Strange that Jose and Seth seem to be the only ones reporting problems with 5.16..... We have 20 boxes from low end Celerons and Durons to AMD X2s running 5.16 and have had no instability problems whatever. We also do NOT run the screensaver as a matter of principle as it eats CPU cycles that can be better used by the science app. We use the 'Blank' screen provided by WinXP.
I'm also very impressed by the patience and the effort shown by Moderator 9 to solve this problem, especially since it seems to be very limited in scope. IMHO, no other Project would work this hard to solve such an isolated glitch and the effort expended should be more appreciated by those affected....Cheers, Rog.

---

Strange that Jose and Seth seem to be the only ones reporting problems with 5.16..... We have 20 boxes from low end Celerons and Durons to AMD X2s and have had no instability problems whatever. We also do NOT run the screensaver as a matter of principle as it eats CPU cycles that can be better used by the science app. We use the 'Blank' screen provided by WinXP.
I'm also very impressed by the patience and the effort shown by Moderator 9 to solve this problem, especially since it seems to be very limited in scope. IMHO, no other Project would work this hard to solve such an isolated problem and the effort expended should be more appreciated by those affected....Cheers, Rog.

We are not the only ones having it. May be the only ones reporting them, but not the only ones having them.

To set the record straight: I like Moderator 9...I think "NUMERO Nueve" knows that many of the comments I make about human sacrifice and voodoo are a joke ( hey I do have a weird sense of humor). "Numero Nueve's"sense of humor is refined and a little nutty too as evidenced by Nueve's comments made about BBQ's and going back to the sacrificial pool .


I do appreciate the time that is being taken to solve "MY"issues and others. That is why even though I keep saying I will detach, I haven't.

---

This and no other is the root from which a Tyrant springs; when he first appears he is a protector.”
Plato

With Rosetta 5.16 (and earlier) I see at the begin of the "relax"-phase the text "Accepted Energy" written into the low-energy-frame. The text is streched during the first steps until it is moved outside of the window. Hopefully the clipping of the graphics routines prevent bad sideeffects.

Norbert

---

Checked the DEP setting and it is OK. I've crunched 24 of the 5.16 units on 2 different systems and the only error I got was the one time I turned on the graphics. I do not use screensavers of any kind and set monitors to shut off after 30 min. I got the Windows error and reporting box
on that one error.

https://boinc.bakerlab.org/rosetta/forum_thread.php?id=1574#16401
Jose: how about running this, as I suggested in the 5.13 thread, and post your hijackthis! log so we can see everything that's actually running on your system; and we (programmers, comp techs, etc) can identify anything that may be conflicting with Rosetta.

---

If your system is set as shown in the picture, then I am wrong and I need to keep looking. If however the button for "Turn DEP on for ALL programs and services ..." is selected, then I think you should change that setting to look like the picture, and try running Rosetta that way.

I am going to go contrary to your advice. I am going to run Windows with the "Turn DEP on for ALL programs and services ..." selected. " and trust my Anti virus and my maintenance utilities ( or the guy that repairs my computer) to repair any damage.

My reasoning: For reasons only known to the Bill Gates, Windows may be deciding a program I am running should be blocked and when I run it along with the ever present Rosetta...the whole mess starts. See, I just noticed the comment in the last error message that I was running a Word 2003 document. I have checked my "records" and in at least 5 occasions I got computing errors, I had that document open. ( I am proof reading/editing and making comments to the computer repair's kid thesis). So lets see what happens now when I run that document again...without allowing Windows decide it may be a threat.

Argh watch my computer crash!!!!!

You may well be on to something. It is possible the there is a conflict between Word and Windows (won't be the first time), I have seen the word documents themselves create problems. In short your test is a valid approach.

I have tried to recreate the problem on a system I have available that is very similar to yours. Here are the specs.

