Posts by robertmiles

41) Message boards : Cafe Rosetta : Breakthrough !? A 1000 cores low power processor are invented (Message 80724)
Posted 9 Oct 2016 by Profile robertmiles
Post:
http://www.zdnet.com/article/say-hello-to-the-worlds-first-1000-core-chip-that-runs-off-an-aa-battery/

Cruwl thing, but where can we will get that, have it disc or something else,
never a x86 architecture i think.

Greets, just spreding the news ;)

Looks suitable for running applications that don\'t require enough memory to need any separate memory chips. The clock speed given in that article is over 100 GHz for each core, and if that is accurate, none of today\'s memory chips are fast enough to avoid leaving most of the cores idle most of the time.

But where are you going to find useful applications small enough to run without any separate memory chips?
42) Message boards : Number crunching : GPU computing (Message 80723)
Posted 9 Oct 2016 by Profile robertmiles
Post:

There is really no way that scientists have to wait hours or even days for their computation results. Personally, I hope BOINC will die because it's a kludge.


I hope not.
If it is "kludge", why you are here?


Because I want to help. That's why.

One of the problems is the heterogeneous architecture of rosetta@home: There are PCs, Macs and tablets/smartphones (seriously?). Why not an internet-connected dual-core toaster?

There are a lot of issues with these devices and this is IMHO a waste of developer resources.

A homogeneous architecture based on AVXx would alleviate all those problems while yielding a higher performance.

The distributed nature of rosetta also introduces latencies: preparing work, zipping it, sending it and collect the results back over a WAN. Being forced to deal with ultra-lame ancient CPUs and the like are another problem.

So you want far fewer processors to be used? None of my computers use a CPU that even has AVXx available, and not enough money is available to replace all the computers available through BOINC with equivalents that have AVXx available. It would be possible, though, to produce separate compiles of the application for computers with AVXx and computers without, and add a shell program that tests what the CPU has available, then starts only the version of the program best for the current CPU.
43) Message boards : Number crunching : GPU computing (Message 80722)
Posted 9 Oct 2016 by Profile robertmiles
Post:
I can't fathom the computing knowledge you need for something like Rosetta. Or anything useful for that matter... I just got into learning Python (I figured an EE should know a good bit of programming) and I'm struggling like mad. MATLAB is the only language I'm proficient at, but it's so user friendly it doesn't count IMO.


If i remember correctly, the public test of rosy on gpu was with and old version of pycl

This is the post one developer wrote about this test. It's a pity that pdfs are not longer available

I've used Fortran for several years, and have taken classes in C++ and CUDA since then. Is any help needed for translating any remaining Fortran code to C++?

I would not be able to travel for this.

I'm still looking for an online OpenCL class aimed at GPUs rather than FPGAs. A CUDA version would work on most Nvidia GPUs, but not on other brands. An OpenCL version should work on other brands of GPUs.

A GPU version REQUIRES that most of the application allows many threads to run in any order, or even at the same time, since they don't use anything produced by the other threads. If this is not satisfied, the GPU version may be as slow as a quarter of the speed of the CPU version.
44) Message boards : Number crunching : GPU computing (Message 80721)
Posted 9 Oct 2016 by Profile robertmiles
Post:
I\'ve just been looking at the performance of the new GTX1080 and for DOUBLE precision calculations it does 4 Tflops!!!! For comparison a relatively high performance chip like an overclocked 5820K will do maybe 350GFlops. So we are talking an order of magnitude difference. In addition the Tesla HPC version will probably be double that at 8 TFlops. (Edit: Looks like it is actually 5.3TFlops) The Volta version of the gtx1080 (next gen on, due in about 18 months time) is rumoured to be 7TFlops FP64 in the consumer version.

There is no way that conventional processors can keep up with that level of calculation. At what point does the gap between serial CPU and parallel GPU have to be before the project leaders decide they can not afford NOT to invest in recoding to parallel processing? Because by 2 years time, HPC GPUs will be around 35 times faster than CPUs. How much will it cost to rewrite the code, $100-150K maybe?? Isn\'t that worth paying for such a huge step up?

