Open Beta test: SoG for NVidia, Lunatics v0.45 - Beta6 (RC again)

Message boards : Number crunching : Open Beta test: SoG for NVidia, Lunatics v0.45 - Beta6 (RC again)
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 22 · 23 · 24 · 25 · 26 · 27 · 28 . . . 32 · Next

AuthorMessage
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 1822489 - Posted: 7 Oct 2016, 14:28:49 UTC - in response to Message 1822471.  

see "did i miss the memo"


ED F



. . OK I will look for it, thanks.

Stephen

.
ID: 1822489 · Report as offensive
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 1823714 - Posted: 12 Oct 2016, 13:10:09 UTC

. . Hello Richard,

. . Since Raistmer has announced that V8.19 has been released does that mean you can also release the Lunatics update?

. . I have been looking forward to it. Only thing missing will be SSE4.2 support :)

Stephen

.
ID: 1823714 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14677
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1823715 - Posted: 12 Oct 2016, 13:14:16 UTC - in response to Message 1823714.  

. . Hello Richard,

. . Since Raistmer has announced that V8.19 has been released does that mean you can also release the Lunatics update?

. . I have been looking forward to it. Only thing missing will be SSE4.2 support :)

Stephen

As already hinted at in message 1823688.
ID: 1823715 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14677
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1823755 - Posted: 12 Oct 2016, 16:22:08 UTC - in response to Message 1809276.  

As the title says. Beta test only, try at your own risk.

I think we should dig out the old "Advanced Users only" tag for this one - I'm not sure whether it will fly or not.

Beta2 now available, in both 64-bit and 32-bit versions. Pick the right one, or test out my new dialogs...

Now updated again with Beta3 (both 64-bit and 32-bit).

And a further update to Beta4, to include r3500 where available - that's the SoG app for NVidia cards, and the ATi app for discrete GPUs (ATi_APU not updated this time).

For NVidia SoG, please note that Raistmer has changed the name of the command-line tuning file from mb_cmdline_win_x86_SSE3_OpenCL_NV.txt to

mb_cmdline_win_x86_SSE3_OpenCL_NV_SoG.txt

A (hopefully) final Beta5 update, containing the r3528 applications released to the Main project as v8.19 this morning. The ATI 'discrete GPU', NVidia SoG, and Intel GPU applications have all been updated, though the iGPU app still hasn't eliminated the problems with Skylake-class GPUs. That's the next job...

This build can be considered a Release Candidate as far as the applications and app_info segments are concerned: I've only had to add 69 <app_version> statements this afternoon, to cover all the new version numbers and upgrade paths. Please check as many of them as you can. This version should upgrade cleanly from

8.19 (opencl_nvidia_SoG)
8.12 (opencl_nvidia_sah)
8.12 (opencl_nvidia_SoG)
8.00 (opencl_nvidia_sah)
8.00 (opencl_nvidia_SoG)
8.00 (cuda50)
8.00 (cuda42)
8.00 (cuda32)
8.00 (cuda23)

and similarly for ATI and iGPU. Fortunately, we've lost some of the experimental ATI apps since May, so the list can start shrinking again next time round.

I need to work through the ReadMe files while y'all test that it works, and then we should be good for a public release.

Download url remains: https://onedrive.live.com/redir?resid=8D83BF774A4A86F5!979&authkey=!AK7t-36P9nAec2Y&ithint=folder%2cexe
ID: 1823755 · Report as offensive
Harri Liljeroos
Avatar

Send message
Joined: 29 May 99
Posts: 4740
Credit: 85,281,665
RAC: 126
Finland
Message 1823772 - Posted: 12 Oct 2016, 17:54:10 UTC - in response to Message 1823755.  

Firefox, Chrome and IE are all trying to block this download. IE allowed me to bypass the warnings and download it anyway. When scanned wit Avast no problems were found.

