Welcome to MilkyWay@home

Posts by Keith Myers

1) Message boards : Number crunching : Any enhanced applications developed for Milkyway to speed up GPU work? (Message 69530)
Posted 10 days ago by ProfileKeith Myers
Post:
I did once at my work switch back on a breaker which I assumed had tripped. Actually a workman had turned it off to rewire something. His fault for not labelling it. He didn't know I had keys to that room :-) He was a bit annoyed when he got a big jolt - 3 phase hurts!

Shame on him! Lock out - Tag out was not followed. Could have been much worse than a jolt.
2) Message boards : Number crunching : Any enhanced applications developed for Milkyway to speed up GPU work? (Message 69483)
Posted 18 days ago by ProfileKeith Myers
Post:
How odd. All my Windows installations upgraded for me. I just got a notice saying would you like free Windows 10, so I clicked yes and it happened. Hell they even upgrade pirated Windows for free.

I reiterate my comment about you being exceedingly lucky and blessed by Microsoft.

That issue of pirated Windows7 was one of the issues I had to overcome. I had no clue that the Windows 7 installation was unlicensed. It said it was a genuine copy of Windows. I had the license key for Windows 7 and I thought from all the articles I read that upgrading to Win 10 with the Win 7 license key was a simple matter. Two thirds into the Win10 install, (2 hours later), I get an error message that the upgrade can't proceed and it needed to roll me back to Win 7. Back to square one except now Windows 7 said it was an unlicensed copy with a watermark in the bottom right of the screen. Fought it for another 8 hours wrangling Win 10 onto the machine. What a effing headache.
3) Message boards : Number crunching : Any enhanced applications developed for Milkyway to speed up GPU work? (Message 69482)
Posted 18 days ago by ProfileKeith Myers
Post:
Latest Ubuntu, 19.10 has the Nvidia 430 drivers in the install package. No questions asked, no hoops to jump through. The installer sees Nvidia hardware and installs the drivers for you, boom.Done.

What about AMD? I am putting together a new machine with an RX 570, and can do either Ubuntu 18.04.4 in a few days, or go with 19.10.
I figured I would stay with 18.04, since I think I have to install the drivers myself anyway, but is there any advantage to 19.10?

I would say yes. AMD is providing new drivers and the compatibility packages like Mesa, Vulkan, DXVK etc. are all dumping new updates fast and furious every day it seems. The latest kernels in 19.10 make use of all the new AMD updates. https://www.phoronix.com/scan.php?page=article&item=linux-54-radeon&num=1 I would say just read the news everyday at http://Phoronix.com to see how much activity is taking place constantly on AMD support.
4) Message boards : Number crunching : Any enhanced applications developed for Milkyway to speed up GPU work? (Message 69474)
Posted 19 days ago by ProfileKeith Myers
Post:
I guess you are lucky and I am unlucky with Windows 10. I never had any issues with Windows 7. Worked all the time, every time.

As far as my hardware is concerned, no issues with the hardware from the system that wouldn't boot Windows 10 or repair or recover it, to simply putting my Ubuntu 18.04 LTS bootable install USB stick in the computer and was up and running the Linux desktop in 10 minutes. That was after fighting Win 10 for several hours. Enough was enough. Goodbye Win 10.

And my opinion of Win 10 was just recently enforced when an ex-coworker asked me to update his Windows 7 computer I built for him to Windows 10 because of EOL declared on Win 7. 12 excruciating hours later I finally had a working Win 10 installation for him that saved his data.

I just don't want to have to work so hard to install a OS.
5) Message boards : Number crunching : Any enhanced applications developed for Milkyway to speed up GPU work? (Message 69472)
Posted 19 days ago by ProfileKeith Myers
Post:
in some versions of Linux even the Nvidia drivers are a pain to get loaded compared to at least one, Linux Mint, that asks you if you want to 3rd party drivers during the install process.

Latest Ubuntu, 19.10 has the Nvidia 430 drivers in the install package. No questions asked, no hoops to jump through. The installer sees Nvidia hardware and installs the drivers for you, boom.Done.
6) Message boards : Number crunching : Any enhanced applications developed for Milkyway to speed up GPU work? (Message 69471)
Posted 19 days ago by ProfileKeith Myers
Post:
May be fine in Windows, but a completely different experience in Linux. And based on the number of posts I have made helping others with AMD driver problems in Linux, it is a common and difficult problem.


ESPECIALLY since Linux deprecated the AMD drivers a few versions back to non crunching drivers as the default ones, in some versions of Linux even the Nvidia drivers are a pain to get loaded compared to at least one, Linux Mint, that asks you if you want to 3rd party drivers during the install process. Several just load the Mesa drivers and then make you hunt for how to install them thru the gui way, MXLinux was very hard to find them. I loaded half a dozen different versions of Linux on a machine I replaced just to see if there are easier ones to use than the one I do use for just crunching and didn't find one easier than Linux Mint for gpu drivers and just working pretty quickly. I always put Nvdia cards in Linux machines though because command line loading is picky and constantly changing file names, as the drivers get updated, means you have to look first.


