16 comments

[ 0.67 ms ] story [ 20.5 ms ] thread
Last time posted was almost a year ago, the information is still very much relevent and interesting.
It is almost 8 years ago. Do we anything new to learn in this area?
Probably, if you started studying or learning software engineering sometime between then and now.

"Get off my lawn!"

this is a good article, I remember it from a while back. Does anyone know what the rule is on duplicates? How do some come through? I always just get redirected to a reference for the old posting.
Some content is evergreen. The software May account for that. With linked documents, the URI may be different as well.

The last meaningful discussion was about three years ago. The first was close to seven. The HN community has grown since both those times, but still deals with CPU memory.

For those complaining about duplicates, if the post is on the first page or if it has any upvotes it means someone found it helpful. Deal with it.
There is more than one side to that story. Not to take away from your argument, but dupes are a sign of stagnation. A chasm between the new users (for whom it's not a dupe) and the old. A high dupe ratio will eventually drive away the older members.

The reason people complain about dupes is because for each of these forums, one can get attached to its evolution. You want to see it grow. Every time it surprises you: yay, it's still becoming better!

Dupes can mean the beginning of the end. Not to say it's a good reason to complain, but it is one. Don't tell the older members who've seen this site grow and want the best for it to "deal with it."

Anyway, upvoted because I'm not one of the older members and I've never seen this before :)

I have no problem with duplicates. I do have a problem with the wildly overstated, linkbaity "every programmer needs to know about" meme.
This is an incredibly useful guide, but it starts out with transistor level stuff that I at least find much less relevant.

The good stuff starts at section 3, CPU Caches.

I found it insightful as it makes you realize the following important takeaway: we can still improve bandwidth, but not latency.