Message boards :
Number crunching :
Getting the most Seti cpu production out of your Amd 7 2600/2700 cpu
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · 6 · Next
Author | Message |
---|---|
Keith Myers Send message Joined: 29 Apr 01 Posts: 13161 Credit: 1,160,866,277 RAC: 1,873 |
Or as Zalster has commented or recommended, just set a max_concurrent statement in app_config so that the total cpu usage is 65-90%. I like around 65% usage as that keeps cpu_time within a couple of minutes of run_time. Follow Juan's formula for a spoofed client, add the desired number of cpu tasks wanted to run to the total amount of spoofed gpus to arrive at the designated max_concurrent statement. Seti@Home classic workunits:20,676 CPU time:74,226 hours 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 |
Or as Zalster has commented or recommended, just set a max_concurrent statement in app_config so that the total cpu usage is 65-90%. I like around 65% usage as that keeps cpu_time within a couple of minutes of run_time. Follow Juan's formula for a spoofed client, add the desired number of cpu tasks wanted to run to the total amount of spoofed gpus to arrive at the designated max_concurrent statement. I had this running previously: <app_config> <project_max_concurrent>48</project_max_concurrent> </app_config> Which is the gpus 6-> 32 spoofed setup. And with 100% of the cores/threads enabled in Boinc Manager, the RAC graph was dead flat. I have just re-patched to this: app_config> <project_max_concurrent>48</project_max_concurrent> <app> <name>setiathome_v8</name> <gpu_versions> <gpu_usage>1.0</gpu_usage> <cpu_usage>1.0</cpu_usage> </gpu_versions> </app> <app> <name>astropulse_v7</name> <gpu_versions> <gpu_usage>0.5</gpu_usage> <cpu_usage>1.0</cpu_usage> </gpu_versions> </app> </app_config> I now have the above with 90% threads running. 6 gpus + 8 cpus. This is the stock setup I have been using. If the RAC graph curves into a steep dive tonight, then we are going to have to come up with a compromise of some type because some over-committing of the cpus may increase total production. Tom A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13161 Credit: 1,160,866,277 RAC: 1,873 |
Do you really expect RAC to change that fast????? I have never seen RAC change overnight. More like a couple of days if that fast/ Seti@Home classic workunits:20,676 CPU time:74,226 hours 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 |
When one of my two biggest boxes crashes and I don't catch it, it takes 5,000-10,000 RAC off the top. But that is when it is down for hours or even overnight. As soon as I tried Tbars proposal, the RAC that had been sliding lower and lower suddenly became nearly flat. That was before I upgraded the cpu cooler and started running 4.0Ghz on the cpu. If there is a massive reduction in total put through then I expect the graph to drop a lot. But if it is "dinky" reduction then like you said it can take days...... The processing times of the cpu tasks track nicely with the level over commitment. So the question is where is the "optimum" over commitment that produces the highest total production. I expect the cpu processing times to mostly return to the 47-57 or less minute range now. Right now many of the cpu tasks are taking over 2 hours. So what I am looking for is more than 8 cpu threads but less than 14 threads that slows the cpu processing to maybe an hour and produces greater total output. Tom A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13161 Credit: 1,160,866,277 RAC: 1,873 |
My rule of thumb is keep the times within a couple of minutes of each other. Ten minutes apart is wasted efficiency for the day because you would do more work with less tasks running. I can see the effect very easily in the BoincTasks reported values per day and per week for each host on the main page of BT. You can see it the BT graphs very readily where peak production is located. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
As soon as I tried Tbars proposal, the RAC that had been sliding lower and lower suddenly became nearly flat. That was before I upgraded the cpu cooler and started running 4.0Ghz on the cpu. I wouldn't read into it that much, because around that same time everyone else's RAC started climbing from the different WU diversity and earning more credit with changing Nothing. You went from negative to flat, while everyone else went from flat to positive. There's more variables to your RAC than just what settings you are using, some you can't control. Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13720 Credit: 208,696,464 RAC: 304 |
