Posts by Jesse Viviano
log in
1) Message boards : Number crunching : Errors (Message 66631)
Posted 17 Sep 2017 by Jesse Viviano
I made a bad post because I failed to read the date on the thread I linked to. Please ignore.
2) Message boards : Number crunching : Nvidia GPU compute client only doing 1/5 of the required work (Message 66374)
Posted 5 May 2017 by Jesse Viviano
Either a new driver or the new application version fixed the bug.
3) Message boards : Application Code Discussion : AVG detecting/blocking Milkyway App due to malware/virus? (Message 66370)
Posted 4 May 2017 by Jesse Viviano
Submit the false positive at https://secure.avg.com/submit-sample.
4) Message boards : Number crunching : Nvidia GPU compute client only doing 1/5 of the required work (Message 66285)
Posted 9 Apr 2017 by Jesse Viviano
I just filed a bug report with Nvidia over this issue.
5) Message boards : Number crunching : Nvidia GPU compute client only doing 1/5 of the required work (Message 66269)
Posted 3 Apr 2017 by Jesse Viviano
There is no sign that Windows 10 reinstalled any driver. Since other projects' GPU work still is being turned in and successfully validating, I would think that the problem is either a bug exists in the driver, or a new optimization in the driver exposed a previously undetected bug in the GPGPU software client (the latter has happened with Folding@home).
6) Message boards : Number crunching : Nvidia GPU compute client only doing 1/5 of the required work (Message 66267)
Posted 3 Apr 2017 by Jesse Viviano
The problem is back. I have had to disable GPU work for MilkyWay@home for now.
7) Message boards : Number crunching : Nvidia GPU compute client only doing 1/5 of the required work (Message 66266)
Posted 3 Apr 2017 by Jesse Viviano
I reinstalled my video driver using a clean install, and that seems to have fixed the problem.
8) Message boards : Number crunching : Nvidia GPU compute client only doing 1/5 of the required work (Message 66262)
Posted 2 Apr 2017 by Jesse Viviano
I also noticed that I am using the SSE4.1 path when my CPU, a Xeon E5-2690 v3, is capable of using AVX. In the work units where the GPU client failed, those who did the cleanup work after my computer messed up were using an AVX path. Could something be wrong with the SSE4.1 code path?
9) Message boards : Number crunching : Nvidia GPU compute client only doing 1/5 of the required work (Message 66261)
Posted 2 Apr 2017 by Jesse Viviano
I recently had to shut off GPU work for this project in my preferences because the Nvidia GPU client was only doing the final work unit in the now-standard 5 work unit bundle, causing my work to be declared as validate errors. The client was working fine until April 1 (No, this is no April fools joke.) My GPU is still generating valid work for other projects. Is anyone else having a similar problem?
10) Message boards : Number crunching : WU restart every 20% (Message 65988)
Posted 7 Dec 2016 by Jesse Viviano
There is a bug in how percentages are calculated by the MilkyWay@home application. As for why your work units are slower, the newer work units bundle 5 of the older work units into one bigger work unit to reduce BOINC server load because double-precision capable GPUs run these tasks much faster than CPUs, causing GPU crunchers to overwhelm the BOINC server.
11) Message boards : News : Scheduled Maintenance Concluded (Message 65812)
Posted 14 Nov 2016 by Jesse Viviano
Actually, 32-bit applications waste plenty of time in the memory system that 64-bit applications can often avoid because AMD doubled the number of registers in the register files from 8 to 16 while creating AMD64. Some programs will be able to fit all of their speed-critical data into the registers like AQUA@home used to do when it was active with its 64-bit application while the 32-bit version kept having to shuffle data into and out of the memory causing it to be significantly slower. However, the speedup a GPU provides overcomes the problems that 32-bit x86 provides like memory system overhead and conversion of 32-bit calls to OpenCL to 64-bit calls to OpenCL, but every little bit of speed helps.
12) Message boards : News : Scheduled Maintenance Concluded (Message 65717)
Posted 12 Nov 2016 by Jesse Viviano
I think that there is no need to suspend the project while these problems persist. Configure your MilkyWay@home preferences to only accept CPU work and not GPU work.
13) Message boards : News : Scheduled Maintenance Concluded (Message 65703)
Posted 12 Nov 2016 by Jesse Viviano
Thanks. Another speculative possibility that I should have thought of is that the Windows Nvidia GPU client might be running the work units in the CPU instead of the GPU because it is using as much CPU as a single core CPU client.
14) Message boards : News : Scheduled Maintenance Concluded (Message 65697)
Posted 12 Nov 2016 by Jesse Viviano
I am getting a 0% load on my GPU as measured by TechPowerUp GPU-Z that can be downloaded from https://www.techpowerup.com/downloads/. I think that your application is wasting a significant amount of time in between the time each work unit is being processed.
15) Message boards : News : Scheduled Maintenance Concluded (Message 65692)
Posted 12 Nov 2016 by Jesse Viviano
Work units that are generated with the old version of MilkyWay@home that need additional processing are causing errors. See https://milkyway.cs.rpi.edu/milkyway/workunit.php?wuid=1386764939 for an example. It was generated for the old version of MilkyWay@home, and it needs additional runs for validation. Either the new version needs to be able to deal with results from old work units, or the new version should be listed as a new application so that old work units continue to be processed with the old version.
16) Message boards : Number crunching : Massive server issues and Wu validation delays (Message 65544)
Posted 28 Oct 2016 by Jesse Viviano
MilkyWay@home has some tasks that cannot be done on GPUs: the N-Body Simulation tasks. I think that this project should feed those first to CPUs and only send other tasks (both of which can be processed on GPUs) to CPUs if the scheduler runs out of N-Body Simulation tasks.
17) Message boards : Number crunching : Poor likelihood????? (Message 61040)
Posted 10 Feb 2014 by Jesse Viviano
Don't worry about it unless you are getting results getting declared invalid.
18) Message boards : Number crunching : R9 290X (Message 61039)
Posted 10 Feb 2014 by Jesse Viviano
That depends on your R9 290X.