Mine
CPU Type - GenuineIntel
Intel(R) Pentium(R) M processor 1.86GHz
CPUs 1
Operating System
Microsoft Windows XP
Professional Edition, Service Pack 2, (05.01.2600.00) (05.01.2600.00)
Memory - 2039.37 MB
Cache - 976.56KB
Floating point - 1643.58 MO/sec
Integer speed - 3401.25 MO/Sec

Jose
GenuineIntel
Intel(R) Pentium(R) 4 processor 2.26GHz
CPUs 1
Operating System
Microsoft Windows XP
Professional Edition, Service Pack 2,
Memory - 991.23 MB
Cache - 976.56KB
Floating point - 1155.89 MO/Sec
Integer speed - 2261.9 MO/Sec

Both system are running BOINC 5.4.9 and Rosetta 5.16. Mine is a laptop and I think Jose mentioned somewhere his is a desktop.

My test system is also running RALPH and is set to go to screen saver mode in 2 min of inactivity. I have never had a windows error message "pop up" appear, and I have only seen three errors. All three were in Ralph, all three were file errors related to a bad group of work units that erred on everyones machines. None of these errors halted the machine, it just moved on to the next work unit. Try as I might I have been unable to MAKE my machine produce any -107 Access violations.

I can even digitize video on this system, use "Go-to-my-PC" and run Rosett/RALPH and it does not drop frames, or even seem to care about all this heavy activity. But more important it does not fail any workunits.

However, there are some interesting differences between these two systems that may or may not be relevant here, but they do raise a question. Why is Jose's machine, that is clocked so much faster than mine, benchmarked so much slower? The answer (at least in part) has to be there is more "stuff" running on Jose's system when the benchmarks are taken. The memory is certainly part of the answer, but not that big a part.

IF this was a Mac, I would tell you to try turning off all the extensions on restart, and run BOINC in a clean environment and see if you get any errors. This would work on a Mac, but I do not think BOINC will run on a PC in safe mode, and I can't test this because I am not where the PC is right now. (any takers, out there with errors want to try it?).

If BOINC would run in safe mode, then we could safely say that Rosetta is having a problem with one of the background applications on Jose's system. It certainly runs well on mine and I have a very limited set of background functions running.

As a personal aside, I am afraid as Jose suspects I have a very active sense of humor, and he manages to tickle it often. As hard as I try to keep all of this as professional as possible, at times I just can't help myself. That is the source of the banter between some of you, like Jose, and I. In Jose's case it was obvious very early on that he was frustrated by the problem (as are we), but still not really upset at anyone here.

I think he is correct that there are others having similar errors who we have not heard from. But the projects current error rate is VERY low. So I would think that less than 50 systems are seeing this problem, and most are not hanging, they fail the work unit and proceed to the next. Most of the time the users may not even be aware they have had an error.

---

Moderator9
ROSETTA@home FAQ
Moderator Contact

With Rosetta 5.16 (and earlier) I see at the begin of the "relax"-phase the text "Accepted Energy" written into the low-energy-frame. The text is streched during the first steps until it is moved outside of the window. Hopefully the clipping of the graphics routines prevent bad sideeffects.

Norbert

What type of Work Unit is running just ahead of the CASP work. If it is a work unit that has a known structure, you could just be seeing image retention on the monitor.

---

Moderator9
ROSETTA@home FAQ
Moderator Contact

Mod9:
CPU Type - GenuineIntel
Intel(R) Pentium(R) M processor 1.86GHz
CPUs 1
Operating System
Microsoft Windows XP
Professional Edition, Service Pack 2, (05.01.2600.00) (05.01.2600.00)
Memory - 2039.37 MB
Cache - 976.56KB
Floating point - 1643.58 MO/sec
Integer speed - 3401.25 MO/Sec

Jose:
GenuineIntel
Intel(R) Pentium(R) 4 processor 2.26GHz
CPUs 1
Operating System
Microsoft Windows XP
Professional Edition, Service Pack 2,
Memory - 991.23 MB
Cache - 976.56KB
Floating point - 1155.89 MO/Sec
Integer speed - 2261.9 MO/Sec

The Pentium M is based on the Pentium III core - and has a much higher Instruction per clock cycle than the Pentium IV core which requires high Ghz speeds to perform the same amount of work as a lower clocked Pentium III/Pentium M, or Athlon XP/Athlon 64 cpu.