After I installed this it didn't restart Boinc but I started it manually. All are running normally now but I was already running these applications (SoG r3528 as 8.12) so that's no surprise.
ID: 1823772 · Report as offensive
Oddbjornik Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 15 May 99
Posts: 220
Credit: 349,610,548
RAC: 1,728
Norway
Message 1823774 - Posted: 12 Oct 2016, 18:11:04 UTC - in response to Message 1823772.  

Firefox, Chrome and IE are all trying to block this download. IE allowed me to bypass the warnings and download it anyway. When scanned wit Avast no problems were found.

After I installed this it didn't restart Boinc but I started it manually. All are running normally now but I was already running these applications (SoG r3528 as 8.12) so that's no surprise.

Same here, with Firefox. Firefox also allowed me to bypass the warning, and Eset Nod32 says the file is ok.

Looks like it might need some sort of signature or whatnot to avoid these dire warnings.
ID: 1823774 · Report as offensive
Profile Jeff Buck Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester

Send message
Joined: 11 Feb 00
Posts: 1441
Credit: 148,764,870
RAC: 0
United States
Message 1823775 - Posted: 12 Oct 2016, 18:12:17 UTC - in response to Message 1823772.  

Firefox, Chrome and IE are all trying to block this download.

I ran into the same problem with Firefox, both with the Win32 and Win64 versions. It actually seems to be the browser that's trying to block it, rather than McAfee. I can't say I've ever seen that before. However, Firefox does allow me to unblock the download, at which point McAfee promptly scans the file and says everything's fine. Later scans of the extracted files are fine, too.
ID: 1823775 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14677
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1823779 - Posted: 12 Oct 2016, 18:35:43 UTC - in response to Message 1823775.  

We often get problems like this when a file is first made available for download. I can't even download it myself using Chrome, but I suspect the reason is, in Google's words,

If your site's downloads are identified as "Uncommonly downloaded"
If Google Safe Browsing hasn't seen a particular binary before, Chrome may warn that it is uncommonly downloaded and could be dangerous. In these cases the warnings are lifted automatically if Google Safe Browsing verifies that it is benign.

So, have a good night's sleep everyone, and we'll try again in the morning.
ID: 1823779 · Report as offensive
AMDave
Volunteer tester

Send message
Joined: 9 Mar 01
Posts: 234
Credit: 11,671,730
RAC: 0
United States
Message 1823794 - Posted: 12 Oct 2016, 20:02:26 UTC - in response to Message 1823779.  

Using Firefox 49.0.1, and Avast 11.2.2262.
    • To download Lunatics_Win64_v0.45_Beta5-for-SoG_setup.exe, I had to disable "browser.safebrowsing.downloads.enabled" in Firefox.

    •  To download Lunatics_Win32_v0.45_Beta5-for-SoG_setup.exe, I had to do as above, and disable "Web Shield" in Avast.


Subsequent scanning by Avast yielded a warning for the 32-bit .exe only ("Win32:Evo-gen [Susp]").

ID: 1823794 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1823798 - Posted: 12 Oct 2016, 20:25:18 UTC - in response to Message 1823794.  
Last modified: 12 Oct 2016, 20:26:15 UTC

Chrome 54 downloaded it and then complained it was malicious. It wanted an acknowledgement to Discard the file at the bottom banner area. Before I closed Chrome I found the downloaded file in my Download directory and moved it elsewhere, then closed Chrome. I renamed the file to what Richard originally named it and everything is fine. One way around the false positive. It tests fine in Kaspersky.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1823798 · Report as offensive
Profile mr.mac52
Avatar

Send message
Joined: 18 Mar 03
Posts: 67
Credit: 245,882,461
RAC: 0
United States
Message 1823830 - Posted: 12 Oct 2016, 23:07:04 UTC - in response to Message 1823779.  

I was able to download the 64 bit version with a simple over ride of Chrome's warning. On my system it is running well with no errors.
ID: 1823830 · Report as offensive
robertmiles
Volunteer tester

Send message
Joined: 16 Jan 12
Posts: 213
Credit: 4,117,756
RAC: 6
United States
Message 1823835 - Posted: 12 Oct 2016, 23:24:01 UTC - in response to Message 1823794.  
Last modified: 12 Oct 2016, 23:33:34 UTC

Using Firefox 49.0.1, and Avast 11.2.2262.
    • To download Lunatics_Win64_v0.45_Beta5-for-SoG_setup.exe, I had to disable "browser.safebrowsing.downloads.enabled" in Firefox.

