The WAN Show: AMD Mantle, Phone Bloatware Banned, Source Engine 2, Jan 31th, 2014

The WAN Show episode from January 31st, 2014 covers a wide range of tech topics with Linus and Slick guiding the discussion through new hardware features, software policy moves, and ongoing hardware benchmarking debates. The crew starts by noting AMD Mantle and the Battlefield 4 patch, clarifying that Mantle is not yet widely playable due to driver availability but promising a potential performance uplift by reducing CPU bottlenecks in CPU-constrained scenarios. They discuss Mantle’s expected impact on multi-core CPUs and mid-range systems, stressing that gains depend on game configuration, resolution, and the exact hardware mix. The hosts propose interactive experiments to compare Mantle against DirectX in different hardware configurations, inviting viewer input via Twitter to determine whether to run high-end CPU with high-end GPU tests or focus on CPU bottlenecks on mid-range hardware. They also touch on Mantle’s broader ecosystem, including OpenCL 2.0 and HSA support, and note LibreOffice adding Mantle support as a surprising cross-application use case. The discussion then shifts to Valve’s Steam In-Home Streaming, with Linus recounting a detailed setup conducted for a video, the challenges of dual-network routing, and the impressive latency and image quality achieved on wired connections. They address edge cases, such as streaming from a low-power system where 1080p could be taxing, and discuss practical accessories like USB-to-Ethernet adapters to improve network reliability for streaming. The hosts highlight Valve’s public engagement, including the Steam Community announcements and ongoing beta feedback, framing the event as a milestone for game streaming beyond the living room. The episode then pivots to hardware reliability and failure rates, featuring Backblaze’s reported drive statistics and a thorough critique by Tweaktown about the testing methodology, sampling bias, and the influence of enclosure, age, and workload on drive failure. They emphasize how variability in consumer storage conditions complicates benchmarking and caution against taking sensationalized charts at face value, advocating for more controlled, apples-to-apples comparisons. Backblaze’s data is contrasted with enterprise-grade drives and the realities of field returns, revealing that consumer drives show wide variation and can be misinterpreted when aggregated without context. The hosts loop back to the broader theme of fair benchmarking and remind viewers that an experiment’s value lies in reproducibility and transparency, not in sensational headlines. The conversation then broadens to Google’s ISP ranking initiative, explaining how performance measurements for service providers are becoming more nuanced and region-specific, and how verification tiers like 720p, 1080p, and HD-verified can help users understand actual streaming performance. The show assesses how this data translates for consumers, noting that ISP speeds advertised may not reflect real-world streaming quality due to network contention, peering, and local infrastructure, and discussing potential regional discrepancies in perceived performance. The Mantle discussion resurfaces with a quick recap of driver releases including OpenCL 2.0 support for Kaveri APUs and the practical implications for non-gaming workloads such as spreadsheet calculations, where HSA acceleration could boost performance for financial and data-processing tasks. The panel also mentions library software like LibreOffice adopting Mantle-like acceleration paths, illustrating a broader shift toward cross-application optimizations beyond traditional gaming engines. The show explores in-home streaming experiences in greater depth, including latency, jitter, and tearing observed under various hardware configurations, and considers how different CPU-GPU mixes influence the perceived quality of the gaming experience when streaming. They close the segment by previewing follow-up benchmarking plans, including a Twitter-driven poll to steer future tests toward mid-range CPUs and entry-level GPUs, and tease additional experiments with Kaveri APUs and multiple GPUs. The episode weaves humor throughout, including lighthearted candy-themed skits and the ongoing banter between Linus and Slick, while maintaining a strong focus on practical, testable takeaways for enthusiasts looking to optimize hardware and software interplay. Finally, the hosts tease upcoming coverage of Oculus Rift topics, the potential for Source Engine 2, and more leaks around Half-Life 3 era discussions, staying true to the show’s habit of balancing current event analysis with speculative industry chatter. They remind viewers of the value of hands-on testing and encourage audience participation through polls, forum posts, and social media to shape the direction of future WAN Show content. The closing segments reflect on personal anecdotes from the hosts’ workflows, including gear recommendations, network gear upgrades, and plans for further home-lab experiments, underscoring the ongoing theme of technology accessibility and experimentation for a broad audience of tech enthusiasts.

Topics · Science & Technology · Technology News · PC Hardware · Software & Apps