I don't mean to be rude, but Linux sounds like a bit of a pain. Windows just works.

Everyone is entitled to their opinion. In my case, Windows was a pain in the ass to work with. Kept blowing up. Finally decided since I had to reinstall an OS again for the umpteenth time, why not just give Windows the old heave-ho and jump to Linux. All my other hosts had been running Linux for a year with no issues. So converted my sole Windows 10 experiment to Linux and never looked back. I don't miss Windows in any way at all. I can do everything I want to do in Linux, just don't need anything from Windows including its grief.

For installing Nvidia drivers, nothing could be simpler than adding the graphics-drivers ppa and choosing what version of the drivers you want to install and one reboot later to load the drivers you are up and running.
7) Message boards : Number crunching : Any enhanced applications developed for Milkyway to speed up GPU work? (Message 69467)
Posted 20 days ago by ProfileKeith Myers
Post:
May be fine in Windows, but a completely different experience in Linux. And based on the number of posts I have made helping others with AMD driver problems in Linux, it is a common and difficult problem.
8) Message boards : Number crunching : Any enhanced applications developed for Milkyway to speed up GPU work? (Message 69465)
Posted 21 days ago by ProfileKeith Myers
Post:
I have always and will always buy AMD.

That is certainly your prerogative as a consumer.

In my case, I will never buy an AMD gpu because the drivers have been, are, and always will be a nightmare to install.

Nvidia on the other hand . . . download, install and boom, up and running in minutes. And never have to struggle with them ever again.
9) Message boards : Number crunching : Any enhanced applications developed for Milkyway to speed up GPU work? (Message 69461)
Posted 22 days ago by ProfileKeith Myers
Post:
Actually, the very first CUDA application for Seti was written by actual Nvidia developers and programmers.

It is up to the individual project scientists to develop CUDA applications or have the project volunteers do it for the project. Since most science projects operate on a shoe-string budget, there are no resources to fund a project programmer.

The fact that we have two volunteer GPUUG team members competent in CUDA programming allowed us to take the legacy CUDA code developed by Nvidia and early Seti project volunteers and morph it into the present special app CUDA10 application. That is why the CUDA10 application can use all of the present day hardware capabilities and is so impressively fast.
10) Message boards : Number crunching : OK, I think I'm stuck.... (Message 69454)
Posted 23 days ago by ProfileKeith Myers
Post:
Click on the task in the running or Active Task list in Manager and then click Properties from the side window panel. You should see the time of the most recent checkpoint and also the Progress rate. As long as the checkpoint is less than your checkpoint interval from present time and the progress rate is not 0, then the task is still crunching. You could also look at the task in the OS's Task Manager and see how much cpu % it is using. If it is not zero, then it is actively crunching.
11) Message boards : Number crunching : Any enhanced applications developed for Milkyway to speed up GPU work? (Message 69453)
Posted 23 days ago by ProfileKeith Myers
Post:
How do I tell which projects use CUDA? I assume the Nvidia icon here could mean using OpenCL: https://boinc.berkeley.edu/projects.php

Actually, the Nvidia logo on the project lists denotes gpu applications, NOT necessarily OpenCL applications nor CUDA applications. You would still have to look at the https://{project]/apps.php page to see if their are CUDA applications.
12) Message boards : Number crunching : Any enhanced applications developed for Milkyway to speed up GPU work? (Message 69450)
Posted 23 days ago by ProfileKeith Myers
Post:
Yip, I always go for double precision cards. I fail to see why they make them without it. I also detest the way Nvidia hate OpenCL

They do make double-precision optimized cards, they are called Quadros and Teslas. They charge a pretty penny for them too since the target clientele are high-speed Wall Street traders and computational scientists.
OpenCL is a generalized computing API. Not best for any one thing, and just mediocre at all things. So why further develop a mediocre programming API when you developed the first "race-horse" computing API called CUDA. The performance of CUDA over OpenCL is not even close. They have plenty of customers who are willing to pay for a thoroughbred. So their focus is to continue to develop CUDA. They joined the OpenCL consortium out of the desire to appear to not block OpenCL and not appear to try and force CUDA down any one else's throat. So there is an alternative choice if so desired.

