It is much, much easier to develop in Linux compared to Windows.

Nothing preventing a good Windows developer from tackling a N-body gpu app.

But no reason to NOT develop an enhanced N-body gpu app for Linux users.

The user always has the option of running Linux.

I have Linux in an Oracle VirtualBox, so I can run the Linux CPDN tasks. Would that run GPU apps? I can't see a way to allow it to compute on the GPU, just pass a "display" across.

Running Linux only on a machine is beyond a joke. It's so over the top on security it's unusable. Not being allowed to edit your own .xml files in Boinc on your own computer on a program you installed yourself for goodness sake. I log onto mine as root so I don't punch the screen.

I'll just keep chugging along until power prices completely cripple me (which isn't far off when you have a mortgage & family to support) Prices are rising again by something like 30% on 1st July :( On that basis my next bill would be about AUD$2200 for the quarter which is ridiculous.

I think they're going down in the UK, not that I pay much attention to the news.

So... The next question is what can I move to that's space or environment related and works on 32-bit Win 7, CPU only and has similarly low memory use? Just 2 cores, so 2 tasks if it's one per core, and SETI@home tended to use about 75 Mb RAM for 2 tasks, Milkyway@home only uses about 10 Mb for Separation, which is just awesome. The others that I tried of the quite few projects from those categories just crippled my computer because of their use of my low amount of RAM, hundreds of MB per task, could hardly crunch while using the computer.

Why not Nbody? You can make that run on 2 cores. I've got it on 4-core machines using only two cores while the other two run other projects.

In addition to some projects listed in previous posts, your CPU is 64-bit capable so an upgrade to a 64-bit OS is possible (you likely already know this but still).
I downloaded some Asteroid@home tasks and they were using very little memory in the short time I ran them (less than 10 MB of memory per task).

Sure it is, but why use 64 bit with 4 GB RAM and no dedicated graphics? Just means increased use of already limited RAM, so crippling performance further. And reinstalling Win 7 now would be a terrible pain.

Asteroids... Seem to have considered it after SETI@home ended, but never actually tried it, not entirely sure why, but since I see that the server was out for over a year, maybe it was already out then? And a glance on the forums now says that even after the project came back up, work is intermittent. So can't be a full replacement.

The problem is as the datasets get bigger there's no room in memory for the numbers required in a 32bit app and still keep it anything close to efficient. Sure they can make one but if it takes 10 days to run on a 32 bit machine and only 1 day on a 64 bit machine that's not really helping anyone. I don't know your budget but Dell often has laptops on sale, I got a gently used one for about $300US that had an I7 cpu in it with 32gb of ram and a gpu that works fine but isn't powerful enough to crunch with, I don't use it for Boinc anyway so it was a great find for me.

I'm moving back to Einstein@home on a couple boxes but will leave one milkyway@home dedicated box running.

A lot will be doing this, as I think it's the most similar project. They're not going to know what hit them. They have 6 months of work queued, that's going to drop very quickly.

Maybe universe@home

Why doesn't n-body output anything except

<search_application> milkyway_nbody 1.76 Windows x86_64 double  OpenMP, Crlibm </search_application>
Using OpenMP 8 max threads on a system with 8 processors

in stderr.txt?

What great news! This is exactly what we have all been working towards! Is there any opportunity to port the n-body work over to Macs now? I would love to continue to support this project with my CPUs.

If you insist on using those overpriced machines, can't you run an emulator to do windows tasks? I emulate linux to run CPDN on my windows machines.

Just please don't do it like SETI with sending out multiple copies of same task and let them run even if the result has been successfully crunched by another computer and waste CPU/GPU cycles for nothing. Make it a nice and clean shutdown, preferably simply stop creating new WUs when you don't need anything more and let BOINC do the rest until all tasks are returned. It might take a bit longer than other ways to do it, but I think it's the best way to do it.

Indeed. Nobody wants to waste time (and money if you pay for electricity) doing something pointless when they could be helping another project. Boinc users are used to swapping between projects for downtime or just to share their power among them, or because they have more than one interest. Only give out what is useful to science, that's what we're here for.

several projects "use" DP, like Einstein and Asteroids. but it's not a large portion of the total computation time like Milkyway Separation is/was. so even though a P100 has decent FP64 performance, it's relatively low FP32 performance makes it fall behind in these hybrid FP32/FP64 situations. far more worth it to go with a card with better FP32 since that's where most of the time is spent.

Indeed. I run folding on two types of cards, One is 4000SP/1000DP, the other is 8000SP/500DP. So it's easy for me to see which a project is using. One is twice as fast on DP projects, the other is twice as fast on SP projects. Folding is SP. The small amount of DP can fit into any card. Well it can on a 1:16, not sure about the Nvidias with 1:a billion DP.

Your pc is hidden so it's hard to be exact but something like this hsoul work for you:
<app_config>

<app_version>
<app_name>milkyway_nbody</app_name>

<plan_class>mt</plan_class>
<avg_ncpus>2</avg_ncpus>
<cmdline>--nthreads 2</cmdline>
</app_version>
<project_max_concurrent>3</project_max_concurrent>

</app_config>

Copy that into Notepad in windows and put that in your, in Windows, c:\program data\boinc\projects\milyway.cs.rpi..edu_milkyway folder and save it is app_config.xml. Be sure it saves it exactly as I wrote and NOT with a ',txt' on the end of it.

