Jack,

I'm a long time Open GL hack, four years working on Open GL drivers in the late 90's and early 2000's. Is there any help whatsoever that I can offer in the quest to resolve the glut32.dll issue? Meaning are there functions in glut32 that you're calling that could be handed to me for local implementation (or something).

Thanks for the offer! I think we've fixed the problem by folding the font drawing code into our code, the same way seti@home did. So we won't have to distributed glut32.dll anymore.

Well, I got another graphics WU (hasn't started yet), and just in case, I put a copy of glut32.dll in my system folder. It can't hurt to have it anyway.

---

yippie! got 3 graphical wu'S in a row :)

---

--
Florian
www.domplatz1.de

Hey! I'd better take a break and see what it looks like. I am running one.

---

<cry> <sob> 2 graphics WU have passed me by in the night. Isn't there an option somewhere in Boinc, 'No graphics at night'? Not to worry, I see 3 more in the queue. Seems like a new flock has been released :-)

---

BOINC.BE: For Belgians who love the smell of glowing red cpu's in the morning
Tutta55's Lair

Yes, I finally got one, so I suspended the other WU's I had, and looked at it. It was funny to watch, it executed as it should and uploaded. But I missed the last minutes of it, so I don't know if it ended up folding like the natural?

---

[b]"I'm trying to maintain a shred of dignity in this world." - Me[/b]

Yeah, I got about 2 days worth of graphics. They look really cool so far.

---

I got a couple. And even got a chance to watch one late in the game ... mostly it was vibrating ...

But, it would probably be more interesting when I can see more so I can get a better "feel" for what I should be seeing.

Still, it was interesting and encouraging to see ...

---

But, it would probably be more interesting when I can see more so I can get a better "feel" for what I should be seeing.

To a new user with no idea what Rosetta is about, it'll be "pretty" - but then, that's about all you can say for the SETI screensaver too, and IMHO, this one is "prettier". As someone learns more about Rosetta, they'll get more out of the screensaver. As Paul says, it helps to watch several results go through, too.

I would like it if the bottom-right corner wasn't "refreshed" when going from ab-initio to relaxation. I'm never sure if the lowest-left point was picked or not. It'd be nice to see that the program tried a whole bunch of random "spots", then picked the "best" one, and is trying to "improve" on that one. I understand it may be a scale-change issue, but even if only the tries that "fit" in the new scale were to remain, I think it'd be easier to grasp what is going on. If it is possible to put in a different-colored "dot" for the "goal", that would be great too! Right now, the only way to know what the whole idea is, is to have looked around the web site and seen the user results graphs. Without that, it's just a bunch of lines jumping around...

---

To a new user with no idea what Rosetta is about, it'll be "pretty" - but then, that's about all you can say for the SETI screensaver too, and IMHO, this one is "prettier". As someone learns more about Rosetta, they'll get more out of the screensaver. As Paul says, it helps to watch several results go through, too.

Cool to hear you think it's prettier. My goal has been to balance aesthetics vs. putting as much of the inner workings of Rosetta on the screen as possible. So I hope it works for both audiences, as you suggest.

I would like it if the bottom-right corner wasn't "refreshed" when going from ab-initio to relaxation. I'm never sure if the lowest-left point was picked or not. It'd be nice to see that the program tried a whole bunch of random "spots", then picked the "best" one, and is trying to "improve" on that one. I understand it may be a scale-change issue, but even if only the tries that "fit" in the new scale were to remain, I think it'd be easier to grasp what is going on. If it is possible to put in a different-colored "dot" for the "goal", that would be great too! Right now, the only way to know what the whole idea is, is to have looked around the web site and seen the user results graphs. Without that, it's just a bunch of lines jumping around...

There is, as you say, a scaling issue with showing the points from the ab intio stage on during the relax stage. They use different energy functions. We could report the lowest RMSD structure found during the ab inito stage, so that you can see how close the one actually picked for refinement is to the lowest.

The goal will always be points on the left side of this plot. And preferably in the lower-left corner (low RMSD and low energy.)

There is a page in the works to explain the graphics in more detail.

The goal will always be points on the left side of this plot. And preferably in the lower-left corner (low RMSD and low energy.)

Rather than "goal", I probably should say "Native"... I assume that is the ideal we're trying to match. Speaking of which, the one I've got running right now is actually somewhat "close" in appearance, the first one I've seen that is... come on, blue part, bend left, not right...

---

Rather than "goal", I probably should say "Native"... I assume that is the ideal we're trying to match. Speaking of which, the one I've got running right now is actually somewhat "close" in appearance, the first one I've seen that is... come on, blue part, bend left, not right...