Remind me to dig out a few of the wonderful articles on places like Anandtech on the various cpu cores and how the length of the pipelines affect how much gets done per clock cycle. Even places like Tom's Hardware have posted articles about how the Pentium III core was back (as the Pentium M) and mentioned the benefits that created for Intel.

So the question is not whether Mod9's stats match Jose's stats, but whether Jose's benchmarks match a similarly clocked Pentium IV system that uses the default Boinc client.

---

Win 98SE system with 256 MB memory frequently processes WUs and shows no CPU time after 8 hours. eg.

https://boinc.bakerlab.org/rosetta/result.php?resultid=20659581

My other systems are all XP and don't exhibit this problem.

---

Win 98SE system with 256 MB memory frequently processes WUs and shows no CPU time after 8 hours. eg.

https://boinc.bakerlab.org/rosetta/result.php?resultid=20659581

My other systems are all XP and don't exhibit this problem.

This is a known issue with Win 98. They all do it. The programers are trying to find a way around the issue.

---

Moderator9
ROSETTA@home FAQ
Moderator Contact

Win 98SE system with 256 MB memory frequently processes WUs and shows no CPU time after 8 hours. eg.

https://boinc.bakerlab.org/rosetta/result.php?resultid=20659581

My other systems are all XP and don't exhibit this problem.

This is a known issue with Win 98. They all do it. The programers are trying to find a way around the issue.

Is there a list of known issues anywhere in the Rosetta FAQ?
-Seth

---

[color=darkred][b]Rosetta Version 5.16 has been released. Please report any problems in this thread.

The servers may be slow until the new application is distributed.

Version 5.16 has the following features;

I am still using the old one but i can not connect message board or webpage from the GUI in linux.

---

... So the question is not whether Mod9's stats match Jose's stats, but whether Jose's benchmarks match a similarly clocked Pentium IV system that uses the default Boinc client.

I agree with everything you wrote, but I have to work with the systems I have, and that one is as close to Jose's set up as I can get.

The programers are looking for a memory leak as the possible cause as well, so something should turn up soon.

---

Moderator9
ROSETTA@home FAQ
Moderator Contact

... So the question is not whether Mod9's stats match Jose's stats, but whether Jose's benchmarks match a similarly clocked Pentium IV system that uses the default Boinc client.

I agree with everything you wrote, but I have to work with the systems I have, and that one is as close to Jose's set up as I can get.

The programmers are looking for a memory leak as the possible cause as well, so something should turn up soon.

During the last 2 hours I submitted Boinc, Rosetta and even my computer to every torture I could think off. This included a moment in which the BOINC application, left my computer to be restored via the restore point function.
Not a single step was lost. The Wu was there , working as if nothing had happened.

Could this be a sign of good thinks to come?

Or Have I just re jinxed the computer? "Numero 9" do you know any good computer-performance restoring voodoo spells?

---

This and no other is the root from which a Tyrant springs; when he first appears he is a protector.”
Plato

With Rosetta 5.16 (and earlier) I see at the begin of the "relax"-phase the text "Accepted Energy" written into the low-energy-frame. The text is streched during the first steps until it is moved outside of the window. Hopefully the clipping of the graphics routines prevent bad sideeffects.

Norbert

What type of Work Unit is running just ahead of the CASP work. If it is a work unit that has a known structure, you could just be seeing image retention on the monitor.

I'm running only CASP-WUs at the moment. And it happens at the 2nd etc. model too.

Norbert

---

With Rosetta 5.16 (and earlier) I see at the begin of the "relax"-phase the text "Accepted Energy" written into the low-energy-frame. The text is streched during the first steps until it is moved outside of the window. Hopefully the clipping of the graphics routines prevent bad sideeffects.

Norbert

What type of Work Unit is running just ahead of the CASP work. If it is a work unit that has a known structure, you could just be seeing image retention on the monitor.

I'm running only CASP-WUs at the moment. And it happens at the 2nd etc. model too.