Then go into the Boinc Manager and click on options, read config files and then you can get the NBody tasks using 2 cpu cores per task and only running 3 tasks at one time on your pc. If you need to change that just be sure to use Notepad in Windows as a word processing program adds hidden stuff that Boinc can't read.

Asking for anymore help should be done in the Number Crunching thread so we don't get bogged down here.

It works fine without any app config. Just not perfectly efficiently.

I have a MacBook Air.

I wasn't directly asking for help on this thread more a general point of when and how when the time comes.

You're going to have to emulate Windows or Linux. There is no Mac Nbody.

If I exclude math projects that I don't have much interest in, the number of CPU projects aren't that abundant these days either, so I am glad there will still meaningful work to do here.

CPUs will run all sorts of useful stuff. GPUs not doing maths is basically now Einstein or Folding. And GPUGrid with a Nvidia.

My CPUs do these non-maths projects:

Physics:

Asteroids
Cosmology
LHC
Universe

Biology:

Sidock
TN Grid
World Community Grid

Nice to see that all the donated computation time did help to come to an end of the research project....

Like others here, I was just about to get a lot of FP64 capable GPUs (I like the thought of putting these otherwise uselss GPUS to a meaningfull purpose)..

Is there really no other GPU FP64 project out there?

Since they sell professional cards with DP, something must be using it. So why isn't it on Boinc?

Asteroids... Seem to have considered it after SETI@home ended, but never actually tried it, not entirely sure why, but since I see that the server was out for over a year, maybe it was already out then? And a glance on the forums now says that even after the project came back up, work is intermittent. So can't be a full replacement.

There is work most of the time. Outages are only about a day, so just set a 3 day queue and you'll crunch all the time.

"AGL said "sustained periods" of higher wholesale electricity prices mean underlying earnings for FY24 are forecast to surge to $1.875b and $2.175b and profits more than double from $580m to $780m."

I'm not a treehugger, but wind. Put wind farms everywhere. It's £0.04 per kWh! Big interconnects between countries, it's always windy somewhere. No fuel to pay for once they're built!

The problem is as the datasets get bigger there's no room in memory for the numbers required in a 32bit app and still keep it anything close to efficient. Sure they can make one but if it takes 10 days to run on a 32 bit machine and only 1 day on a 64 bit machine that's not really helping anyone. I don't know your budget but Dell often has laptops on sale, I got a gently used one for about $300US that had an I7 cpu in it with 32gb of ram and a gpu that works fine but isn't powerful enough to crunch with, I don't use it for Boinc anyway so it was a great find for me.

2nd hand computers are most cheaply bought privately from Gumtree, Ebay, etc. Sometimes even Freecycle! I have a 4 core laptop from freecycle. The disk was broken, so I shoved in a 2nd hand SSD, and upgraded the RAM to the 8GB max. It runs 64 bit stuff nicely.

Maybe universe@home

Not for their GPUs they won't. It's GPU work which has gone from here.

The problem is as the datasets get bigger there's no room in memory for the numbers required in a 32bit app and still keep it anything close to efficient. Sure they can make one but if it takes 10 days to run on a 32 bit machine and only 1 day on a 64 bit machine that's not really helping anyone. I don't know your budget but Dell often has laptops on sale, I got a gently used one for about $300US that had an I7 cpu in it with 32gb of ram and a gpu that works fine but isn't powerful enough to crunch with, I don't use it for Boinc anyway so it was a great find for me.

2nd hand computers are most cheaply bought privately from Gumtree, Ebay, etc. Sometimes even Freecycle! I have a 4 core laptop from freecycle. The disk was broken, so I shoved in a 2nd hand SSD, and upgraded the RAM to the 8GB max. It runs 64 bit stuff nicely.

Eh, have this computer for over 8 years now, been meaning to upgrade for about half that time (not counting the fact that I initially only meant to use it for one year, then turn it into my father's work computer and get a much better one right away). Budget for quite what I'd want never was there, but there's enough for something decent, and has been for 2-3 years now, but the problem is the OS. I will not use any Windows past 7, period, and my two attempts on Linux so far made me say I can't do that. Not because of Linux itself, but because of some other software that I've been using for decades not having Linux versions, and for the lack of the kind of security software of the kind I use on Windows, in particular application-level outbound firewall that prompts about everything not matching rules (and don't say it's not needed on Linux, I want it for control and monitoring of the system, not really for safety in itself - I mean, if I'd be so concerned about safety in itself I wouldn't still be on Win 7 long after EoS now). So sticking to this to be able to stick to what seems to be the last OS I'll ever feel comfortable using for however long it'll be until it'll fall apart or basically nothing will still work on it anymore I guess. Budget has too little to do with it.

Thanks for the info about Asteroids. Will need to attach to it soon and check it out.

And sorry for the off topic. Should probably end this discussion here.

This page might have saved you alot of spreadsheet time https://en.wikipedia.org/wiki/List_of_Nvidia_graphics_processing_units#Compute_capability_table
scroll down to see the FP64/FP32 ratio's

Thanx, I now the chart as and have taken it as a starting point as well as other sources (techpowerup, Reddit, etc..) to make my table. I added stuff like efficiency, numer of GPus on the card, prices, etc... This way, I can quickly decide how much power I can get per dollar or watt out of a certain card for a specific task in different applications. The table includes stuff like CUDA Version, Compute Shader type, etc, which come in handy for inference tasks, general computation, etc...