With that kind of performance increase, you can make calcs more accurate. You no longer have to use approximations like LJ potentials, you can calculate the energy accurately and get a better answer in a quicker time than now. Whats not to like?

It seems like so many projects, everyone is comfortable with what they are doing now. Revolution has been forsaken for evolution. Understandable, but the best way to do things?

Be bold and take the leap!

More computing performance is not a good answer if the limit comes from available memory limits rather than from computing limits. Rosetta@Home has already looked into GPU versions, and found that they would require about 6 GB of graphics memory per GPU to get the expected 10 times as much performance as for the CPU version. The GPU version would run each workunit at about the same speed as the CPU version, and would therefore need to run 10 workunits at the same time, using 10 times as much memory, to get 10 times as much performance.

Rather few of the high-end graphics boards have that much memory.
45) Message boards : Cafe Rosetta : Another suspected cause for Alzheimer's (Message 80600)
Posted 7 Sep 2016 by Profile robertmiles
Post:
Air Pollution Leaves Significant Traces Of Magnetic Metals In Your Brain

air-pollution-magnetic-particles_us
46) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 80587)
Posted 1 Sep 2016 by Profile robertmiles
Post:
This is my second post on this issue. I have two desk top computers. One has 8 processors the other four. For Some reason...you program refuses to send my 4 processor computer any work. So until you can tell me what I\'m doing wrong.. I\'ll just go to another project, which is a shame since I enjoy working with Rosetta@home I\'ll leave it to you to figure out what is going wrong.

John Yeatts


Not enough information to tell much. You need to show more of your settings for those computers, both the settings on the computers and the settings on the server. Be sure to include the amount of memory BOINC is allowed to use on each computer, and the amount of memory that is installed.
47) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 80581)
Posted 31 Aug 2016 by Profile robertmiles
Post:
Hello, is it normal to get a couple of tasks with \"Validate Error\"? Is it an issue on my end? I\'ve had 10 of them in the past 4 days or so.

Here\'s a few:
869154381
868770993
868649273

Thanks.


Not a problem I\'ve seen before, on any project. It could mean a problem with the validator, rather than anything on your computer.
48) Message boards : Cafe Rosetta : A protein that might help stroke recovery (Message 80559)
Posted 24 Aug 2016 by Profile robertmiles
Post:
This article names a protein that might help recovery from strokes. Could someone at Rosetta@home check if there has been enough work to determine the 3D structure of this protein?

Mouse Study Suggests Stem Cells May Reverse Stroke Damage
https://consumer.healthday.com/cardiovascular-health-information-20/misc-stroke-related-heart-news-360/mouse-study-suggests-stem-cells-may-reverse-stroke-damage-714075.html

This still needs tests to determine if it also works in humans.
Looks unlikely to help recover memories lost in strokes, but may recover the ability to form those memories.
49) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 80349)
Posted 13 Jul 2016 by Profile robertmiles
Post:
I have ten completed tasks, with six of them due tomorrow, and some have been sitting in my queue for over 24 hours trying to upload.


Five of the six tasks I earlier had waiting for uploads have now uploaded and automatically reported, so it looks like the problem is now PARTLY fixed.


It looks like the project is trying to persuade users to support other projects instead by not fixing the long delays before clients can contact the server again.
50) Message boards : Cafe Rosetta : An Alzheimer\'s reversal? (Message 80295)
Posted 26 Jun 2016 by Profile robertmiles
Post:
This looks worth investigating to determine how accurate it is.

The first study to ever show Alzheimers can be reversed has just been published
Alzhreverse

http://www.dailykos.com/stories/2016/6/21/1540972/-The-first-study-to-ever-show-Alzheimers-can-be-reversed-has-just-been-published?detail=email&link_id=5&can_id=04cf7618a5486f0ccaca597540d8df5e&source=email-church-whose-pastor-praised-orlando-shootings-is-being-asked-to-leave-by-landlord&email_referrer=church-whose-pastor-praised-orlando-shootings-is-being-asked-to-leave-by-landlord&email_subject=church-whose-pastor-praised-orlando-shootings-is-being-asked-to-leave-by-landlord
51) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 80286)
Posted 24 Jun 2016 by Profile robertmiles
Post:
I have ten completed tasks, with six of them due tomorrow, and some have been sitting in my queue for over 24 hours trying to upload.


