256 GB Is Not 256 GB

Researched and written in cooperation with Claude.ai. Sources and caveats are listed at the end. Here are two ways to put 256 GB of accelerator memory on your floor. The first is an NVIDIA DGX-1: eight Tesla V100s at 32 GB each, a petaFLOPS of FP16 tensor throughput, 7.2 TB/s of aggregate HBM2 bandwidth, and a list price…

Continue reading 256 GB Is Not 256 GB

Why EIGRP is the only IGP that does unequal-cost load balancing

(This is a REALLY old post that never got finished because for the most part the research never was complete. I finish it now because I feel that the answers are, for the most part, readily available and the topic is not complete speculation.) -ORIGINAL POST IDEA- I’d always kind of wondered about a few…

Continue reading Why EIGRP is the only IGP that does unequal-cost load balancing