Message boards :
Number crunching :
Linux CUDA 'Special' App finally available, featuring Low CPU use
Message board moderation
Previous · 1 . . . 6 · 7 · 8 · 9 · 10 · 11 · 12 . . . 83 · Next
Author | Message |
---|---|
Sidewinder Send message Joined: 15 Nov 09 Posts: 100 Credit: 79,432,465 RAC: 0 |
Question, It'll probably depend on what distro you're using. Arch has a pre-compiled boinc in its community repo. You could try manually compiling it? I have not tried this though. http://boinc.berkeley.edu/wiki/Compiling_the_core_client |
Gene Send message Joined: 26 Apr 99 Posts: 150 Credit: 48,393,279 RAC: 118 |
@Brent ... I'm using the Debian (Linux) distro and a couple of months ago I got BOINC 7.6.22 (and boincmgr) from their "testing" repository. It is now 7.6.33 . That version will likely migrate to the "stable" release around the first of April. Other Linux distros will follow their own update schedule. The downside of running a "testing" application is that some of its required libraries will differ from the (already installed) "stable" versions. It can make a mess of things to try to install/upgrade the "testing" libraries so the safe approach is to load such libraries into their own directory, leaving all the rest of the installed system unchanged, and point the loader to that directory only when starting the boinc, and boincmgr, applications. If you're not quite ready to be "system administrator" to that degree then stick with the current distro version (for whatever flavor of Linux you choose) and be alert to update BOINC whenever later versions are available. |
Brent Norman Send message Joined: 1 Dec 99 Posts: 2786 Credit: 685,657,289 RAC: 835 |
Thanks guys, I will do some more reading on compiling from source. |
Michel Makhlouta Send message Joined: 21 Dec 03 Posts: 169 Credit: 41,799,743 RAC: 0 |
Hi, I have slackware 14.2 up and running with boinc 7.6.33 and already crunching (CUDA60 on linux compared to SoG?). Can you send me the latest special app? and some high level instructions :) |
Michel Makhlouta Send message Joined: 21 Dec 03 Posts: 169 Credit: 41,799,743 RAC: 0 |
i just noticed i have cuda60, opencl SoG an opencl sah in the queue. |
rob smith Send message Joined: 7 Mar 03 Posts: 22200 Credit: 416,307,556 RAC: 380 |
During the first few weeks of operation you will find various (GPU) applications will run. This is done to establish which one works best on your system. You will not be "offered" the "Petri/TBar Special" application by the servers as this still hasn't been declared fit for public consumption since it (still) gives a very high rate of "inconclusive" results. Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
Michel Makhlouta Send message Joined: 21 Dec 03 Posts: 169 Credit: 41,799,743 RAC: 0 |
in that case, i might stick to whatever the server offers... does the same apply for linux when it comes to commandline & # of WUs? on win10, i was running commandline and 2WU, should i be replicating that for cuda60 (assuming it ends us being what i run)? |
Brent Norman Send message Joined: 1 Dec 99 Posts: 2786 Credit: 685,657,289 RAC: 835 |
Sure you can use the 2WU/GPU w/CommandLine like you did in Widows, but if/when you install the 'special' app, it is designed to make use of the entire GPU resources and therefore only 1 task per card. |
Michel Makhlouta Send message Joined: 21 Dec 03 Posts: 169 Credit: 41,799,743 RAC: 0 |
out of curiosity, which programming language are people using to create their own special apps? is it a matter of improving code or optimizing during compiling? |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