There's more variables to your RAC than just what settings you are using, some you can't control.Only some? Your wingman & their hardware, their application, the WU being processed, Credit New and what it does to determine Credit, all of those combined in all possible permutations. Grant Darwin NT |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
very true Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
There's more variables to your RAC than just what settings you are using, some you can't control.Only some? Not forget the most important player, SETI credit system is ruled by Credit Screw, no matter what you do, at the end he and his upside down random number generator will give you less credit than expected. If was possible and you could crunch the exact set of WU a day after the other, each day it wil give you a diferent credit, simple like that. |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
All that said, my RAC graph is curving up this morning..... :) The cpu MB tasks whose naming convention I recognize have reverted to under 1 hour processing times. The "other" MB tasks are still taking longer. And the wallclock time vs. cpu time are once again tracking nearly together..... 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 |
I ran across this review of the Amd 2700 vs. the Amd 2700x..... and it explains my OC issues. It was the Amd stock cooler..... https://www.tomshardware.com/reviews/ryzen-7-2700-2700x-review,5606.html The 2700 stock cooler is a LOT less competent than the 2700x stock cooler. Article recommended an after market cpu cooler for the 2700..... Tom A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13161 Credit: 1,160,866,277 RAC: 1,873 |
I can't remember, what did you end up with for an aftermarket cooler . . . . other than some "big-ass air cooler" Seti@Home classic workunits:20,676 CPU time:74,226 hours 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 |
I can't remember, what did you end up with for an aftermarket cooler . . . . other than some "big-ass air cooler" The Noctura that you recommended and posted a link on. 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 |
All that said, my RAC graph is curving up this morning..... :) Yes it is still going up. OBTW, I only keep last 90 days of report info set in cc_config.xml and it seems to have made my RAC more response. A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13161 Credit: 1,160,866,277 RAC: 1,873 |
All that said, my RAC graph is curving up this morning..... :) That setting only applies to the graphing of statistics in the Manager. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Kissagogo27 Send message Joined: 6 Nov 99 Posts: 715 Credit: 8,032,827 RAC: 62 |
rac rac rac ... http://www.teamocuk.co.uk/cpartcred2.php?s=e814c4fa2914443ac95b714dc911e936&p=SAH&u=190117 not so flat ^^ |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13161 Credit: 1,160,866,277 RAC: 1,873 |
rac rac rac ... http://www.teamocuk.co.uk/cpartcred2.php?s=e814c4fa2914443ac95b714dc911e936&p=SAH&u=190117 not so flat ^^ Interesting plot. Never seen that one before. Mine is a general downslope which doesn't match the trend in BOINC Manager statistics graphs. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Wiggo Send message Joined: 24 Jan 00 Posts: 34744 Credit: 261,360,520 RAC: 489 |
Now Tom you just have to do the same with your Ryzen 5 2600 to get its CPU tasks run time vs. CPU time closer together. ;-) Cheers. |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
Now Tom you just have to do the same with your Ryzen 5 2600 to get its CPU tasks run time vs. CPU time closer together. ;-) I would agree except it is about to get back up to its all-time-high RAC score and I am reluctant to touch it yet. I am running a 0.50 cpus to gpu since it is also a 6c/12t setup.... I would like the upward slop to show signs of "petering out" or the end of the WOW contest before I make many changes. Since this is the system rack that will get the Intel upgrade to the server motherboard it is going to be changing no matter what. I probably won't have enough gpus leftover to fill it completely up in a regular computer case Infact it may end up with a gtx 750ti because I don't have anything else left..... Tom A proud member of the OFA (Old Farts Association). |
rob smith Send message Joined: 7 Mar 03 Posts: 22160 Credit: 416,307,556 RAC: 380 |
Now, for the next MONTH at least sit back and watch that computer. Don't change anything, just monitor it and let it run. Doing so will get you a half decent baseline from which to see what effect changes have. Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
©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.