Message boards :
Number crunching :
Don't know where it should go? Stick it here!
Message board moderation
Previous · 1 . . . 14 · 15 · 16 · 17 · 18 · 19 · 20 . . . 147 · Next
Author | Message |
---|---|
Unixchick Send message Joined: 5 Mar 12 Posts: 815 Credit: 2,361,516 RAC: 22 |
Any thoughts on how the deadline dates for the WUs are chosen?? It looks like the APs have a shorter time window, but the blc WUs are variable. I'll get a nice 6 week out pattern going and then it will throw some about 10 weeks out. Just curious. |
RickToTheMax Send message Joined: 22 May 99 Posts: 105 Credit: 7,958,297 RAC: 0 |
WUs that need longer time to compute will have a longer deadline AFAIK, At least if i sort them by Remaining (estimated), the one that takes more time to complete will have a longer deadline |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
I am trying out pure "gpu" processing on a couple of my machines for a few weeks. I want to see if I can come up with a "reliable" formula per Gpu task that gives you an estimate of the RAC based on the cuda core# and gpu task. The two machines I am using both appear to have "stable" RAC's. So they could go down, stay the same or (gulp) go up. Examples: Gtx 750Ti, Windows, gpu task "SOG", RAC#? So far this one looks like it may go down. Gtx 1060 3GB, Linux, gpu task CUDA91, RAC#? So far this one looks like it may stay the same. Anyone else want to join it this (probably useless) research ;) Tom A proud member of the OFA (Old Farts Association). |
RickToTheMax Send message Joined: 22 May 99 Posts: 105 Credit: 7,958,297 RAC: 0 |
I was wondering, sometimes we get very small WUs, and still get full credits.. Idk how regular those are for each host, so that may impact the average i guess |
Bruce Send message Joined: 15 Mar 02 Posts: 123 Credit: 124,955,234 RAC: 11 |
Hi Siran, You might try one of these web pages to try and tame Windows 10. Disable updates Hide updates Hope this helps. Bruce |
Siran d'Vel'nahr Send message Joined: 23 May 99 Posts: 7379 Credit: 44,181,323 RAC: 238 |
Hi Siran, Hi Bruce, I checked those pages and found that the setting to disable driver updates was removed in build 15019, I am at build 17xxx. Just another example of Micro$oft's push to "own" ALL Windows PCs in the universe. They are removing any control we have over our own PCs bit by bit. This is the last Windows OS I will have on this PC. When they start with their pay-as-you-go OS, I'm outta here with Micro$oft. I just hope that the rest of the world will follow suit. But, most users are the type to just use it and not think about the OS and what Micro$oft is doing. When they come out with the "new" OS it'll be installed without any knowledge as an upgrade and then to run the PC, Micro$oft will demand a credit card number to continue. If one is not given, PC won't run. I can really see that scenario happening in 2020 or whenever the "new" OS is to be launched. Have a great day! :) Siran CAPT Siran d'Vel'nahr - L L & P _\\// Winders 11 OS? "What a piece of junk!" - L. Skywalker "Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
I was wondering, sometimes we get very small WUs, and still get full credits.. Excellent question. It has been a topic of discussion for several months. I do know that I have been getting extermely short tasks on both my gpu's and my cpus. They run down in the 5-7 second range. They have been characterized as "potential noise" bombs that may not much or any useful data. As far as I can tell they are assigned "credit" just like the longer tasks and are sent out to at least 1 other Setiathome system for processing. I also think they raise your raw/daily RAC. Since the RAC has a significant time frame, they won't raise the regular RAC much unless a lot of them are processed over an extended period. HTH, Tom A proud member of the OFA (Old Farts Association). |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
@RickToTheMax Congratulations of having the RAC's going up very nicely on your main crunchers. I remember a couple odd weeks ago you were asking :) Tom A proud member of the OFA (Old Farts Association). |
RickToTheMax Send message Joined: 22 May 99 Posts: 105 Credit: 7,958,297 RAC: 0 |
Thanks! Yes it is going up at a nice pace! It should settle somewhere around 120-140k as i estimate my 8core 1080 system to reach 75-80k and my 16core 1060 system 60-65k. I am slowly increasing my overclock on both cards too. The 1060 that i bought used last weekend is reaching nice clocks currently running 2151mhz at only 55c after upping the fan speed a tad! I am tempted to buy a second one but I'll try to resist! Edit: Forgot about my i7 3770 NAS server without gpu that is contributing a little from time to time =) |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
I am trying out pure "gpu" processing on a couple of my machines for a few weeks. e. Interesting. The i7-3770-Gtx 750Ti running Win10 (I haven't checked the video driver yet) is running the stock SOG app. And based on the task manager it is using 15-17% of the cpu. The task manager under Lubuntu claims that each of my gtx 1060's is using 3% but that is both a different OS and an entirely different gpu app. Tom A proud member of the OFA (Old Farts Association). |
RickToTheMax Send message Joined: 22 May 99 Posts: 105 Credit: 7,958,297 RAC: 0 |
But your system with 1060s has way more cpu cores, probably why the usage is much lower! |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14654 Credit: 200,643,578 RAC: 874 |
Be careful of your tools and your terminology. CPUs, or cores? On an 8-core i7, the SoG app using 100% of a single core will register as using 12.5% of the whole CPU. |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