out of curiosity, which programming language are people using to create their own special apps? is it a matter of improving code or optimizing during compiling? Primarily hand coded optimisations in both the GPU Cuda/OpenCL and Host CPU C/C++ code. In the case of Cuda special apps, that tends to be a combination of higher level structural optimisations, such as adapting to use Cuda Streams, along with low level PTX assembly (&/or intrinsics). The higher level structural enhancements tend to allow better scaling on newer hardware, while the low level assembly tends to improve memory access patterns and instruction level paralellism. Compiler options tended to decrease in importance after about Cuda 4, since nVidia switched to an LLVM based compiler. That relatively modern type of compiler tends to do a reasonable job without much twiddling. "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
petri33 Send message Joined: 6 Jun 02 Posts: 1668 Credit: 623,086,772 RAC: 156 |
PLEASE DO FORGET THE TEXT BELOW. Seems to have an error still in the gasuss-finding. Do not try this version. Sorry. For those who are running any of my older Linux versions I'd strongly suggest that you use this new version. Version zi3o for sm_35, sm_50, sm_52 and sm_61 ====================================== The executable: https://drive.google.com/open?id=0B9PYeBxtfMjaZVoxY3hXWkh5aDg The source: https://drive.google.com/open?id=0B9PYeBxtfMjaT3paRFF5anpQT28 command line options: 1) To reduce CPU usage (possibly at the cost of performance) use option -bs meaning blocking sync. 2) To process guppi/vlar faster use -unroll N together with -pfp N. N is the number of SM/SMX units on your (lest capable) NVIDIA GPU on your machine. Too high N may cause you 'out of GPU memory error'. For N the GTX780 has 12 SM, GTX980 has 16 SMX and GTX1080 has 20 SMX. 3) You can also try -pfb 8 or higher. The default is 4. Petri To overcome Heisenbergs: "You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones |
petri33 Send message Joined: 6 Jun 02 Posts: 1668 Credit: 623,086,772 RAC: 156 |
PLEASE DO FORGET THE TEXT BELOW. There'll be more in beta until it (the gaussian search) is fixed. All my offline benchmarks finish 'clean'. To overcome Heisenbergs: "You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones |
Brent Norman Send message Joined: 1 Dec 99 Posts: 2786 Credit: 685,657,289 RAC: 835 |
You know you are running Beta tasks when .... Cuda60 tasks have been running over 1 hour! |
Brent Norman Send message Joined: 1 Dec 99 Posts: 2786 Credit: 685,657,289 RAC: 835 |
I found something for those with temperature issues in Ubuntu (likely others) sudo nvidia-xconfig --thermal-configuration-check --cool-bits=4 --enable-all-gpus *then reboot coolbits - adds manual fan control to NVIDIA control panel all gpus - more than 1 thermal check - not sure what if anything that did/does **Read manual** ..... man nvidia-xconfig |
petri33 Send message Joined: 6 Jun 02 Posts: 1668 Credit: 623,086,772 RAC: 156 |
I found something for those with temperature issues in Ubuntu (likely others) With cool-bits you gain access to CPU fan, GPU and VRAM clock offset etc. Just use google. Remember the --enable-all-gpus (-a ?) if you have more than one. Here is an example (edit it to suit your needs/get rid of warnings) script to run after a reboot: /usr/bin/nvidia-smi -pm 1 #gtx 1080s nvidia-settings -a "[gpu:0]/GPUPowerMizerMode=1" nvidia-settings -a "[GPU:0]/GPUOverVoltageOffset=16000" /usr/bin/nvidia-settings -a "[gpu:0]/GPUMemoryTransferRateOffset[3]=1100" -a "[gpu:0]/GPUGraphicsClockOffset[3]=190" nvidia-settings -a "[gpu:0]/GPUFanControlState=1" -a "[fan:0]/GPUTargetFanSpeed=96" /usr/bin/nvidia-smi -i 0 -pl 215 /usr/bin/nvidia-smi -i 0 -ac 5005,1911 nvidia-settings -a "[gpu:1]/GPUPowerMizerMode=1" nvidia-settings -a "[GPU:1]/GPUOverVoltageOffset=16000" /usr/bin/nvidia-settings -a "[gpu:1]/GPUMemoryTransferRateOffset[3]=1100" -a "[gpu:1]/GPUGraphicsClockOffset[3]=190" nvidia-settings -a "[gpu:1]/GPUFanControlState=1" -a "[fan:1]/GPUTargetFanSpeed=96" /usr/bin/nvidia-smi -i 1 -pl 215 /usr/bin/nvidia-smi -i 1 -ac 5005,1911 nvidia-settings -a "[gpu:2]/GPUPowerMizerMode=1" nvidia-settings -a "[GPU:2]/GPUOverVoltageOffset=16000" /usr/bin/nvidia-settings -a "[gpu:2]/GPUMemoryTransferRateOffset[3]=1100" -a "[gpu:2]/GPUGraphicsClockOffset[3]=190" nvidia-settings -a "[gpu:2]/GPUFanControlState=1" -a "[fan:2]/GPUTargetFanSpeed=100" /usr/bin/nvidia-smi -i 2 -pl 215 /usr/bin/nvidia-smi -i 2 -ac 5005,1911 nvidia-settings -a "[gpu:3]/GPUPowerMizerMode=1" nvidia-settings -a "[GPU:3]/GPUOverVoltageOffset=16000" /usr/bin/nvidia-settings -a "[gpu:3]/GPUMemoryTransferRateOffset[3]=1100" -a "[gpu:3]/GPUGraphicsClockOffset[3]=180" nvidia-settings -a "[gpu:3]/GPUFanControlState=1" -a "[fan:3]/GPUTargetFanSpeed=96" /usr/bin/nvidia-smi -i 3 -pl 215 /usr/bin/nvidia-smi -i 3 -ac 5005,1911 # gtx980s #/usr/bin/nvidia-smi -i 0 -pl 230 #/usr/bin/nvidia-smi -i 2 -pl 230 #/usr/bin/nvidia-settings -a "[gpu:1]/GPUMemoryTransferRateOffset[3]=200" -a "[gpu:1]/GPUGraphicsClockOffset[3]=30" #/usr/bin/nvidia-settings -a "[gpu:2]/GPUMemoryTransferRateOffset[3]=200" -a "[gpu:2]/GPUGraphicsClockOffset[3]=20" #/usr/bin/nvidia-smi -i 0 -ac 3605,1321 #/usr/bin/nvidia-smi -i 2 -ac 3605,1324 #/usr/bin/nvidia-settings -a "[GPU:1]/GPUOverVoltageOffset=16000" #/usr/bin/nvidia-settings -a "[GPU:2]/GPUOverVoltageOffset=16000" #nvidia-settings -a "[gpu:1]/GPUFanControlState=1" -a "[fan:1]/GPUTargetFanSpeed=90" #nvidia-settings -a "[gpu:2]/GPUFanControlState=1" -a "[fan:2]/GPUTargetFanSpeed=90" // 12 virtual CPUs, distribute the work ... for (( ; ; )) do schedtool -a 1,2,3,4 `pidof setiathome_x41zc_x86_64-pc-linux-gnu_cuda65` schedtool -a 1,2,3,4 `pidof setiathome_x41zc_x86_64-pc-linux-gnu_cuda65_v8` schedtool -a 1,2,3,4 `pidof ap_7.01r2793_sse3_clGPU_x86_64` schedtool -a 6,7,8,9,10,11 `pidof MBv8_8.05r3345_avx_linux64` schedtool -a 6,7,8,9,10,11 `pidof setiathome_8.04_i686-pc-linux-gnu` schedtool -a 6,7,8,9,10,11 `pidof ap_7.05r2728_avx_linux32e` schedtool -a 5 `pidof compiz` sleep 2 rmdir ~petri/Downloads/BOINC/slots/1?* 2>/dev/null rmdir ~petri/Downloads/BOINC/slots/2?* 2>/dev/null rmdir ~petri/Downloads/BOINC/slots/3?* 2>/dev/null rmdir ~petri/Downloads/BOINC/slots/4?* 2>/dev/null rmdir ~petri/Downloads/BOINC/slots/5?* 2>/dev/null rmdir ~petri/Downloads/BOINC/slots/6?* 2>/dev/null rmdir ~petri/Downloads/BOINC/slots/7?* 2>/dev/null rmdir ~petri/Downloads/BOINC/slots/8?* 2>/dev/null rmdir ~petri/Downloads/BOINC/slots/9?* 2>/dev/null and the xorg.conf with cool-bits "28" looks something like this: @Linux1:~/Downloads/BOINC# cat /etc/X11/xorg.conf # nvidia-xconfig: X configuration file generated by nvidia-xconfig # nvidia-xconfig: version 378.09 (buildmeister@swio-display-x86-rhel47-04) Sat Jan 14 22:12:34 PST 2017 Section "ServerLayout" Identifier "Layout0" Screen 0 "Screen0" Screen 1 "Screen1" RightOf "Screen0" Screen 2 "Screen2" RightOf "Screen1" Screen 3 "Screen3" RightOf "Screen2" InputDevice "Keyboard0" "CoreKeyboard" InputDevice "Mouse0" "CorePointer" EndSection Section "Files" EndSection Section "InputDevice" # generated