Very delayed model runs in recent weeks

[Ken]

Just wondering why the runs of models like ECMWF, ACCESS-C, etc seem to have been consistently coming into Windy noticeably later over the past few weeks than they used to?

For example, in the case of ECMWF, its 12z forecasts consistently used to become available in Windy from around 2010z.
This was late enough as it was, given that many other weather websites start making 12z ECMWF data available from about 1800z while the run is still in progress, before the full run becomes available by around 1900z on those sites.
But now, the 12z ECMWF doesn't usually start appearing in Windy until after around 2040z (more than 2 hours after the unfinished run starts appearing on other weather websites and 1.5 hours after the full run appears on those other sites). This is also a full 9 hours after the 12z run time which is a long time when weather is evolving rapidly.

The ACCESS-C model used in Australia is another example where its 18z run used to become available in Windy from around 2035z but now, it's more like around 2050z or so.

[Suty]

@Ken Hello, I forwarded this feedback to our colleagues and they are aware of this. The reason is getting much more new parameters and data from these models. They are working on optimizations but it can't be handled immediately, rather with the continuous progress.

Thank you for potential patience.

[Ken]

@Suty thanks for that.