But your system with 1060s has way more cpu cores, probably why the usage is much lower! I have an experiment underway that involves only processing with the gpu's to see what happens to the RAC. I am using cpu cores to drive the gpu apps (SOG or CUDA91). In about 2 weeks or less, I will start running CPU tasks again. The gtx 750Ti box runs a heterogeneous (Seti, Seti Beta, Rosetti, World Community Grid, Einstein at home for gpu backup) mix since it never would be in the running for a top seti producer. I suspect the two Gtx 1060's will stay pretty much where they are which leads to the question of how much further will the CPU tasks push the system? Tom A proud member of the OFA (Old Farts Association). |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
Be careful of your tools and your terminology. CPUs, or cores? Sorry, I was quoting both the total CPU % used as displayed by the Windows 10 Task manager and the process % as displayed by the Task manager for the BOINC processes. In Windows 10, I am not sure if you can get at the per CPU core or logical cpu in the Taskmanager. Yes, I fondly remember the Win7 task manager. Apparently, it is not available in the Resource Monitor either. I know an app named TThrottle used to track the Win7 task manager fairly closely. And it is, in fact, saying that the CPU logical core is running at 100% driving the gpu. Tom A proud member of the OFA (Old Farts Association). |
Brent Norman Send message Joined: 1 Dec 99 Posts: 2786 Credit: 685,657,289 RAC: 835 |
Well Zal, the old P6X58EPRO shouldn't have been so jealous of seeing a new board arrive :\ Got the Asus P9X79-E WS (2011-R1) fired up and no post, swapping cards, PSU, RAM, CPU and nadda. Finally tried an old ATI R7-240 GPU and it posted. Cool I can see the BIOS rev now! Checked Asus and 1704 has "improved compatibility for TitanX and 980" hmm old, but the newest. As before I went though all my 10xx cards I was trying, still nadda. Got in the 'old' pile and popped in a 980 and posted, 750Ti and 780Ti the same. WTF this is a high end ASUS board and all their others seem to have no problems. This thing is driving me nuts! Searching I see others have had problems but no definite solution (that I see), and there has been benchmarks post by 'testers' with this board and 1080s, so it must work. AAAARRRGGGGG |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
Well Zal, the old P6X58EPRO shouldn't have been so jealous of seeing a new board arrive :\ have you tried (just as a test) using something that you know will POST, like the 750ti as the display adapter, and a 10-series card as a secondary card so see if it's visible/useable in the OS? you said the BIOS is the latest already, but have you tried reflashing it anyway? i've has BIOS issues in the past where reflashing it to the same version fixed some issues. Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
Brent Norman Send message Joined: 1 Dec 99 Posts: 2786 Credit: 685,657,289 RAC: 835 |
Interesting idea to try a 10xx card with a 750Ti, I will try that now that I'm awake again, long night it was. I updated the BIOS to the latest 1704 (was 12xx I think), and could certainly re-flash that to. |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
I did a little upgrade on my gaming computer. this machine started as a 1080 build, and became a 1080ti build at the end with hardline tubing and 2x280mm rads all internal to the case. it was a super tight fit. the ASUS z270-G board started having problem with the PCIe ports, and wouldnt even post with something in the second slot, and it was giving computational errors from cards in other slots when running seti or mining, so the board has to be RMAed. I knew i wanted to move to a 1080ti SLI setup, and I've also been eyeing up the impending i7-9700k, so i decided to rebuild the system on z370. using an i3-8100 now just as a placeholder, then when the 9700k is released i'll swap it in. i've also decided to move away from hardline tubing. i did it, it was cool, but it caused too many issues for system maintenance. draining the loop was necessary to reamice nearly anything except the ram, and draining was a nightmare because everything was packed so tight. and i dont think i could have gotten it all in, with hardline, with the pump and rads in the case, that wouldnt be even worse for system maintenance. and i dont think only 2x280mm thin HWL GTS rads would be enough cooling for the 9700k + 1080ti SLI. so now it's soft tubing. and i'll be able to swap the CPU without draining the loop :) Computer parts: EVGA z370 micro i3-8100 (just a placeholder until the 9700k is released) 2x EVGA 1080ti SC2 Hydrocopper 16GB Geil EVO X DDR4-3200 250GB Samsung 960 EVO Nvme with EK heatsink 2x 250GB Samsung 850 EVO SSD (Raind-0 for 500GB games drive) Corsair TX850M PSU Watercool parts: Watercool MO-RA3 360 Pro with 4x Phobya 180x25mm 900 rpm fans Watercool Heatkiller Tube 150 D5 w/Laing D5 pump (full speed, no speed control) Koolance panel mount QD3 quick disconnects mix of bitspower (left over from old build) angle adapters and EK 10/13mm compression fittings Primochill Advanced LRT 3/8"x1/2" in White (internal) and Black (external) Distilled water + Primochill liquid utopia Pics: but sadness, the RGB panel on the bottom card has failed since taking these pics. first it was just blue, then red, and now the green doesn't seem to work either. green works when i put some pressure on the panel, but it's intermittent :(. i really don't want to tear it all apart, and the card is still working properly from a games/compute performance and cooling part, just the RGB isnt working. Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
Thank for the lovely pictures. A proud member of the OFA (Old Farts Association). |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13751 Credit: 208,696,464 RAC: 304 |
A working Apple I computer has just sold at auction. When it originally went on sale back in 1976, the machine sold for $666.66. Yesterday, it was bought with a winning bid of $375,000. Article. Grant Darwin NT |
©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.