Norbert

Right, it happens for CASP WUs (probably due to how the boxes are arranged, as there is no "native" box anymore) I reported it a few posts earlier.

---

Best UFO Resources
Wikipedia R@h
How-To: Join Distributed Computing projects that benefit humanity

...During the last 2 hours I submitted Boinc, Rosetta and even my computer to every torture I could think off. This included a moment in which the BOINC application, left my computer to be restored via the restore point function.
Not a single step was lost. The Wu was there , working as if nothing had happened.

Could this be a sign of good thinks to come?

Or Have I just re jinxed the computer? "Numero 9" do you know any good computer-performance restoring voodoo spells?

Well... I am guardedly optimistic at this report. However if you fixed it could you please tell us what you did? It is just remotely possible that you have found an answer for more than just your system if it is working now.

I have been thinking about the DEP, and what you said you were going to try. In essence what you were proposing was to force windows to handle any errors that came up. Assuming windows is smart enough to know what to do (am assumption I would never make) it just might work. The Work Unit might still be terminated by windows, but it "should" be a soft landing.

With the current Rosetta application, as long as the computer does not crash, and it processes some of the workunits, or at least finishes some models, then there is no harm from any failed work units.

Although, you will understand if I just tread water here in the pool for a while to see what happens.

---

Moderator9
ROSETTA@home FAQ
Moderator Contact

With Rosetta 5.16 (and earlier) I see at the begin of the "relax"-phase the text "Accepted Energy" written into the low-energy-frame. The text is streched during the first steps until it is moved outside of the window. Hopefully the clipping of the graphics routines prevent bad sideeffects.

Norbert

What type of Work Unit is running just ahead of the CASP work. If it is a work unit that has a known structure, you could just be seeing image retention on the monitor.

I'm running only CASP-WUs at the moment. And it happens at the 2nd etc. model too.

Norbert

Right, it happens for CASP WUs (probably due to how the boxes are arranged, as there is no "native" box anymore) I reported it a few posts earlier.

Yes but the CASP work units are not supposed to have a native box. The normal work units still do. The project is running both kinds, and the graphics are SUPPOSED to be different. So if you catch it graphic just after it changes from a normal work unit (with a native box and RMDS graph), to a CASP work unit (without these two features), and particularly if you are using a "Glass" CRT monitor, you could see some after image as he describes.

---

Moderator9
ROSETTA@home FAQ
Moderator Contact

With Rosetta 5.16 (and earlier) I see at the begin of the "relax"-phase the text "Accepted Energy" written into the low-energy-frame. The text is streched during the first steps until it is moved outside of the window. Hopefully the clipping of the graphics routines prevent bad sideeffects.

Norbert

What type of Work Unit is running just ahead of the CASP work. If it is a work unit that has a known structure, you could just be seeing image retention on the monitor.

I'm running only CASP-WUs at the moment. And it happens at the 2nd etc. model too.

Norbert

Right, it happens for CASP WUs (probably due to how the boxes are arranged, as there is no "native" box anymore) I reported it a few posts earlier.

Yes but the CASP work units are not supposed to have a native box. The normal work units still do. The project is running both kinds, and the graphics are SUPPOSED to be different. So if you catch it graphic just after it changes from a normal work unit (with a native box and RMDS graph), to a CASP work unit (without these two features), and particularly if you are using a "Glass" CRT monitor, you could see some after image as he describes.

No no... you guys don't understand. You display the graphic. It shows the word "Low Energy" in the righthand box. I also think I've seen it say "Accepted Energy"... and then each time the model hits a lower energy and it draws a new graphic there... the words stretch longer... and longer and off the righthand side of the box. The font goes stretched one further each time it redraws the low energy graphic. I think it is the WUs that say "accepted energy" in that box that have the problem. I've got two now that say "Low Energy" and they don't seem to be doing it.

In my case, whenever I have the graphic up, it's pretty much a given that I'm going to grab that bugger and spin it around some. Perhaps that has something to do with it?

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/

Here, I just caught one on Ralph. These two snaps were taken within 10 seconds of each other:




The words get stretched out lower and longer to the point that it just becomes a straight line.

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/