Working with modern architectures for high performance applications is increasingly more difficult for programmers as the complexity of both the system architectures and software continue to increase. The level of hand tuning and native adaptations required to achieve high performance comes at the cost of limiting the portability of the software. For instance, we show that a compute intensive DCT algorithm performs better on graphic processors than the best algorithm for x86. In particular, limited portability is true for cyclic multimedia workloads, a set of programs that run continuously with strict requirements for high performance and low latency. An example of a typical multimedia workload is a pipeline of many small image processing algorithms working in tandem to complete a particular task. The input can be videos from one or more live cameras, and the output is a set of video frames with elements from several of the source videos, for example as stitched panorama frames or 3D warped video. Such a setup runs continuously and potentially needs to adapt to various degrees of changes in the setup without interruptions or downtime. To reach the performance goal required by multimedia pipelines, modern, heterogeneous architectures are considered instead of the traditional symmetric multi-processing architectures. We also investigate variations between recent microarchitectures of symmetric processors to identify differences that a low-level scheduler must take into account. Further, since multimedia workloads often need to adapt to various external conditions, e.g., adding another participant to a video conference, we also investigate elastic and portable processing of multimedia work- loads. To do this, we propose a framework design and language, which we call P2G. In the age of Big Data, this idea differs from the typical frameworks used for distributed processing, such as MapReduce and Dryad, in that it is designed for continuous operation instead of batch process- ing of large workloads. We emphasise heterogeneous support and expose parallel opportunities in workloads in a way that is easy to target since it is similar to sequential execution with multidimensional arrays. The framework ideas are implemented as a prototype and released as an open source platform for further experimentation and evaluation.

Segmented adaptive HTTP streaming has become the de facto standard for video delivery over the Internet for its ability to scale video quality to the available network resources. Here, each video is encoded in multiple qualities, i.e., running the expensive encoding process for each quality layer. However, these operations consume both a lot of time and resources, and in this paper, the authors propose a system for reusing redundant steps in a video encoder to improve the multi-layer encoding pipeline. The idea is to have multiple outputs for each of the target bitrates and qualities where the intermediate processing steps share and reuse the computational heavy analysis. A prototype has been implemented using the VP8 reference encoder, and their experimental results show that for both low- and high-resolution videos the proposed method can significantly reduce the processing demands and time when encoding the different quality layers.

P2G is a framework designed to integrate concepts from modern batch processing frameworks into the world of real-time multimedia processing, where we seek to scale transparently with the available resources. P2G consists of a compiler and run-time that analyzes dependencies dynamically and merges or splits kernel instances based on resouce availability and performance monitoring.

De fleste datamaskiner har i dag en multikjerneprosessor. Kanskje du også har et grafikkort i datamaskinen din, eller kanskje du har en Playstation 3? Da har du også en asymmetrisk multikjerneprosessor - som stort sett bare utnyttes når du spiller spill!

Even though shared-memory concurrency is a paradigm frequently used for developing parallel applications on small- and middle-sized machines, experience has shown that it is hard to use. This is largely caused by synchro- nization primitives which are low-level, inherently nondeterministic, and, consequently, non-intuitive to use. In this paper, we present the Nornir run-time system. Nornir is comparable to well-known frameworks such as MapReduce and Dryad that are recognized for their efficiency and simplicity. Unlike these frameworks, Nornir also supports process structures containing branches and cycles. Nornir is based on the formalism of Kahn process networks, which is a shared-nothing, message-passing model of concurrency. We deem this model a simple and deterministic alternative to shared-memory concurrency. Experiments with real and synthetic benchmarks on up to 8 CPUs show that per- formance in most cases scales almost linearly with the number of CPUs, when not limited by data dependencies. We also show that the modeling flexibility allows Nornir to outperform its MapReduce counter-parts using well-known benchmarks.

When used efficiently, modern multicore architectures, such as Cell and GPUs, provide the processing power required by resource demanding multimedia workloads. However, the diversity of resources exposed to the programmers, intrinsically requires specific mindsets for efficiently utilizing these resources - not only compared to an x86 architecture, but also between the Cell and the GPUs. In this context, our analysis of 14 different Motion-JPEG implementations indicates that there exists a large potential for optimizing performance, but there are also many pitfalls to avoid. By experimentally evaluating algorithmic choices, inter-core data communication (memory transfers) and architecture-specific capabilities, such as instruction sets, we present tips, tricks and troubles with respect to efficient utilization of the available resources.

Today, major newspapers and TV stations make live and on-demand audio/video content available, video-on-demand services are becoming common and even personal media are frequently uploaded to streaming sites. The discussion about the best transport protocol for streaming has been going on for years. Currently, HTTP-streaming is usual although the transport of streaming media data over TCP is hindered by TCP's probing behavior, which results in the rapid reduction and slow recovery of the packet rates. On the other hand, UDP has been criticized for being unfair against TCP, and it is therefore often blocked by access network providers. To exploit benefits of both TCP and UDP, we have implemented a proxy that performs transparent protocol translation in such a way that the video stream is delivered to clients in a TCP-compatible and TCP-friendly way, but with UDP-like smoothness. The translation is related to multicast-to-unicast translation and to voice-over-IP proxies that translate between UDP and TCP. Furthermore, it is also similar to the use of proxy caching that ISPs employ to reduce bandwidth demands. The unique advantage of our approach is that we avoid full-featured TCP handling on the proxy server but still achieve live protocol translation at line-speed in a TCP-compliant, TCP-friendly manner. Although we discard packets just like a sender of non-adaptive video over TCP, we achieve higher user-perceived quality because our proxy can avoid receive queue underflows in the proxy, while also achieving the same average bandwidth as a TCP connection between proxy and client. In this demo, we present our prototype implemented on an Intel IXP2400 network processor. The prototype proxy does not buffer outgoing packets, yielding data loss in case of a congested TCP side. Comparing HTTP-streaming from a web-server and RTP/UDP-streaming from a video server shows that, in case of some loss, our solution using UDP from the server and a proxy that translates to TCP delivers a smoother stream at playout rate while the end-to-end TCP stream oscillates heavily.

The transport of streaming media data over TCP is hindered by TCP's probing behavior that results in the rapid reduction and slow recovery of the packet rates. On the other side, UDP has been criticized for being unfair against TCP connections, and it is therefore often blocked out in the access networks. In this paper, we try to benefit from a combined approach using a proxy that transparently performs transport protocol translation. We translate HTTP requests by the client transparently into RTSP requests, and translate the corresponding RTP/UDP/AVP stream into the corresponding HTTP response. This enables the server to use UDP on the server side and TCP on the client side. This is beneficial for the server side that scales to a higher load when it doesn't have to deal with TCP. On the client side, streaming over TCP has the advantage that connections can be established from the client side, and data streams are passed through firewalls. Preliminary tests demonstrate that our protocol translation delivers a smoother stream compared to HTTP-streaming where the TCP bandwidth oscillates heavily.