from default Identifier "Mouse0" Driver "mouse" Option "Protocol" "auto" Option "Device" "/dev/psaux" Option "Emulate3Buttons" "no" Option "ZAxisMapping" "4 5" EndSection Section "InputDevice" # generated from default Identifier "Keyboard0" Driver "kbd" EndSection Section "Monitor" Identifier "Monitor0" VendorName "Unknown" ModelName "Unknown" HorizSync 28.0 - 33.0 VertRefresh 43.0 - 72.0 Option "DPMS" EndSection Section "Monitor" Identifier "Monitor1" VendorName "Unknown" ModelName "Unknown" HorizSync 28.0 - 33.0 VertRefresh 43.0 - 72.0 Option "DPMS" EndSection Section "Monitor" Identifier "Monitor2" VendorName "Unknown" ModelName "Unknown" HorizSync 28.0 - 33.0 VertRefresh 43.0 - 72.0 Option "DPMS" EndSection Section "Monitor" Identifier "Monitor3" VendorName "Unknown" ModelName "Unknown" HorizSync 28.0 - 33.0 VertRefresh 43.0 - 72.0 Option "DPMS" EndSection Section "Device" Identifier "Device0" Driver "nvidia" VendorName "NVIDIA Corporation" BoardName "GeForce GTX 1080" BusID "PCI:10:0:0" EndSection Section "Device" Identifier "Device1" Driver "nvidia" VendorName "NVIDIA Corporation" BoardName "GeForce GTX 1080" BusID "PCI:9:0:0" EndSection Section "Device" Identifier "Device2" Driver "nvidia" VendorName "NVIDIA Corporation" BoardName "GeForce GTX 1080" BusID "PCI:6:0:0" EndSection Section "Device" Identifier "Device3" Driver "nvidia" VendorName "NVIDIA Corporation" BoardName "GeForce GTX 1080" BusID "PCI:5:0:0" EndSection Section "Screen" Identifier "Screen0" Device "Device0" Monitor "Monitor0" DefaultDepth 24 Option "Coolbits" "28" SubSection "Display" Depth 24 EndSubSection EndSection Section "Screen" Identifier "Screen1" Device "Device1" Monitor "Monitor1" DefaultDepth 24 Option "Coolbits" "28" SubSection "Display" Depth 24 EndSubSection EndSection Section "Screen" Identifier "Screen2" Device "Device2" Monitor "Monitor2" DefaultDepth 24 Option "Coolbits" "28" SubSection "Display" Depth 24 EndSubSection EndSection Section "Screen" Identifier "Screen3" Device "Device3" Monitor "Monitor3" DefaultDepth 24 Option "Coolbits" "28" SubSection "Display" Depth 24 EndSubSection EndSection To overcome Heisenbergs: "You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
Hi, . . OK, I am pretty much a raw Linux novice. I have been looking for a flavour of Linux to use. I thought there were few and only two desktop choices (I haven't even looked at Linux for about 12 years). But I have found the choice is very confusing, being faced now with not just Gnome or Kde but also Kfce, cinamon and other third party versions. Not to mention nearly a dozen implementations of Ubuntu itself, Kubuntu, Lubuntu and whatever else someone has dreamt up. . . So what are the recommendations for a) desktop and b) install version of Ununtu? I don't think I want Edubuntu or Lubuntu or Mythbuntu. I thought Kde was the desktop to go for but what is LXDE? Or should I still consider Gnome? . . I am working from a Win10 machine to create a "Boot Disk" on a Sandisk Ultra Fit 16GB flash drive. What do I need to do to turn the ISO into a bootable USB? Most references are by Linux gurus and couch everything in Linux terminology. And what format does Linux like on the USB drive? Is it happy with NTFS or does it have to be Fat32? . . The new rig is not put together yet but I would like to have the OS sorted out before it is. . . Any help will as always be greatly appreciated. Stephen ?? [Edit - PS - It is going on an i5-6600 with 16GB DDR4-2133 Ram, Intel or Samsung NVMe SSD boot drive, and 2 x GTX970s. Hopefully this hardware is all compatible with Linux] . |
