Piggybacking as you may find it useful. I built https://flopper.io and aim to do this differently that OP. We started with datasheets and individual GPUs. We're trying to get as many as we can.
Currently trying to work to have a plugin of similar to post back to get accurate benchmarks.
Yes, we're trying to get those ingested. We have quite a few but you raise a good point that inference at home is rising!
Part of the problem we have is some of those desktop GPUs don't have power and flops data. We're toying with the idea of writing a CLI to run/publish these but obviously a lot of work.
These three are measured against the old model of Llama 2 7B, but it also means the numbers are without any per-model quirks that reduce compute and memory BW usage, thus more or less generically applicable.
Added models and pricing there recently too. Aiming to be a huge place for all GPU data when it comes to mathematical performance. Things like FP8 for AI inference etc.
Tempted to make the infrastructure a bit more sound.
Aside from that I'm working on https://solarable.org - which is aiming to provider understanding of solar panels before finding an installer and plotholes.app (https://plotholes.app) which is a consensus network for potholes as reported and determined by cyclists.
Would be cool if you could add coverage, for each GPU, of both rental price as well as purchase price (i.e. buy price := one of the spectrum of completed transactions on ebay, liquidator price, second hand price, new price, MSRP).
Always found the relationship between rental price and purchase price a bit peculiar.
I worry about the safety element of this. I'm not sure that having people walking around completely oblivious to their environments is a good thing.
In a multiple of ways: you become disconnected from nature and closed off to environments, and you also reduce safety because you don't know what is going on around you (high % of population live in cities).
People drive cars every day, isolated in a box with windows rolled up, playing music or radio, not being able to hear what's going on around them. So I don't buy the safety argument of noise canceling headphones being dangerous.
If we’re going there, cars are a leading cause of death and the leading cause of death for children and young adults.
That said, the masses have had noise cancelling earbuds for years at this point and while there definitely are cases of people getting hurt or killed from being unaware of their surroundings, it’s not really been that big a deal.
I wonder if that stat is true if removing the kids in gangs from the data. One cause of death is relatively undiscerning and applicable to all, but the other cause of death might only apply to a few. But I don’t know, I’d like to see the actual stat.
Well, that’s exactly why I worry when walking around with noise canceling phones. Because all those people driving cars oblivious to their surroundings
Yeah, I get your point, I jusr don't think the solution is to police pedestrians. I'll stop wearing noise canceling headphones in the city the day cars stop being loud, and when the drivers of those cars drive with no music and their windows rolled down ;)
>In a multiple of ways: you become disconnected from nature and closed off to environments
I live in a city. Nature for me means the squealing breaks of cars and trucks + people trying to get my attention so I'll take their religious flyer. No thanks.
That's also why there are different designs, e.g. in ear, over ear, open over ear, bone conduction, and technologies, e.g. noise cancelling, active noise cancelling, non voice filtering.
Typically there are visuals though, e.g. crossing lights for pedestrian, so the risk from the lack of hearing in a city is quite low IMHO.
FWIW in Belgium you can even legally ride a bike with earphones, which shows the risk is estimated to be very low.
Currently trying to work to have a plugin of similar to post back to get accurate benchmarks.
Take a look if you're interested!
reply