logicchains

If I do that, though, performance degrades each run as the cpu heats up. On 21 December 2014 at 01:26, Erwan Ameil [email protected] wrote: > I'm looking at _D_'s results...

I've found the same; on ARM it's around 3.5 seconds for this implementation, compared to around 3.0 for the previous one. On 3 January 2015 at 09:39, Cosmin Gorgovan [email protected]...

Cool, I've added it to the makefile and run script. I found the Use Highbit one to be fastest.

I added them. I don't suppose you have a different D compiler handy? Dmd is generally the slowest by a significant margin, so it's not the fairest for benchmarking While...

> It would be interesting to have many versions written in each language to > show the spread of performance within that language for plausible > implementations, and then plot...

That would be great! I was originally planning to add one but then I found the language seemed like it would take me quite a while to learn, as I...

Thanks! Is there some way I can generate a prompt of exactly 1024 tokens? E.g. maybe some character sequence that I could repeat 1024 times?

They've just updated their Huggingface to confirm that the models are now available under Apache 2.0: https://huggingface.co/tiiuae .

Could you submit it as a separate file, e.g. cpppruned.cpp, and I'll include that. Makes it clear that it's using a different algorithm to the other implementations. On 21/12/2014 1:56...

Thanks for pointing that out. It must have crashed while computing it; it should be there. On 22 December 2014 at 20:02, gasche [email protected] wrote: > The first version of...