[AF>EDLS]GuL Send message Joined: 15 Feb 06 Posts: 10 Credit: 27,125,503 RAC: 0 |
So what are the recommendations for a) desktop and b) install version of Ununtu? Hi Stephen, From my point of view, the easiest version to use and configure is Ubuntu with Gnome. I really not recommend Fedora, which I am using, because it is too hard to configure. Moreover, Ubuntu has a very important and reactive community. In order to burn your USB key, I will suggest you to use rufus and keep default options. Cheers |
tazzduke Send message Joined: 15 Sep 07 Posts: 190 Credit: 28,269,068 RAC: 5 |
Greetings Stephen Yes there are many flavours, but after trying a few, ie, Debian OS, Mint KDE, Mint Ubuntu (Mate and Cinnamon) and a couple of other flavours that have dropped out of the Distrowatch Top 10. Back on track, I used a program called Universal USB Installer, to make a bootable usb Linux installer, which I found at https://www.pendrivelinux.com/universal-usb-installer-easy-as-1-2-3/ Instructions are on the website. Personaly, I use Mint 18.1 Cinnamon edition, (I have it on a spare HDD for when I want to try it out again) but that is my choice. (I am still having trouble breaking away from Windows 7) lol. Another thing I learnt is that is when you have two NVIDIA cards in the same system, only connect the 2nd once your system up and running with the NVIDIA drivers. Have fun. Mark PS, I ended up with 2 x GTX 960's and a GTX 780 in my systems, the 1050's didnt make it lol. Oh the GTX 580 is a part timer cruncher nowadays. |
Brent Norman Send message Joined: 1 Dec 99 Posts: 2786 Credit: 685,657,289 RAC: 835 |
Making a bootable USB from an ISO seems pointless to me (unless you don't have a DVD) just burn the ISO to a DVD, it's just the install disk (and usually a test run of the package). You won't be able to save anything anyways until you actually install the OS to a drive or USB -after you have a bootable disk made from an ISO. Grab any old drive, or USB to test it out on your current hardware with a dual boot. And yea, 16G is enough. The main OS's people are using seem to be Ubuntu or Mint (keep away from v16, use 14 ... I think it is the same for Mint) Crimson seems to be a Mint choice, I used, LOL I forget for Ubuntu. It was ubuntu-14.04.5-desktop-amd64.iso Google will find it. The OS itself is very simple to get use to, but for me not as a daily computer. Sorry I just don't care for the MAC OSX look. The ONLY thing I don't like about Ubuntu is the BOINC version 7.2 which doesn't have process priority control so you have to be careful not to work the CPU too hard and starve the GPUs. So yea I could probably get more out of Mint, which I do plan on at least testing. It's easy to install and get running on SETI 'stock' in probably a little over 30 minutes without having to know anything about computers. From there it's manually adding the app files, setting permission, make an app_info, and your off. Not really a big deal. Ohh if it's a cruncher only, set auto login at install - it's just easier. |
tullio Send message Joined: 9 Apr 04 Posts: 8797 Credit: 2,930,782 RAC: 1 |
I am running SuSE Leap 42.2 on an old SUN WS with only 8 GB RAM and both SETI and SETI Beta run on it. There is a GTX 750 GPU board and its driver has recently been updated by SuSE, which frequently sends me updates asking me if I want to install them, not like Microsoft on my Windows 10 PC where I have no control over updates, except the GPU board updates, which I get from Geforce. It is a Palit GTX 1050. Tullio |
©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.