    •  To download Lunatics_Win32_v0.45_Beta5-for-SoG_setup.exe, I had to do as above, and disable "Web Shield" in Avast.


Subsequent scanning by Avast yielded a warning for the 32-bit .exe only ("Win32:Evo-gen [Susp]").


I was able to download the Win64 version with Firefox 49.0.1 without that setting, but it took opening the downloads window of Firefox, seeing that it was marked as possibly malicious, and telling Firefox to download it anyway. Without that last step, it was a .exe.part file instead of a .exe file, and Windows didn't know how to run it.

Scanned with both Windows Defender (under Windows 10) and Microsoft Security Essentials (under Windows Vista); no problems found.

Now installed on both computers.

A problem seen with beta4 - r3500 told BOINC Manager it only used 0.16 of a CPU, but it actually used closer to 1.00 of a CPU. Beta5 makes the same claim but also uses much more than it claims. This causes slow console response. at least under 64-bit Windows Vista, if you expect 1 CPU core to be shared between GPU tasks and any console operations.
ID: 1823835 · Report as offensive
Profile Zalster Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 27 May 99
Posts: 5517
Credit: 528,817,460
RAC: 242
United States
Message 1823838 - Posted: 12 Oct 2016, 23:38:00 UTC - in response to Message 1823830.  

I was able to download the 64 bit version with a simple over ride of Chrome's warning. On my system it is running well with no errors.


I too was able to download it using the above mentioned override. Installed and running.
ID: 1823838 · Report as offensive
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 1823844 - Posted: 12 Oct 2016, 23:58:46 UTC - in response to Message 1823755.  


A (hopefully) final Beta5 update, containing the r3528 applications released to the Main project as v8.19 this morning. The ATI 'discrete GPU', NVidia SoG, and Intel GPU applications have all been updated, though the iGPU app still hasn't eliminated the problems with Skylake-class GPUs. That's the next job...

I need to work through the ReadMe files while y'all test that it works, and then we should be good for a public release.

Download url remains: https://onedrive.live.com/redir?resid=8D83BF774A4A86F5!979&authkey=!AK7t-36P9nAec2Y&ithint=folder%2cexe



. . Thanks Richard,

. . But what is it with Windows10/Avast lately? I had to wrestle them to the ground to get that download. They reported that "this app is disguised as a friendly app but could make unexpected changes to your system" or words to that effect. It seems anything that is not Microsoft is a virus :(

Stephen

.
ID: 1823844 · Report as offensive
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 1823845 - Posted: 13 Oct 2016, 0:08:28 UTC - in response to Message 1823835.  


A problem seen with beta4 - r3500 told BOINC Manager it only used 0.16 of a CPU, but it actually used closer to 1.00 of a CPU. Beta5 makes the same claim but also uses much more than it claims. This causes slow console response. at least under 64-bit Windows Vista, if you expect 1 CPU core to be shared between GPU tasks and any console operations.


. . There is a lot of information about this in Raistmers threads.

. . But if you use sleep, ie: add the following command to your commandline file, this will break the CPU hold and let it run only parttime to support the GPU task.

-high_prec_timer -use_sleep

. . I hope that helps.

Stephen

.
ID: 1823845 · Report as offensive
EdwardPF
Volunteer tester

Send message
Joined: 26 Jul 99
Posts: 389
Credit: 236,772,605
RAC: 374
United States
Message 1823862 - Posted: 13 Oct 2016, 1:03:33 UTC
Last modified: 13 Oct 2016, 1:06:26 UTC

downloaded via Opera through McAfee no prob's reported.

running on Win-7 core-7 (gen 1) on a nvidia 770

Will report anything if I see it.