Ah, yes. This would be nice, however it is not currently something that is in Rosetta. It knows the native structure as determined by experiment, but not its energy. It could calculate this energy, but the structure has to be "relaxed" first, using our energy function. This is how the native points are generated on the top predictions and results pages. This extra information would have to be included in the work unit, and read into Rosetta, so it probably won't happen in the short term.

Nevertheless, 0.00 RMSD, all the way on the left, is always the goal.

come on, blue part, bend left, not right...

Jack, you said you wanted "visual" feedback, based on the comment that the human eye can see things that the computer cant.

Bill's comment about the "blue" part being wrong is interesting, because if you look at those two screenshots I took in another thread, in the first of them it looks like it's the blue part that's in the wrong place.

There's other smaller errors (like a minor tweak needed to the actual hairpin between red and orange), but for the most part it's fairly close to the mark. Except for that blue string that's way off in left field.

I might add that this is why I was talking about the ability to rotate the native with the mouse, so that I can see how bad the error really is.

---

come on, blue part, bend left, not right...

Jack, you said you wanted "visual" feedback, based on the comment that the human eye can see things that the computer cant.

Bill's comment about the "blue" part being wrong is interesting, because if you look at those two screenshots I took in another thread, in the first of them it looks like it's the blue part that's in the wrong place.

There's other smaller errors (like a minor tweak needed to the actual hairpin between red and orange), but for the most part it's fairly close to the mark. Except for that blue string that's way off in left field.

I might add that this is why I was talking about the ability to rotate the native with the mouse, so that I can see how bad the error really is.

Yes, rotation would definitely be good. Perhaps in iteration 3 of the screen saver. We are very close to sending out iteration 2.

The most useful information that boinc users can provide at the moment are overall impressions of the trajectories. For example, if you often see something that looks "wrong" forming early and getting stuck. We know that any particular structure is quite likely to have flaws. What's especially difficult for us to analyze at the moment are the details of the trajectories that lead to those flaws.

The most useful information that boinc users can provide at the moment are overall impressions of the trajectories. For example, if you often see something that looks "wrong" forming early and getting stuck. We know that any particular structure is quite likely to have flaws. What's especially difficult for us to analyze at the moment are the details of the trajectories that lead to those flaws.

So, as an example, in this one:



you might care about the fact that the blue strand never seemed to make it to the correct place. Not even after several sets of ab init - instead it always seemed to land about where it shows here, with an RMSD in the 6.0 to 7.0 range

---

well i still havent seen a grapic one yet.. im crunching on as many machines as i can and still none have come through

---

Today I received my first two WU's with graphics. The WU's were distributed to two seperate machines, a n Intel 2.6 ghz celeron, and an Intel Pentium 4 dual core 3.0 ghz. The celeron is displaying the graphics perfectly with no problems. However, the dual core will not display the graphics at all. When the screensaver engages, the screen just goes blank. This happens even when I suspend everything else that is running in BOINC. I cannot start the graphics even if I try to do it manually by clicking on the 'show graphics' button.
...thought you might find that interesting.


Also, on a different topic, can you direct me to a page that will explain the screensaver in some detail, along with some of the science involved, so that I can understand what I am looking at. Thanks

---

Ray

Today I received my first two WU's with graphics. The WU's were distributed to two seperate machines, a n Intel 2.6 ghz celeron, and an Intel Pentium 4 dual core 3.0 ghz. The celeron is displaying the graphics perfectly with no problems. However, the dual core will not display the graphics at all. When the screensaver engages, the screen just goes blank. This happens even when I suspend everything else that is running in BOINC. I cannot start the graphics even if I try to do it manually by clicking on the 'show graphics' button.
...thought you might find that interesting.

Are you running an ATI card on the dual core? Boinc graphics do not run reliably on them. I have a dual core machine at work with an nVidia card in it and the graphics run perfectly (which didn't when I had an ATI)

Also, on a different topic, can you direct me to a page that will explain the screensaver in some detail, along with some of the science involved, so that I can understand what I am looking at. Thanks

They are currently working on a page to explain the graphics, but there is plenty of info in this thread:
https://boinc.bakerlab.org/rosetta/forum_thread.php?id=252

---

Hi genes,
The display is an Intel 82915g/gv/910gl pci Express Chipset family.(that came directly out of the device manager). I have again tried to get the graphics working, but still no go. I am at a loss.
Thanks also for the link to the thread.

---

Ray

got my first grapics unit last night.. looked neet...

im sure there are lots of tweeks and improbments that can be made but it lookg good to me

---