Is it a full-custom card or a reference card with a custom cooler? If so, look for a project that needs lots of integer performance. The cryptocurrency craze shows that this GPU is the integer king. AMD was intending to have high integer performance to speed up video encoding which requires lots of integer performance and some floating point performance, but hit the cryptocurrency market instead.

Is it a reference card with a reference cooler? Either forget about BOINC on this card and replace this card before it burns itself out or water cool the card. The inadequate cooler on this card will run the card at 95 degrees Celsius and throttle the card's performance when that threshold is tripped. AMD stated that this card was designed for high temperatures, but I don't believe it, much as my disbelief at Nvidia's claim that the GTX 480 was designed for high temperatures was validated when GTX 480s started failing according to what I remember reading in forums. If you want to keep the card, water cool it.

In all cases, monitor the GPU and the VRM temperatures. Some custom cards have been shown to cool the VRM adequately only when the fans are allowed to run at performance/uber mode but let the VRM overheat in quiet mode.
19) Message boards : News : Some new work units running (Message 61037)
Posted 10 Feb 2014 by Jesse Viviano
I do not think that anybody can get any of them. When I try to update MilkyWay@home in BOINC, I get a feeder not running error.
20) Questions and Answers : Windows : OpenCL tasks on Haswell? (Message 60792)
Posted 22 Jan 2014 by Jesse Viviano
No. OpenCL below version 2.0 requires that either the developer compile for each GPU vendor's architectures using the GPU vendor's OpenCL library or ship source code that would be used by the platform's compiler. The second approach would be ideal if all operating systems came with a compiler, but most Windows installations do not have any compilers installed. Therefore, you would need the developers to compile using Intel's OpenCL library before you can use your GPU for MilkyWay@home.


Next 20

Main page · Your account · Message boards


Copyright © 2017 AstroInformatics Group