Remote profile

@camelcdr yes, right, I wasn't looking closely enough at the actual op.

I guess I was thinking of that as a different issue: that's sort of the long-decoder issue, where you have to decide how well the implementation is going to handle instructions that generate more uops than it can deal with in a group (which may or may not be equal to your rename bandwidth). That's what I've seen be the issue with long-LMUL performance, and that doesn't change when you move around the V rename point. Renaming V later just lets you know what the actual uop requirement is without predicting.

and FWIW: it's definitely possible to decouple the rename granularity from the architectural register size. I agree it's probably not a good idea ;)

@camelcdr

> Consider scalar1, m8rvv, scalar2. In a 4-wide design, the "latency" focuses RVV implementation would decode scalar2 after 3 cycles, but the "throughput" one on the first.

That's assuming the latency-focused design doesn't predict the vector config, which is where the rabbit hole of edge cases comes from...

@camelcdr IMO where you place the vector config handling in the pipeline really depends on what performance characteristics you're looking for. Handling it after scalar rename adds latency in the common case, but gets rid of a bunch of complexity that can result in really bad issues in the uncommon cases.

The way the ISA is designed certainly favors the higher-latency designs, but IMO the right decision depends on what sort of workload characteristics you're expecting to see. Unfortunately it's a hard tradeoff to make: the latency-sensitive workloads also tend to be the irregular ones that are harder to cover pre-silicon, so the chance of missing an edge case that destroys performance is high.

So designers are sort of forced to decide between rolling the dice on the lower-latency implementation or sticking with the higher-latency implementation and accepting the inherent performance penalties on some workloads.

I suspect we're going to see two very different styles of RVV implementation show up over the next few years as vendors optimize for one style or the other, and then various ISA changes go through that work around the issues in each.

@CosmicGoat ya, I've been doing a few years of winter riding and sort of built it up over time as my commutes got longer and I got more used to riding in the cold (and thus did longer/colder rides). Each year I basically just rode a bit farther into the winter until I just didn't stop.

There's a lot of cheaper options, too. Toe covers are a much cheaper substitute for the shoes, for example. As long as it's not too wet out they're fine, just get some cheap toe covers to block the wind and then socks for warmth.

Just don't give up because people say you're crazy, cold isn't really an issue for riding.

@CosmicGoat The coldest it gets around here is about 35F and I used to do 2 hours before sunrise when my office was farther away. I ended up with a pretty comfortable setup, it's not cheap but you can find this stuff on pretty steep sales if you look around for it.

I have the 45NRTH boots. They're basically snowboard boots that you can put two-bolt cleats on. They're kind of clunky, but you get used to them. If it's dry I'll just use toe covers on regular shoes, but it's frequently wet in the winter around here and the boots really help.

I've found a full face helmet works really well. It keeps the air off my face so I don't need anything over my mouth, which makes sweat management much easier. I also have a winter multi-sport helmet, but I find the full face works better for longer rides. I just use a regular summer full face helmet and wear light weight ski gear under it.

I haven't found a great setup for gloves. I have a few pairs meant for skiing, but they're too bulky to really be happy on my road levers. They're fine on MTB levers, though, and I usually use those when it's wet out anyway so it's not such a big deal.

I don't do anything particularly special for the rest of the clothes, basically just layers on top of regular bike kit. It doesn't really rain around here, it just sort of mists, so that makes things simple.

@ascentale @iris @bikenite I briefly tried an Ortlieb "Backpack Carrying System for Panniers" but didn't like it. I think the general idea is viable, their implementation just didn't fit my use case.

@SnoopJ it's a 6V DC LED bulb attached straight to an unregulated dynamo? That sounds like it's asking for trouble, IIUC dynamo voltage is all over the place in a way that really just eats modern electronics.

Looks like there are some parts for 6V motorcycles that might be a good option here?