[JohnBooty]

What I have found is that even among talented senior engineers there is massive Dunning-Kruger effect when it comes to performant architecture. They don't know how to do it, and they don't know that they don't know how to do it.

    I have always wanted reasonable performance (though this 
    might appear like “performance programming” to a concerning 
    proportion of the software industry),

This hit me right in the heart.

I'm often the only person on the team who cares about performance, so I am drawn to these performance-related challenges... and it has really hurt my career, because I am then often perceived as some kind of person focused on optimization rather than delivering features.

Like the author I do not focus on performance for performance's sake. Nearly all of the time the right course of action is "do not optimize this piece of code; focus on maintainability and readability instead."

However, sometimes, you really need to design for performance from the beginning.... or you don't have a product.

At my most recent job they were trying to push lots of data through RabbitMQ/Celery for scientific work. This worked for trivial jobs (tens of megabytes) but not for moderate or large ones (hundreds of gigabytes)

To make such a product viable, you really need to consider performance from the start. Celery explicitly tells you not to pass non-trivial amounts of data around: you should be passing pointers to data (database ID's, file paths, S3 URIs, whatever) rather than the actual full-fat data.

The team really struggled with this. Their next proposed solution was "well, okay, we'll store intermediate results in the database and 'optimize' later" Great idea, but this involved 1B+ result objects. Wrong again. You are not serializing 1B+ Python objects, sending them over the wire, and performing 1B+ Redis or Postgres inserts in any reasonable amount of time or memory. Optimize and bulk insert all you want, but that's an absolute dead end.

There aren't a whole lot of options for performantly slinging around hundreds of gigabytes of data. Assuming you can't just run on a monster server with hundreds of GB of RAM (which honestly is often the right answer) you are generally going to be looking at fast on-disk formats like Parquet etc. In any event that's something you really need to design around from the start, not something you sprinkle on at the end.

They're on their second iteration of the architecture right now, and it's slower than the first iteration was. Still no viable product. Shame.

[kstrauser]

> I'm often the only person on the team who cares about performance, so I am drawn to these performance-related challenges... and it has really hurt my career, because I am then often perceived as some kind of person focused on optimization rather than delivering features.

Pro tip: always turn that kind of thing into a dollar value you can put on your annual review. "My update to X let us use Y fewer EC2 instances, saving us $Z per year." Then it's not some Don Quixote obsession, but a clear fiscal benefit to the company.

[JohnBooty]

This is kind of the gold standard and well worth aspiring to for any given situation.

I've often found it hard to achieve in practice.

A typical issue is that for non-trivial performance improvements (the kind that will take several days to several months to achieve) it's hard to estimate the speedup (and therefore, the the corresponding dollar amount) without actually doing a substantial portion of the work and estimating the corresponding speedup.

Another typical issue is where the performance speedup doesn't correspond directly to a dollar amount. At a recent employer we were working on a scientific product. The optimizations would have affected the size of the datasets able to be handled by our product. So there would have been a dollar impact, but it would have not been some kind of relatively simple AWS math. It would have been measured in potential sales dollars.

It's like making your game run at 60fps instead of 30fps. 60fps is clearly better (all other things being equal) and better games tend to sell better, and we'd like to make the best game possible and sell the most copies possible... but how do we quantify the expected return on this work? Many times, I don't think you can.

[stephc_int13]

I think that software performance is the good kind of vanity metric, speed almost always translate to better end-user experience, but there is a point where it can turn into a pointless rabbit hole.