i feel so lucky, i just get 6 new WU!!! Enough for less than a minute.... but is a progress. 6 is better than nothing.

I am amazed a Linux machine can process 6 work units in under a minute. Did I read that correctly?

Depends how many GPUs you have, and what type. High end cards can do a WU in 30sec or so.

System I am referring to has a RTX 2070 running Linux. I have an RTX 2070 and tasks are currently taking between 2.5 and 3 minutes running Windows 10 Not sure if it makes a difference but I am using AMD 3900X

---

System I am referring to has a RTX 2070 running Linux. I have an RTX 2070 and tasks are currently taking between 2.5 and 3 minutes running Windows 10 Not sure if it makes a difference but I am using AMD 3900X

With the Special Application my RTX 2060 Super was doing WUs in around 50sec (Arecibo VLARs in around 75 sec).

---

Grant
Darwin NT

System I am referring to has a RTX 2070 running Linux. I have an RTX 2070 and tasks are currently taking between 2.5 and 3 minutes running Windows 10 Not sure if it makes a difference but I am using AMD 3900X

With the Special Application my RTX 2060 Super was doing WUs in around 50sec (Arecibo VLARs in around 75 sec).

Man that's fast. I only wish such a thing could be done on Windows. I am aware why Windows is slower

---

Man that's fast. I only wish such a thing could be done on Windows. I am aware why Windows is slower

It could, but there was no one around able to do so (Jason G left before they had the Special App sorted out, and no one came along able to take his place).

---

Grant
Darwin NT

Man that's fast. I only wish such a thing could be done on Windows. I am aware why Windows is slower

It could, but there was no one around able to do so (Jason G left before they had the Special App sorted out, and no one came along able to take his place).

Interesting I thought it was to do with the way Linux was written. I am not wanting to start a debate about that in this thread

---

Unless the above can be done remotely it will not be getting done until Berkely is out of lockdown.

Anything but directly messing up with the hardware can be done remotely.

I can do pretty much anything but running games with my computer remotely and this is a desktop pc, not a server. I can even run graphical applications over internet.

I once tried out of curiosity if I could open a web browser running on my home computer in Finland when I was in Saudi Arabia. This was long ago and back then Saudi Arabian network connectivity went through satellite, so there was a huge latency. The browser was Netscape Navigator. Opening the window took several hours but when it had finally opened, it was actually usable. Just a bit laggy. I could browse the web and because the browser process was actually running in Finland and only its X11 window was on my local computer in Saudi Arabia, I wasn't affected by Saudi web censorship ;)

Interesting I thought it was to do with the way Linux was written. I am not wanting to start a debate about that in this thread

Linux is more efficient than Windows but that difference fits in a single digit percentage. You just need a bit more expensive GPUs to do the same on Windows.

Linux is more efficient than Windows but that difference fits in a single digit percentage. You just need a bit more expensive GPUs to do the same on Windows.

Windows 10 Outperforming Linux On A ~$5000 Laptop
Regardless of OS, it generally comes down to the application- how well does it take advantage of the OS and deal with it's limitations, along with the hardware it is supporting & how well it takes advantage of it & deals with it's limitations.

---

Grant
Darwin NT

Linux is more efficient than Windows but that difference fits in a single digit percentage. You just need a bit more expensive GPUs to do the same on Windows.

Windows 10 Outperforming Linux On A ~$5000 Laptop
Regardless of OS, it generally comes down to the application- how well does it take advantage of the OS and deal with it's limitations, along with the hardware it is supporting & how well it takes advantage of it & deals with it's limitations.

You just need to find a guy like petri on the windows universe and let him do the magic.
I agree find one like him, is not an easy task. But find an ET signal is not an easy task too.

---

. . Mostly only getting work from E@H, but since stupid BOINC only polls SETI for work every few hours those machines are unlikely to get much or anything from SETI unless they get very, very lucky. So I tried to look at my machine specs and see how many outstanding tasks I have waiting on wingmen, the database is reallly, reaaalllyy slow so I was unable to find out. I was curious as to how many might have a chance of clearing in the next 48 to 72 hours. I guess i will just have to wait and see. :(

Stephen

:(

Nothing for several hours, then a few WUs on a couple of requests, then back to nothing for several hours...
I've moved one system to Rosetta and will keep the other just for Seti & will retire it from BOINC once the last WU for Seti is done (or they shut up shop, which ever comes first).

---

Grant
Darwin NT

Linux is more efficient than Windows but that difference fits in a single digit percentage. You just need a bit more expensive GPUs to do the same on Windows.

Windows 10 Outperforming Linux On A ~$5000 Laptop
Regardless of OS, it generally comes down to the application- how well does it take advantage of the OS and deal with it's limitations, along with the hardware it is supporting & how well it takes advantage of it & deals with it's limitations.

