Posted 11 Jan 2021 by
marmot
Depends on how much CPU cache is used. On a ryzen 3950x there is a limit of 10 then it will be run slower.
3950x has 64mb lvl 3 so we should set the thread count to use no more than about 6mb per thread?
My older server Xeons only have 20MB (still better than my 2700X GPU host) so will have to limit the number of long2 WU's.
1 WU at 3 threads per CPU; currently they are running 2 WU at 6 threads per CPU package. (there's probably a slowdown of all other project WU's as they compete for cache w/ the long2's)
Are all the WU's needing this amount of cache?
If you have some data about max cache per WU type; I'll put it into a spreadsheet chart.
Posted 10 Jan 2021 by
marmot
Noticed my machines were running cooler than expected and checked the process manager and saw the long2's are only using 5 threads for long periods. They can drop down to 2 threads and increase thread count but never seem to ever make it to 6.
They are set to:
<!-- srbase11 = Sierpinski / Riesel Base - long2 -->
<app>
<name>srbase11</name>
<max_concurrent>1</max_concurrent>
<fraction_done_exact/>
</app>
<app_version>
<app_name>srbase11</app_name>
<avg_ncpus>6</avg_ncpus>
<cmdline>-t6</cmdline>
</app_version>
Is optimal 4 threads or less to maximize CPU usage?
I ran over deadlines leaving these long2's at 1 or 2 threads.
Posted 21 Dec 2020 by
marmot
Task from list are purged when work unit are validated
@rebirther
what's delay before a validated WU is purged from a clients WU list?
I'm guessing 16-18 hours since all WU from yesterday are now purged from the TF machines' current work database but yesterday the last 90 TF WU's of the day (11-12 mins each so appr 16-18 hours worth) were visible.
Is purge happening once a day in a database cleanup?
Posted 20 Dec 2020 by
marmot
Boy, points sure went to shit the past day or 2.
Something odd in the points awarding is happening.
See my message:
http://srbase.my-firewall.org/sr5/forum_thread.php?id=1498&postid=7130#7130
UPDATE: I was seeing a WU results delay and likely not related to your concern.
Posted 20 Dec 2020 by
marmot
Task from list are purged when work unit are validated but list work done is stored in account for each application. As it looks now you would see:
Sierpinski / Riesel Base credit 834,753 total (5930 tasks)
.
-
.
TF credit 180,000 total (100 tasks)
This stored on your account and not disappear.
I appreciate your response but been doing work at SRBase and do know how that works. I posted because this seemed abnormal to my last 3 years of service here.
I was seeing a delay of about 3 hours in the account WU database.
Went and checked the client event log before going to work and it showed 60+ tasks done but my account here only showed 5 TF done, in total, a few minutes after I posted my previous message. (It shows 89 completed atm and w/o a delay I should have seen 60+ when starting this thread).
Not sure if the reporting delay has always existed or is just a temporary delay.
UPDATE: I checked the time stamps of the current WU's and one complete at 21:32 UTC (event log on client) and shows in progress here on the account. The last one to complete shows 21:22 and shows completed in my account WU database so the delay is approximately 7-8 minutes.
The 3 hour delay appears to have been a glitch of yesterday.
Posted 19 Dec 2020 by
marmot
Normally not, you have 127k credits and 71 WUs are done.
Huh, I can only see 2 completed and 1 in progress from my tasks management.
Are these bundled somehow?
EDIT: Oh it's up to 4 finished 1 in progress.
Is there some reporting delay?
EDIT2: Got to get ready for work. Have a good weekend.
Posted 19 Dec 2020 by
marmot
Boy, points sure went to shit the past day or 2.
Something odd in the points awarding is happening.
See my message:
http://srbase.my-firewall.org/sr5/forum_thread.php?id=1498&postid=7130#7130
Posted 19 Dec 2020 by
marmot
I shouldn't look a gift horse in the mouth but somehow I have the 100k silver TF badge and only completed 2 WU for 1800 points each.
GPUGrid was down and the TF's beat Einstein@Home to the GTX1060.
I'll stop doing GPUGrid after 2 more days (will hit 1000 CPU hours) and take up TF full time for badge and 1000 CPU hours (or summer heat returns) but didn't expect a silver badge so soon.
The GPU host machine's PS died months back and got it working last month, and I just allowed SRBase work units last night, so it's not possible the machine ran TF WU's before last night.
