WTF is this thing? - RAM on a PCI Card??

The video opens by situating the Gigabyte IRAM RAM-on-a-PCI-card in its historical context, describing a time when consumer SSDs were not yet mainstream and DDR2 memory was driving a major upgrade path for gamers. The host explains the problem: boot speed and system responsiveness were bottlenecked by the slow, traditional hard drives of the era, creating demand for faster boot devices. The concept behind the IRAM is to repurpose leftover DDR1 memory as a boot drive by placing it on a PCI card, with an FPGA acting as a translator between the SATA interface and the memory controller. The discussion then delves into the hardware design: PCI’s limited 133 MB/s shared bus, the role of the XYLink FPGA, and the decision to avoid custom silicon due to a tiny initial production run. The host details how RAM differs from hard drives in data organization, the volatile nature of DRAM, and why preserving data across power cycles required supplementary battery back-up logic on the card. He also contrasts the theoretical speed benefits with the practical constraints, noting that the IRAM could outperform early HDDs and approach SSD-like performance only in certain scenarios, but suffered from limited capacity and high costs. The segment concludes with a practical teardown, showing Windows installations on the IRAM, the necessity of power retention to maintain boot state, and final reflections on why the concept ultimately did not take off. The video frames the IRAM as a historical stepping stone toward faster storage technologies like Intel Optane, while also acknowledging its limitations, such as volatility, lack of ECC support, small capacity, and steep price. The host closes by inviting viewers to explore Optane further and hints at upcoming reviews of SSDs, underscoring the IRAM’s place in the evolution of memory and storage in consumer PCs.

Topics · technology · computing · hardware