Hacker Newsnew | past | comments | ask | show | jobs | submit | p5a0u9l's commentslogin

Are there statistics on the scale of compute available to GIMPS for this search? Is there any evidence that by crowdsourcing the clients, we are searching faster than, eg, a dedicated cluster financed by a government or a corporation? What is the impact of GIMPS as a distributed problem solving tool? Like, if there was a practical application, how much money would it take to exceed GIMPS throughput, that curious people provide for free?

I’d like it to be astronomical, but given the niche of this, and the low cost of cloud compute, the answer is predicable depressing, like, “$50k/year in AWS costs would equal current GIMPS search throughput”


https://www.mersenne.org/primenet/ suggests 127 PFlop/s average over the last month, which would put it in the top 10 on https://top500.org/lists/top500/2024/06/

I used a random estimate online for computing cost which had 5.6e17 Flops per dollar on A100s gives about a dollar every 4.4 seconds or ~$7 million per year.

Sadly, I do not vouch for the correctness of any part of this, though I did try.


> “$50k/year in AWS costs would equal current GIMPS search throughput”

I think you may be wrong by at least 3 orders of magnitude.


> none of them seem to have that basic "New Window with Profile ..."

iTerm has exactly that.


Yes, you are right! iTerm2 does -- as long as you have more than the default profile set up (and you have to set them up yourself, which is easy enough).


Umm. Well, mostly because tiktok is the only major social media company associated with a government that has a history of trying to steer US elections one way or another.


What is your definition of bare metal?


No virtual machines, zones/containers/jails.

Physical deployment with an operating system that ran on hardware directly.

The only management interface being the very anaemic out of band management controller.


Yes it can. Diffraction limit relates to resolution in space, not scattering.


Prose writing in any form is super useful. The challenge is having a unified tool to impose some structure. Instead, thoughts are divided across email, slack, confluence, quip, etc.


This is the way. I’ve used zettelkasten, which is similar philosophy. Dump a thought and tag it for retrieval later. I prefer making this work with simple files and markdown. I have a colleague who uses email drafts well. My problem with OneNote and similar is the bloat of it. But, having images alongside text, when needed, is super nice.


You can have images alongside text in VS Code extensions too. Yes there's still bloat but you may already have VS Code open.


I don’t think I could ever switch to a windowed app as editor, vs a TUI, eg neovim. The remote story is never great for me. It forces your editor to slowly bloat to become your entire IDE. Native remote dev using tmux is so nice. Can anyone persuade me otherwise?


> Native remote dev using tmux is so nice. Can anyone persuade me otherwise?

I sure as hell can't. SSH + Tmux has consistently been the only good pair-programming solution I've used in the past decade.


Tips please, esp configure vscode like nvim.


It comes down to using the vim extension and making use of the context it adds when setting key binds. Both in settings and keybind json files you set commands for certain vim modes, or bind native VSCode commands to your leader. Zed does almost the same but with no defined leader key so you just have to be more specific about the command and the context they are executed in.


Evidently this is all very new to you, sounding slightly histrionic.

The zed complaint is purely about it be auto enabled. For each language there is usually a standard and at least one tool. Most people want formatting and can’t stands code bases where sometimes it’s a single quote, sometimes a double quote.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: