Posts by Haris Dublas

21) Message boards : Number crunching : Wow! CPU WUs done in just 160 secs! (Message 41964) Posted 7 Sep 2010 by Haris Dublas Post: As I was browsing through my completed but inconclusive tasks, I found this Anonymous wingman. No GPUs but have 119K RAC. http://milkyway.cs.rpi.edu/milkyway/results.php?hostid=119318 All his tasks that need a quorum are invalid.
22) Message boards : Number crunching : Aaargh! Server out of new work! (Message 41789) Posted 27 Aug 2010 by Haris Dublas Post: Collatz probably went down due to the mass switching of people from MW and DNETC.
23) Message boards : Application Code Discussion : MILKYWAY_NBODY_0.03_WINDOWS_X86_64.EXE is a SpyWare?!!! (Message 41717) Posted 24 Aug 2010 by Haris Dublas Post: You are just assuming that we didnt do a full system scan on our PCs. I even tried the online scanner of Bitdefender. MBAM also didnt detect anything.
24) Message boards : Application Code Discussion : MILKYWAY_NBODY_0.03_WINDOWS_X86_64.EXE is a SpyWare?!!! (Message 41699) Posted 23 Aug 2010 by Haris Dublas Post: PC Tools, Avast and Avira didnt detect anything.
25) Message boards : Number crunching : The app-info.xml to ignore a GPU Core (Message 41051) Posted 24 Jul 2010 by Haris Dublas Post: After making or editing the cc_config (Data folder), clicking the read config file option on the boinc manager menu doesn't always work especially the gpu options. You have to restart boinc client or service. 7/24/2010 12:42:27 PM Starting BOINC client version 6.10.19 for windows_x86_64 7/24/2010 12:42:27 PM Config: use at most 4 CPUs 7/24/2010 12:42:27 PM Config: ignoring ATI GPU 0 7/24/2010 12:42:27 PM log flags: file_xfer, sched_ops, task 7/24/2010 12:42:27 PM Libraries: libcurl/7.19.4 OpenSSL/0.9.8l zlib/1.2.3 7/24/2010 12:42:27 PM Running as a daemon 7/24/2010 12:42:27 PM Data directory: D:\Important\BOINCDATA 7/24/2010 12:42:27 PM Running under account boinc_master 7/24/2010 12:42:27 PM Processor: 4 AuthenticAMD AMD Phenom(tm) FX-5000 Quad-Core Processor [AMD64 Family 16 Model 4 Stepping 2] 7/24/2010 12:42:27 PM Processor: 512.00 KB cache 7/24/2010 12:42:27 PM Processor features: fpu tsc pae nx sse sse2 7/24/2010 12:42:27 PM OS: Microsoft Windows XP: Professional x64 Edition, Service Pack 2, (05.02.3790.00) 7/24/2010 12:42:27 PM Memory: 1.97 GB physical, 5.84 GB virtual 7/24/2010 12:42:27 PM Disk: 100.00 GB total, 16.22 GB free 7/24/2010 12:42:27 PM Local time is UTC +8 hours 7/24/2010 12:42:27 PM ATI GPU 0 (ignored by config): ATI Radeon HD 2300/2400/3200 (RV610) (CAL version 1.4.696, 42MB, 40 GFLOPS peak) 7/24/2010 12:42:27 PM ATI GPU 1: ATI Radeon HD 4700/4800 (RV740/RV770) (CAL version 1.4.696, 512MB, 1064 GFLOPS peak) If only 1 wu is processed at a time then only 1 gpu core is used. If using this option in DNETC, you have to detach and reattach otherwise it will still use all cores per wu.
26) Message boards : Number crunching : nVidia GPU and CUDA 1.3 (Message 41029) Posted 22 Jul 2010 by Haris Dublas Post: Assign a different project prefs for that host and set Collatz to just run nvidia and set MW to only run ati. That way, MW will have the sole use of the ati card and Collatz will only use the nvidia cards.
27) Message boards : Number crunching : possible to use less than 100% of GPU? (Message 41008) Posted 20 Jul 2010 by Haris Dublas Post: You can also try experimenting with kernel frequency. Even with a high wait factor, the games I play while crunching are very sluggish. With a higher kernel frequency even with default wait factor, screen lag in my games was gone even if I increase the gpu load to 100% (w0.8-w0.9). kernel frequency: f (default 30) The application determines the size of the work packages sent to the GPU in a dynamic manner. It tries to get the number of executed packages per second above the value specified here by splitting them to smaller ones. The OS is completely blocked from accessing the graphics card for the duration of one package leading to a somewhat sluggish behaviour of the user interface. Limiting the size and therefore reducing the execution time per package creates more opportunities for the OS to slip in between two work packages and to react to user input. The default value of 30 Hz is tuned for usability of the system. Reducing it could increase the throughput of the app slightly. You can try <cmdline>f40</cmdline> or higher. You can also use both the F and W settings and try different values and see what works for you.
28) Message boards : Number crunching : The app-info.xml to ignore a GPU Core (Message 40742) Posted 30 Jun 2010 by Haris Dublas Post: Glad to be of help. Three of my machines in the house are configured like this. The onboard gpus are used for games, office works, web browsing, watching videos , etc. while the discrete cards are crunching collatz, mw, and dnetc. I can even play games or even online 3d games without lag or stutters. Wife doesn't complain anymore when she play her facebook games.
29) Message boards : Number crunching : The app-info.xml to ignore a GPU Core (Message 40734) Posted 29 Jun 2010 by Haris Dublas Post: Is there a monitor or dummy plug attached to the 2nd gpu? With XP, boinc can crunch on the primary display even without monitor or dummy plug but will not detect the other gpus if there are no monitors or dummy plugs attached on the 2nd/3rd/4th gpu and the desktop extended.
30) Message boards : Number crunching : Collatz - where has it gone? (Message 40689) Posted 26 Jun 2010 by Haris Dublas Post: There was a time when I saw them adding new lines. I asked politely if they can remove those unused wires since one of them is already up the post. They told me that they cant do that because its not included in their job order for that day. They told me that I have to file a complaint at the office so that another job order can be issued for cleaning their own mess. If I had pole long enough at hand, I might have poked him off that post.
31) Message boards : Number crunching : Collatz - where has it gone? (Message 40684) Posted 26 Jun 2010 by Haris Dublas Post: Seems everywhere there are lazy ass linemen. Slicker said that comcast lines are a mess in his area. Its like where I live. My house is near the entrance of the subdivision and the post holding the lines (cable tv and phone) is near my fence almost beside the gate. The linemen just keep adding new cables whenever someone applies for a subscription but neglects removing the old lines when someone stopped their subscription. The cables looks like a bird's nest on that post sometimes the disconnected lines from the distribution boxes hangs over my gate. Nothing was ever done when I complain to the subdivision association so I end up complaining personally to 3 phone companies and 2 cable tv providers almost every year.
32) Message boards : Number crunching : Unable to reattach to MW. (Message 40660) Posted 25 Jun 2010 by Haris Dublas Post: Don't forget to thank Crunch3r guys. :)
33) Message boards : Number crunching : No New Work (Message 40635) Posted 24 Jun 2010 by Haris Dublas Post: Probably due to settings, my 2 machines downloaded tasks immediately as soon as work became available. I was watching when the downloads started because I was about to disable NNT on Collatz. Both machines ran out of work about 1 hour or so before work started flowing again.
34) Message boards : Number crunching : No New Work (Message 40630) Posted 24 Jun 2010 by Haris Dublas Post: Using 6.10.18 and 6.10.19, no problems getting work everytime boinc reports finished tasks automatically or manually.
35) Message boards : Number crunching : GTX 295 (Message 40592) Posted 23 Jun 2010 by Haris Dublas Post: There's a few listed in eBay, a couple are even auctioned.
36) Message boards : Number crunching : The app-info.xml to ignore a GPU Core (Message 40590) Posted 23 Jun 2010 by Haris Dublas Post: put this into cc_config: <ignore_ati_dev>n</ignore_ati_dev> n = device number ex. <ignore_ati_dev>0</ignore_ati_dev> tells boinc to ignore device 0 (primary display) check boinc messages for the correct device numbers. Note: Only works on boinc version 6.10.19 and higher, if using older version try the X in the cmdline (ex. x0 x1 x2 x3) of app_info
37) Message boards : Number crunching : Keeping on board GPU active when adding a video card (Message 40553) Posted 20 Jun 2010 by Haris Dublas Post: To use the onboard gpu alongside the discrete gpu, try enabling SurroundView in the northbridge settings of the bios then select which one will the bios detect first as the primary display. I dont know if the 4200 works with an nvidia card using this setup cause my 785 is paired with an ati (doing collatz on the IGP, MW on the 4850). I have an nvidia card but its a lowly 8500gt and it is paired with an nvidia 8200 board. I can make the nvidia IGP crunch with an ati card but I cant monitor the gpus when I'm crunching on both (i get bsods when I open gpu-z). At dnetc, i have to disable the onboard gpu in the boinc cc_config or else dnetc will grab all gpus available and I end up with messed up times or even errors.
38) Message boards : Number crunching : Not Impressed (Message 40439) Posted 16 Jun 2010 by Haris Dublas Post: The wus are quite small. Only the stars**.txt files are 3mb in size, but those are not downloaded every time boinc connect to the project. Only if its the first time you downloaded those batch. So if you have already crunched de_16 wus before, it will not download the stars-16.txt anymore until the next time you do a project reset. The de_.* files are downloaded in mere seconds just like your example, the wu was downloaded in just 2 secs. Since you are on broadband, a few kb is negligible. Now if you have multiple fast gpus (ie. radeon 5970s), then tasks are downloaded every 1 1/2 mins or so. If you really want to have full control over the downloads and uploads, on the boinc menu under Activity select "Network activity suspended" whenever you are working then un-suspend when you are not working.
39) Message boards : Number crunching : Simultaneously CPU & GPU (Message 40431) Posted 16 Jun 2010 by Haris Dublas Post: Since you are using a 5850, the tasks finish in less than 2 mins so I guess you need a lower setting on the cpu time.
40) Message boards : Number crunching : Simultaneously CPU & GPU (Message 40428) Posted 16 Jun 2010 by Haris Dublas Post: If you set boinc to "Use at most 80-90% CPU time", your gpu times wont be affected much. There's no need to leave a cpu core/thread free. Doing concurrent wus also lessens the cpu usage spikes a little bit, leaving your cpu cores free is just a waste of resources (unless you want to save power consumption or for keeping your overall system temps low).

Previous 20 · Next 20