Five of the six tasks I earlier had waiting for uploads have now uploaded and automatically reported, so it looks like the problem is now PARTLY fixed.
52) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 80260)
Posted 24 Jun 2016 by Profile robertmiles
Post:
Could something be added to the Server Status to show the amount of free disk space available available for uploading? I\'ve seen some BOINC projects that had to delay uploads when such space ran short.
53) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 80245)
Posted 23 Jun 2016 by Profile robertmiles
Post:
I see a growing number of pending upload jobs in my queues, is there an issue with the disks or access at R@H ?

\"Temporarily failed upload\" ... \"project communication failed\", \"Internet access OK\". My internet access is ok, I can download new packages, but not upload them. Those failing are usually around 400 Kb in size.

Regards,
BP


If it were an issue with the disks, I\'d expect it to block downloads as well.

It looks more likely to be an issue with a server program handling uploads but not downloads.

BP, you may want to change the sort option at the top right of the thread - it show the messages oldest first by default.
54) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 80226)
Posted 23 Jun 2016 by Profile robertmiles
Post:
Rosetta@Home appears to have had some problem that block uploads for the last several hours.

This workunit is affected:

http://boinc.bakerlab.org/rosetta/workunit.php?wuid=759093541

It appears that downloads and getting new workunits to clients are probably not affected.
55) Message boards : Number crunching : Minirosetta 3.73-3.78 (Message 80164)
Posted 7 Jun 2016 by Profile robertmiles
Post:
Hi Sid,

Thanks for the alert! looks like these jobs require lots of memory. We have a way to specify how much memory to use. It will corrected in the next round of submission!

That would kind of explain why the task runs for a reasonable while before crashing out, and I\'ve seen occasional tasks using 1.2Gb, but I\'m running with just short of 10Gb free of 16Gb total.

I set Boinc to run 60% of memory when the computer is in use (90% when not in use). Do people routinely allocate more than that? What can I safely adjust that setting to, or is it just trial and error?


I\'ve found that 64-bit Windows Vista is rather inefficient at handling memory for running 32-bit applications, so I set that computer to use 30% to 40% of the memory for BOINC out of 8 GB. 64-bit Windows 7 and Windows 10 are more efficient, so I set that computer to use 70% out of 16 GB. 64-bit BOINC is not very good at giving up memory when the computer is in use, so these settings are the same for when the computer is in use as when not in use.
56) Message boards : Number crunching : Minirosetta 3.73-3.78 (Message 80021)
Posted 4 May 2016 by Profile robertmiles
Post:
I\'m seeing the same issue that\'s been reported above: an intermittent failure to get new workunits accompanied by this message in the event log.

Rosetta Mini for Android is not available for your type of computer.

Do Network Communication successfully reports the task.

I\'m running 2 R@h workunits/Ubuntu 14.04 LTS/Boinc 7.2.42/No workbuffer


I just discovered: ANDROID is a 32 bit operating system. Is it possible that Rosetta@Home cannot provide for a 32bit operating system? My computer is 32bit Linux -- I could easily load in the 64bit and try again. What does one think? :^)


I\'ve been seeing this problem under 64 bit Windows, so don\'t expect 64 bits alone to be a cure.
57) Message boards : Number crunching : Minirosetta 3.73-3.78 (Message 80011)
Posted 3 May 2016 by Profile robertmiles
Post:

Is there an ETA for the resolution of this issue?


As I understand it, the issue is that the project briefly exhausted it\'s supply of work available for non-Android devices. When only Android tasks are found in the scheduler, and you are coming in with a non-Android host, it reports the message \"Rosetta Mini for Android is not available for your type of computer.\"

This problem is resolved in ebbs and flows as large number of new hosts are coming in to the system. The tasks that make new work available are running continuously, but seem to hit periods of time where they still are barely keeping up with the incoming requests for work.


Could the number of tasks ready to send be divided by what host types they are suitable for, so that users can easily tell when only Android tasks are found in the scheduler?