The better double-precision math of AMD cards compared to Nvidia for the consumer cards has nothing to do with the programming API but rather in the choices of how to configure the silicon of their gpus for their target gamer audience.
13) Message boards : Number crunching : Any enhanced applications developed for Milkyway to speed up GPU work? (Message 69443)
Posted 26 days ago by ProfileKeith Myers
Post:
The Nvidia CUDA app you are mentioning is euphemistically called "the special app" or "the special sauce" app. It comes in CUDA9 and CUDA10.2 flavors currently. Only usable in Linux with cards of at least C.C. 5.0 capability which means Maxwell generation onwards. It is not the SoG app. The SoG and the SAH apps are OpenCL based and can run on any card with the OpenCL API in the drivers. A couple of our GPUUG Team members are the developers. It is not released as an official stock application. Much like the older Lunatics installer optimized applications. It is somewhere between 5X to 10X faster than the stock SoG app depending on the card generation. Seti tasks are typically finished in under a minute, usually closer to 30 seconds. The application is bundled into a All-in-One BOINC package which is a ready to run BOINC platform environment once unpacked onto the Desktop or somewhere in the /home folder. Just double click the client boinc file and it is up and running. The package is at the Crunchers Anonymous website. It can be thought of as the Lunatics Installer Gen. 2.
http://www.arkayn.us/lunatics/BOINC.7z
14) Message boards : Number crunching : Combining computers and credits (Message 69438)
Posted 18 Jan 2020 by ProfileKeith Myers
Post:
You can't on different accounts. You can merge similar hosts on the same account. The newest host absorbs the older host and gains its historical credits.
15) Message boards : Number crunching : Is there a server-limited maximum at 100 CPU workunits? (Message 69425)
Posted 6 Jan 2020 by ProfileKeith Myers
Post:
I recently asked the same question at Seti. The answer was that it was not a BOINC issue but rather a Project level server side constraint. I questioned whether the cpu job limit of 100 was practical anymore when modern processors have 128 or more now.

The Seti cpu job limit was recently doubled. So all it takes is a project scientist to change the limits.
16) Message boards : Application Code Discussion : Need help to build the Milkyway apps? Look here (Message 69409)
Posted 27 Dec 2019 by ProfileKeith Myers
Post:
Yes, the code changes for calculating GFLOPs was broken for AMD cards running the AP tasks at Seti. The tasks were all erroring out because the calculated time to compute was ridiculously short and they had "exceeded time limit errors" after 14 seconds or so. This was merged into the 7.16.3 branch back in February. #2988

The introduction of the Nvidia Turing cards introduced another issue with correctly calculating GFLOPs because of the cores_per_proc = 64 being changed from the previous standard of 128 for all previous generations. This got changed in #2706

If they are simply reusing the OpenCL API library from Nvidia as you state, then I can see that if the S9000 series had a sufficiently different architecture from the previous AMD generation, it too will calculate the wrong peak flops and probably get back the wrong device parameters from that API interrogation.
17) Message boards : Application Code Discussion : Need help to build the Milkyway apps? Look here (Message 69408)
Posted 27 Dec 2019 by ProfileKeith Myers
Post:
deleted
18) Message boards : Application Code Discussion : Need help to build the Milkyway apps? Look here (Message 69406)
Posted 27 Dec 2019 by ProfileKeith Myers
Post:
The S9000 series of cards can't be too common. Do you think it possible that the AMD API is not returning valid device parameters for the statement in gpu_amd.cpp?
 p_calDeviceGetAttribs =(ATI_ATTRIBS)GetProcAddress(callib, "calDeviceGetAttribs" );


I see that they had to change various cards SIMD counts explicitly in the file to correct incorrect returns from the API.
        case CAL_TARGET_610:
            gpu_name="ATI Radeon HD 2300/2400/3200/4200 (RV610)";
            attribs.numberOfSIMD=1;        // set correct values (reported wrong by driver)
            attribs.wavefrontSize=32;
            break;
        case CAL_TARGET_630:
            gpu_name="ATI Radeon HD 2600/3650 (RV630/RV635)";
            // set correct values (reported wrong by driver)
            attribs.numberOfSIMD=3;
            attribs.wavefrontSize=32;


Could this be the case with the S9000 cards also?
19) Message boards : Number crunching : Finally getting new tasks only seconds after running out. May not be worth the hassle. (Message 69369)
Posted 19 Dec 2019 by ProfileKeith Myers
Post:
I have never seen a cc_config that big. I am surprised that boinc did not choke on it. I certainly did.

That is the stock cc_config.xml that is created and fully populated the first time you change the logging options of the Event Log in the Manager. BOINC ships with no cc_config.xml initially.

Simplest way to get a correctly formatted one with all the possible parameters is change a logging flag, even just temporarily and it writes out the cc_config.xml file.
20) Message boards : Number crunching : Radeon R9 280X for MilkyWay@Home (Message 69349)
Posted 16 Dec 2019 by ProfileKeith Myers
Post:
I believe they are great producers for this project. They just use a lot of power compared to later generations.


Next 20

©2020 Astroinformatics Group