EDIT: Free-DC shows me getting 200k points today so maybe the problem is in the individual WU reporting DB?
Posted 6 Mar 2020 by
marmot
Been sick last 4 weeks and didn't notice.
BOINCStats shows me losing 816,352 on Feb 22.
https://www.boincstats.com/stats/-1/user/detail/272536/charts
BTW, I'll hit 100k total hours put into long-1 WU's in 2 days.
Posted 22 Nov 2019 by
marmot
hmmm?
AMD FX-8350 66 °C
Max safe operating temp for that CPU is 62C.
So you need to run it cooler.
https://safetemp.blogspot.com/2015/12/amd-fx-series-fx-8350-max-temp.html
My 1090T also would run at 65C and pass AMD Overdrive stress test for an hour but then it would fail SRBase WU's at high rates and this project forced me to redo my 1090T's overclock.
SRBase WUs run the CPU hotter than most all other BOINC WU's I've tested and push the CPU thermal limits.
So try to run at slightly lower voltage if you wish to run SRBase. It's more cost efficient to run downvolted with slower results but much less electricity used.
Besides, with a climate crisis upon humanity, it's not a great idea to run massively inefficient overclocks. I've downvolted my 1090t as far as possible and so it can still host my GPU crunchers till I can afford an new build.
https://www.neoseeker.com/Articles/Hardware/Reviews/amd_phenom_x6_1090t/4.html
An 8th gen Intel i5 8250u has as many threads, can do about the same amount of SRBase work and runs at 15 watts at max load and I've found laptops with it going for under $300 used. The FX-8350 draws 92W when it's doing no work at all...
https://www.tomshardware.com/reviews/fx-8350-vishera-review,3328-16.html
Posted 28 May 2019 by
marmot
<--- Oil painting. Majestic Alpine Marmot Scans His Domain.
Breaks the 4k rule.
Posted 27 May 2019 by
marmot
The maximum size is 100x100, your pic is 97x110
Someone may wonder why I didn't see the 100x100 size limit or the Gravatar link.
It's because there are two locations to upload a profile avatar and I rarely edit the Community Preferences section.
I'd always uploaded my avatar picture from the Edit Your Profile section:
http://srbase.my-firewall.org/sr5/create_profile.php
Posted 18 May 2019 by
marmot
Uploaded the smaller avatar and downloaded it from my own page and it's now the right size. Let's see if it's working.
Success!
Or try to upload the file to gravatar.com
Gravatar's tracking of my movements is not acceptable to me.
Especially points 1, 4 and 5:
1. The entire reason Gravatar offers their service is to collect internet usage data across multiple sites. It is not offered free out of the goodness of their heart. The entire purpose of the service is to analyse the way YOU navigate the internet.
4. The most egregious part of Gravatar’s service is the inability to stop them from collecting your data. I have in the past tried to cancel a Gravatar registration. Gravatar does not allow this and will continue to track your e-mail address for the rest of time.
5. Gravatar does not provide any details about how they use your personal information and does not respond to any queries relating to privacy issues.
https://meta.stackexchange.com/questions/44717/is-gravatar-a-privacy-risk
Posted 15 May 2019 by
marmot
Edited down to 24kb, 87x99 pix jpg.
Deleted the old pic and confirmed no avatar was showing then uploaded the 87x99 version.
Let's see if it shows on this post (should show on all past posts).
Maybe there is a server update delay period.
EDIT: Downloaded the avatar pic, from the browser, currently showing on my profile page, edited it, it's the 97x110 version even though I deleted it.
I'll delete the profile pic and leave it deleted for a day then upload the new version tomorrow.
Posted 13 May 2019 by
marmot
The maximum size is 100x100, your pic is 97x110
Ah, I'll edit it.
Thanks.
Posted 12 May 2019 by
marmot
Used the 25k jpg pic that works elsewhere but my posts still have no avatar.
Posted 12 May 2019 by
marmot
this is a general boinc problem, which doesn't only show up here at srbase.
Yes, seen BOINC client do this many times this week, alone.
It seems to do it when there are two WU that are not going to meet their deadlines so it overbooks the cores.
There is a grace period on SRBase WU deadlines; don't worry if a WU, which has progress, is past the deadline.