A lot of people don't know this, but the actual Windows Kernel, since NT4.0, was built on top of Microsoft's Xenix Kernel, so at the very core, even windows is Unix. The machine calls are nearly the same on every modern OS today. It's the wrappers around it all that slow it down. In fact, the account "Local Machine" is actually the root account.

. . Off topic but to me poignant ...

. . Here is OZ, the ABC network just aired a program called "The Farthest" made in 2017 with NASA, JPL et al. It covers the Voyager missions, 1 and 2, and the final departure from our solar system of Voyager 1 in 2012 travelling much faster and launched much earlier than Voyager 2 (because 1 continued on its original trajectory while VGR2 turned 90 deg and headed out at right angles to the plane of our solar system). VGR1's launch in 1977 helped to accentuate the time scale of such projects (the Cosmos itself if you like) and increased my sadness that this part of Seti@Home is ending. But it's been nice playing a small part in the process. And the search goes on, bring on Breakthrough Starshot.

Stephen

:(

. . Off topic but to me poignant ...

. . Here is OZ, the ABC network just aired a program called "The Farthest" made in 2017 with NASA, JPL et al. It covers the Voyager missions, 1 and 2, and the final departure from our solar system of Voyager 1 in 2012 travelling much faster and launched much earlier than Voyager 2 (because 1 continued on its original trajectory while VGR2 turned 90 deg and headed out at right angles to the plane of our solar system). VGR1's launch in 1977 helped to accentuate the time scale of such projects (the Cosmos itself if you like) and increased my sadness that this part of Seti@Home is ending. But it's been nice playing a small part in the process. And the search goes on, bring on Breakthrough Starshot.

Stephen

:(

What I'm trying to wrap my head around, is the fact the dataset they're going to be looking at is more than just a look in space, but also time. To try to extrapolate possibilities for signals from millions, perhaps billions of years ago, and try to pinpoint where, and when, the probability of such signals would have existed. Somehow, I get the feeling we will be at this again with sensors which will be looking at different points in time, from different vantage points from locations throughout the solar system and galaxy at other points in the future because we are inevitably missing a lot of points, even with the 20 years we've put into this so far. It's less than a blip in time which we've explored.

Windows 10 Outperforming Linux On A ~$5000 Laptop

Using a brand new top end laptop is a bad way to compare the performances of different operating systems. What you really end up comparing is how fast the support for the latest proprietary hardware quirks gets added to different operating systems. Laptop power saving features are a notoriously fast moving target.

Commercial actors like Microsoft have an edge here because hey have the money to make exclusive deals with the hardware manufacturers and are also aggressively doing that to ensure the information needed to properly support the hardware won't 'leak' to competitors. So the linux kernel coders have to obtain the information by reverse engineering and this is slow.

A lot of people don't know this, but the actual Windows Kernel, since NT4.0, was built on top of Microsoft's Xenix Kernel, so at the very core, even windows is Unix.

This is in disgreement with what I heard back when this actually happened: Microsoft hired Digital to write a kernel for them and it was done by modifying the VMS kernel. So NT wouldn't be a Unix but a VMS.

A lot of people don't know this, but the actual Windows Kernel, since NT4.0, was built on top of Microsoft's Xenix Kernel, so at the very core, even windows is Unix.

This is in disgreement with what I heard back when this actually happened: Microsoft hired Digital to write a kernel for them and it was done by modifying the VMS kernel. So NT wouldn't be a Unix but a VMS.

I've confirmed it inside MS. It is the sole reason Microsoft paid SCO $100K to actually license Xenix. They had future options in mind.

Edit: It was also the source of a few hacks which Microsoft didn't quite prepare for. That Local Machine account was not nearly as guarded as it is today.

Edit2: Which version? NT4 was a lot different that 3.5 and 3.51. That was their biggest leap.

What I'm trying to wrap my head around, is the fact the dataset they're going to be looking at is more than just a look in space, but also time. To try to extrapolate possibilities for signals from millions, perhaps billions of years ago, and try to pinpoint where, and when, the probability of such signals would have existed. Somehow, I get the feeling we will be at this again with sensors which will be looking at different points in time, from different vantage points from locations throughout the solar system and galaxy at other points in the future because we are inevitably missing a lot of points, even with the 20 years we've put into this so far. It's less than a blip in time which we've explored.

. . My personal analogy is like a person coming out of a coma for a moment, opening their eyes and looking at the part of the room in front of them for a few seconds then lapsing back into a coma. It could be highly significant to the person for those few seconds but it won't reveal a helluva lot. But at least we have had a peep.

Stephen

< shrug >

Seems like a really large lag in deleting AstroPulse validated work units. I've never held this many tasks there. Guess I should have expected it with the large number waiting to purge from the SETI@Home climbing so high recently.

Also trickling in GPU tasks, no more than two at a time seemingly, and that seems to be either AP or SETI tasks, haven't seen any mixing happening.