia on remote servers, but also brings challenges to developers who must deal with added delays that degrade interactivity. A fundamental action for many computer-based multimedia applications is selecting a moving target with the mouse. While previous research has modeled both moving target selection and target selection with delay, there have not been models of moving target selection with delay. Our work presents a user study that measures the effects of delay and target speed on the time to select a moving target with a mouse, with analysis of trends and derivation of a model. The analysis shows delay and speed impact target selection time exponentially and that selection time is well-represented by a model with three terms - two with exponential relationships for delay and speed and one an important interaction term.

Is Today’s Public Cloud Suited to Deploy Hardcore Realtime Services?

“Cloud computing” is a popular way for application providers to obtain a flexible server and network infrastructure. Providers deploying applications with tight response time requirements such as games, are reluctant to use clouds. An important reason is the lack of real-time guarantees. This paper evaluates the actual, practical soft real-time CPU performance of current cloud services, with a special focus on online games. To perform this evaluation, we created a small benchmark and calibrated it to take a few milliseconds to run (often referred to as a microbenchmrak). Repeating this benchmark at a high frequency gives an overview of available resources over time. From the experimental results, we find that public cloud services deliver performance mostly within the requirements of popular online games, where Microsoft Azure Virtual machines give a significantly more stable performance than Amazon EC2.

How to demonstrate delay?: let's play!

The speedy interactions of many computer games demand prompt reactions from player and system alike. This demonstration invites participants to experience in-game latency. We present variations of two familiar games that use either mouse or key inputs to control actions seen on the monitor. While playing the games, participants can make changes to the temporal offset between a motor input and a resulting visual action, providing them with a direct experience of motor-visual delays. In turn, they will gain an understanding of how performance may suffer when trying to compensate for such delays.