I am a devops and software engineer. So I like trying to write performant software.
I like to see what performance numbers I get from my CPU.
What do you use your laptop for that you need such a powerful laptop?
Just web browsing. 64GB ought to be enough for anybody to run Chrome.
2 Likes
I got the i7, 32 GB of memory, and the 1 TB SSD because I’m a computer science graduate student. In the past on cheap laptops, I’ve hit limits when doing academic projects where I wanted to run three VMs or process a 50 GB dataset (saving a 50 GB intermediate result after each stage). My school will give me access to a high performance computing cluster as needed, but it’s nice to be able to do some things locally.
99% of the time, of course, I’m using only a tiny fraction of this machine’s resources to browse the web and read PDF papers. I’ve always said that most people overestimate what they need in terms of computing power, memory, bandwidth, etc. I got the i5 with 8 GB of memory for my wife, and it flies despite her running Windows and Chrome with dozens of open tabs.
1 Like
Browse alternate news and videos for boots-on-ground viewpoints, play with computer languages for fun learning, some photography, read and write articles though try to stay above political fray, videos on how to fix things, of course email, some social media, jigsaw puzzles to keep left and right brain cooperating. Etc.
My laptop is going to be a desktop replacement
I have an eGPU ready for gaming and a NAS for storage + servers
I mostly wanted a laptop to take with me to work
At home I’m too distracted to code but I have tons of free time at work to do so
1 Like
@JRLarsen I think eGPUs are a great idea. I would also have a dock and attach it to monitors and SSD enclosures or eSATA.
Someone on Quora told me the problem with eGPUs is bandwidth. Thunderbolt doesn’t have enough bandwidth to run a truly powerful GPU.
I have been doing rough calculations of how many billions of records can a computer search fast and how long does it take from a CPU performance perspective.
The problem is it takes 100 nanoseconds to read from main memory to fill a cache line of 64 bytes at a time. And 0.5 nanoseconds to retrieve a value from cache. So to read 1 billion record locations can be read 1E9/64* 100 is 1.5 seconds per billion bytes. Multiply by 8 cores and you have 8 billion bytes cache miss performance.
So in 1 second you can read 12 billion bytes or 12 gigabytes per second.
Fortunately we can divide this search over cores and split the search over multiple machines at the same time.
Most of the time the number of records being fetched is below a billion so you can search through a lot less to find what you’re looking for…especially if your app is paginated.
Professional chocolatier. I travel to Central and South America and the Caribbean a fair amount, and teach masterclasses over Zoom. Currently using it to write and publish a research paper on cacao origins.
1 Like
Except on alternating Thursdays. Yep, you found me out.
1 Like