Ed F

Edit First observation: WU startup takes 141 sec's compared to r3500's 28 sec's
ID: 1823862 · Report as offensive
Profile Jeff Buck Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester

Send message
Joined: 11 Feb 00
Posts: 1441
Credit: 148,764,870
RAC: 0
United States
Message 1823880 - Posted: 13 Oct 2016, 2:27:17 UTC

I suppose I'm beating a dead horse here (although I sincerely hope not), but once again the device-specific configuration files appear to have been omitted from the installer. I've previously raised this inconsistency in the treatment of configuration and command line files regarding AP, but with SoG we now have the same problem with Multibeam.

A relevant section of one of the Multibeam OpenCL ReadMe files states:
For device-specific settings in multi-GPU systems it's possible to override some of command-line options via
application config file.

Name of this config file:
MultiBeam_<vendor>_config.xml where vendor can be ATi, NV or iGPU.

IIRC, dummy configuration files get distributed with the stock applications, but so far those files haven't carried into the Lunatics installer. The upshot of that is, if you're already using one of those files, it'll get wiped out of the app_info by the installer. Or, if you're not yet using one but decide to do so in the future, you'll have to manually edit it into the appropriate places in your app_info. That negates what, to me, is the main benefit of the installer, avoiding having to manually edit app_info, a sometimes dark and scary experience.
ID: 1823880 · Report as offensive
robertmiles
Volunteer tester

Send message
Joined: 16 Jan 12
Posts: 213
Credit: 4,117,756
RAC: 6
United States
Message 1823894 - Posted: 13 Oct 2016, 4:11:53 UTC - in response to Message 1823845.  


A problem seen with beta4 - r3500 told BOINC Manager it only used 0.16 of a CPU, but it actually used closer to 1.00 of a CPU. Beta5 makes the same claim but also uses much more than it claims. This causes slow console response. at least under 64-bit Windows Vista, if you expect 1 CPU core to be shared between GPU tasks and any console operations.


. . There is a lot of information about this in Raistmers threads.

. . But if you use sleep, ie: add the following command to your commandline file, this will break the CPU hold and let it run only parttime to support the GPU task.

-high_prec_timer -use_sleep

. . I hope that helps.

Stephen

.

I'll try that. Currently, Beta5 appears to be trying to use more than a whole CPU core, and therefore cannot keep the GPU busy.
ID: 1823894 · Report as offensive
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 1823909 - Posted: 13 Oct 2016, 7:03:13 UTC - in response to Message 1823880.  


IIRC, dummy configuration files get distributed with the stock applications, but so far those files haven't carried into the Lunatics installer. The upshot of that is, if you're already using one of those files, it'll get wiped out of the app_info by the installer. Or, if you're not yet using one but decide to do so in the future, you'll have to manually edit it into the appropriate places in your app_info. That negates what, to me, is the main benefit of the installer, avoiding having to manually edit app_info, a sometimes dark and scary experience.


. . I find app_config.xml a much less scary place to do configuring.

Stephen

.
ID: 1823909 · Report as offensive
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 1823910 - Posted: 13 Oct 2016, 7:12:10 UTC - in response to Message 1823894.  


I'll try that. Currently, Beta5 appears to be trying to use more than a whole CPU core, and therefore cannot keep the GPU busy.


. . If you address your concerns in the SoG_v8.19(r3528) support thread Raistmer has started he may answer your concerns when he is able. But most of the info you need is covered in the now closed, but still readable, thread V8.12 support. Look for messages with command line options to tweak the performace of the app. The defaults are pretty successful but there are things like -period_iterations num and -tt. The default for the iterations is 50 but you can try dropping it to boost the GPU utilisation, or increase -tt from 500 up to 1500 but it can cause lag in your user interface doing these things.

. . But there is a great deal of info in those threads.

Stephen

.
ID: 1823910 · Report as offensive
Previous · 1 . . . 22 · 23 · 24 · 25 · 26 · 27 · 28 . . . 32 · Next

Message boards : Number crunching : Open Beta test: SoG for NVidia, Lunatics v0.45 - Beta6 (RC again)


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.