Also, could thing be adjusted so that when no tasks are available for the type of computer requesting them, but many are available for other types of computers, the delay is set much below 24 hours?
58) Message boards : Number crunching : Minirosetta 3.73-3.78 (Message 79982)
Posted 30 Apr 2016 by Profile robertmiles
Post:
I have the same problem here on my Win7 Ultimate, no updates / changes from my side. Everything works fine up to now.

30.04.2016 11:19:55 | rosetta@home | Sending scheduler request: To fetch work.
30.04.2016 11:19:55 | rosetta@home | Requesting new tasks for CPU and NVIDIA GPU and Intel GPU
30.04.2016 11:19:58 | rosetta@home | Scheduler request completed: got 0 new tasks
30.04.2016 11:19:58 | rosetta@home | No work sent
30.04.2016 11:19:58 | rosetta@home | Rosetta Mini for Android is not available for your type of computer.

Wtf, since when have i a Android? What comes next? A Alien? A Zombie?


Looks like someone with access to the source code for the Windows application should add a check at the end for whether the operating system setting still indicates Windows.
59) Message boards : Number crunching : Minirosetta 3.73-3.78 (Message 79976)
Posted 30 Apr 2016 by Profile robertmiles
Post:
I\'m seeing the same issue that\'s been reported above: an intermittent failure to get new workunits accompanied by this message in the event log.

Rosetta Mini for Android is not available for your type of computer.

Do Network Communication successfully reports the task.

I\'m running 2 R@h workunits/Ubuntu 14.04 LTS/Boinc 7.2.42/No workbuffer

I just discovered the same thing in a machine I just assembled. I have a hunch while I am waiting for the Volunteer at BOINC to get back to me: My box has only one \'new\' component: the SSD. I purchased an economy SSD at Amazon which works perfectly otherwise. This is just a hunch. Could the memory chips in the SSD somehow be marked as for use in a cellphone? Forgive me if that is stupid. All my other components have crunched Rosetta before (all the other parts I used in this build.) I also have NO gripes pending on this SSD.


What operating system did you see this under? Windows, Linux, Android, or something else?

I\'ve had many Rosetta Mini workunits complete on one of my computers, which came with an SSD as its main drive. It uses Windows 10.

It\'s happening everywhere, Win7 and no SSD here. Manually updating often fixes it, but in the meantime Boinc puts in a 24hr delay before retrying, which is very annoying if you don\'t notice.


Something you might want to try: Check if restarting Win7, with no updates, fixes the problem, with a possible exception for the 24hr delay.

Also, you might want to check for a problem I\'ve seen in my last three completed workunits: For two of them, the application apparently completed with no problem seen. But then, the upload process for the output files gave a compute error of the upload error type. For the third one, a wingmate had this upload error, but then my computer got the workunit, completed it properly, then uploaded the output properly, then had one more validated workunit. The computer appears to about 5 hours into a 24hr delay; the other 19 hours should be more than enough to finish the other three Rosetta@home task it currently has.
60) Message boards : Number crunching : Minirosetta 3.73-3.78 (Message 79958)
Posted 28 Apr 2016 by Profile robertmiles
Post:
I\'m seeing the same issue that\'s been reported above: an intermittent failure to get new workunits accompanied by this message in the event log.

Rosetta Mini for Android is not available for your type of computer.

Do Network Communication successfully reports the task.

I\'m running 2 R@h workunits/Ubuntu 14.04 LTS/Boinc 7.2.42/No workbuffer

I just discovered the same thing in a machine I just assembled. I have a hunch while I am waiting for the Volunteer at BOINC to get back to me: My box has only one \'new\' component: the SSD. I purchased an economy SSD at Amazon which works perfectly otherwise. This is just a hunch. Could the memory chips in the SSD somehow be marked as for use in a cellphone? Forgive me if that is stupid. All my other components have crunched Rosetta before (all the other parts I used in this build.) I also have NO gripes pending on this SSD.


What operating system did you see this under? Windows, Linux, Android, or something else?

I\'ve had many Rosetta Mini workunits complete on one of my computers, which came with an SSD as its main drive. It uses Windows 10.


Previous 20 · Next 20



©2019 University of Washington
http://www.bakerlab.org