Posted 25 Apr 2019 by
marmot
Longs really shouldn't be run on a 2 core machine, which would generally be 8 year old Core 2 Duos, so excluding them would be setting an appropriate boundary.
Most modern Core i3 CPUs are 2c/4t.
All but the highest end laptops tend to be 2c/4t. Most Mobile Core i5 CPUs are 2c/4t, unlike their desktop counterparts. Unless I'm mistaken, many mobile Core i7s are also 2c/4t.
You don't have to go back 8 years to find dual core CPUs in common use.
When I said Core2 Duo, the implication was 2c/2t. But to be specific, 2c/2t machines. The benchmarks of 6th -> 9th gen 2c/4t are much greater than the Core2 Duos and can do the work of a long1 or long2, slowly.
BOINC will see 4 threads unless an owner actually turned off hyperthreading.
I wouldn't want to be doing long SRBase work units on even a modern 2c machine with hyper threading turned off and no available processing power to even use Chromium. LLR is pretty aggressive even set to idle priority.
BTW, only 1st gen i7 600 series were 2c/4t, all later have been 4+ real cores. 2nd, 3rd, 4th and 5th gen i3/i5s had quite a lot of low power 2 cores sold (desktop i5's were 4 core), 6th gen and later (after Jan 2015) are 2/4, 4 and 6.
We'll just have to agree to disagree. There's a lot in what you said that I completely disagree with.
Unless you are referring to the earlier suggestions about default MT on LLR for SRBase long WU's, the only opinion I present in the last message was not wanting to run SRBase on even a modern 2 core machine with no hyper-threading.
The rest of my response was all factual taken from Intel database sources and sales figures presented in Wikipedia from tech articles.
All the machines I'm referring to are less than 6 GFlops (http://srbase.my-firewall.org/sr5/cpu_list.php) and it's unclear how many have completed a long WU in the last year (doubtful any completed one until the recent grace period was added). It would take some DB analysis.
EDIT: I see about 1200 of 61000 machines in the database that would be effected, most of which are likely not doing long WU's anyway, so that the rest of the workforce can see a 3x to 4x boost in speed. Not seeing a downside.
So, I assume you would love people to run SRBase on 2 core machines as that's the only thing you could be disagreeing with.
Posted 24 Apr 2019 by
marmot
Longs really shouldn't be run on a 2 core machine, which would generally be 8 year old Core 2 Duos, so excluding them would be setting an appropriate boundary.
Most modern Core i3 CPUs are 2c/4t.
All but the highest end laptops tend to be 2c/4t. Most Mobile Core i5 CPUs are 2c/4t, unlike their desktop counterparts. Unless I'm mistaken, many mobile Core i7s are also 2c/4t.
You don't have to go back 8 years to find dual core CPUs in common use.
When I said Core2 Duo, the implication was 2c/2t. But to be specific, 2c/2t machines. The benchmarks of 6th -> 9th gen 2c/4t are much greater than the Core2 Duos and can do the work of a long1 or long2, slowly.
BOINC will see 4 threads unless an owner actually turned off hyperthreading.
I wouldn't want to be doing long SRBase work units on even a modern 2c machine with hyper threading turned off and no available processing power to even use Chromium. LLR is pretty aggressive even set to idle priority.
BTW, only 1st gen i7 600 series were 2c/4t, all later have been 4+ real cores. 2nd, 3rd, 4th and 5th gen i3/i5s had quite a lot of low power 2 cores sold (desktop i5's were 4 core), 6th gen and later (after Jan 2015) are 2/4, 4 and 6.
Posted 21 Apr 2019 by
marmot
Having to play around with config file (and I remember this one is really not easy to do) really limits the potential number of users willing to try to do this.
Agreed.
Longs really shouldn't be run on a 2 core machine, which would generally be 8 year old Core 2 Duos, so excluding them would be setting an appropriate boundary.
So assuming minimum machine is a 4 thread CPU, and the user might be wanting to surf and do other activities, longs could default to MT 3. This is how I set up my app_config. On 4 core machine it leaves 1 core open to capture the short, average or regulars. On 8 core machine, 2 longs run and leave 2 cores available for the other WU's or another project.
Having longs default to MT 4 is also a decent option given a majority of commonly owned CPU's thread counts are divisible by 4.
Certain BOINC server software version have the preferences ability to decide core count of multi thread work units. You had mentioned your server software was an older version, so maybe this option isn't available.