Josh Allmann

> In the provided link I can't find any problems reported, can you point me to what they're reporting? @darkdarkdragon See the note on "Exceeding VRAM" : > Exceeding the...

> How can I distinguish between streams? I usually have a file naming convention. Something like: `out___.ts` and so forth. > So it's looks like it was one streams that...

> I'm using random names. Just took piece of code from our transcoder.go For the benchmarking, wouldn't it be useful to have a more consistent naming scheme to help diagnose...

> it was using around 7M memory on the card with 16M. Just 7MB ? Is that right? Also, how is the measuring done? `nvidia-smi dmon` and `pmon` only show...

Possibly related: cuda-memcheck is indicating issues within `cuvidDestroyDecoder` . This is also present in ffmpeg itself, which would seem to exclude issues on the golang / cgo side. If the...

> the network should expect failures and have resiliency built in. His suggestion was that you actually have n (5?) transcoders encoding each segment Curious how the incentives for that...

Does that come in handy? Usually I'm looking at the logs between adjacent peers, which usually don't match the stream ID (bootnodes etc). If we do need to print the...

> you lose the ability to request content with JUST a content identifier since the streamID encodes nodeID info that can be used to route to the node serving the...

> The ability to always route to the node serving content given a content identifier seems like the right architecture. While we should be able to access content given an...

60fps downsampled to 30fps seems reasonable for ABR, since it's a big savings. Naturally, a 30fps video shouldn't be upsampled to 60fps in this case.