Posts by IDEA

1) Message boards : Number crunching : Hundreds of tasks failing (Message 3220)
Posted 18 Mar 2021 by Profile IDEA
Post:
Hey, IDEA, glad to see you are still hanging around here.


I've still got a couple of old iMacs crunching Collatz because I can't find much else for them to crunch :(

Looks like this problem is still persisting though as boincstats shows everybody has 0 credit for the last 24 hours and the server is responding with "temporarily down for maintenance".

PS. Hope the hatch battening was successful and the weather bypassed you?
2) Message boards : Number crunching : Hundreds of tasks failing (Message 3209)
Posted 17 Mar 2021 by Profile IDEA
Post:
UK Times:

Wed 17 Mar 05:35:31 2021 | collatz | Started upload of collatz_sieve_d4fa341a-8993-4b45-987b-2f1fb5d622b2_0_r179307990_0
Wed 17 Mar 05:35:33 2021 | collatz | [error] Error reported by file upload server: can't write file collatz_sieve_d4fa341a-8993-4b45-987b-2f1fb5d622b2_0_r179307990_0: No space left on server
Wed 17 Mar 05:35:33 2021 | collatz | Temporarily failed upload of collatz_sieve_d4fa341a-8993-4b45-987b-2f1fb5d622b2_0_r179307990_0: transient upload error

Later followed by:

Wed 17 Mar 06:09:56 2021 | collatz | Computation for task collatz_sieve_d25e0415-4a43-43ce-9941-4eff51679842_0 finished
Wed 17 Mar 06:09:56 2021 | collatz | Output file collatz_sieve_d25e0415-4a43-43ce-9941-4eff51679842_0_r1141664431_0 for task collatz_sieve_d25e0415-4a43-43ce-9941-4eff51679842_0 absent

And then all tasks failing after that :(
3) Message boards : Number crunching : Multiple card species question.... (Message 3000)
Posted 1 Dec 2020 by Profile IDEA
Post:
Keep the settings low.

I store 0.3 days and an additional 0.1 days on a host that uses a slow ATI GPU and fast Nvidia GPU.

Both GPUs have been running constantly and haven't run out of work units at any point.
4) Questions and Answers : Macintosh : Insights on crunching using a MacMini 8,1 (Message 2997)
Posted 30 Nov 2020 by Profile IDEA
Post:
If you're tracking the Macs then you'll notice that both ID: 817196 and ID: 825690 have stopped crunching...

These are the two most prolific Mac hosts on BOINC (74th & 81st) -- with you currently in 88th position?

Both are now running Ubuntu and crunching other projects (and still doing Collatz every so often).

The drives are stored in a safe place, so they may return again one day...

Top Mac spot is yours for the taking now. Could be yours in two months ;-)
5) Questions and Answers : Macintosh : Insights on crunching using a MacMini 8,1 (Message 2989)
Posted 27 Nov 2020 by Profile IDEA
Post:
Using a pass-through port interests me -- even with 3 pairs of TB2 ports on the 6,1 I can run out of ports. I've been running a 6,1 Mac Pro with two eGPUs attached. One with a 1080ti and the other an RX580. They perform exactly as expected even when the other TB 2 ports are busy with RAID arrays attached to them. It's currently in 3rd place, behind your two Darwin hosts:

https://boinc.thesonntags.com/collatz/show_host_detail.php?hostid=817142

Didn't seem to be any way of using different .config files for different types of ATI GPUs, so disabled the built-in FirePros and switched them to a different project.

Q: Do you or anybody else know if it is possible to run multiple .config settings when using two different GPU models from the same manufacturer?

I've repurposed my older Mac Minis in the last fortnight and they are now running Ubuntu. Annoyingly, Ubuntu works perfectly with an eGPU attached to a Mac Mini, but I can't get BOINC to recognise and use the attached GPU. They were CPU crunchers anyhow and I decided to make the switch because crunching in Linux earns a lot more credit than the Mac OS in one of my projects ;-)

But, if anybody else has got experience of eGPUs and Linux, I'd like to know how it's done?

I do clean the Mac Minis regularly by taking the fan out and giving the blades a brush. Haven't had to disassemble the fans yet, but will probably need to at some point to oil the bearings.
6) Message boards : Number crunching : Error: GPU steps do not match CPU steps. (Message 2959)
Posted 29 Oct 2020 by Profile IDEA
Post:
More info:

My other 1080ti is running on a Mac, but it looks like this 256 thread limit is a driver issue.

The Mac stderr reads:

Driver Version      10.33.0 387.10.10.10.40.139
actual threads 512
Max Kernel Threads  1024


Whereas the Windows stderr reads:

Driver Version      456.71
actual threads      64
Max Kernel Threads  256


So it looks like this new Windows driver is imposing a 256 thread limit.

Interesting that removing the thread line in the config causes the app to use 64 threads but the difference in processing time is only increased by around 15 seconds.
7) Message boards : Number crunching : Error: GPU steps do not match CPU steps. (Message 2958)
Posted 29 Oct 2020 by Profile IDEA
Post:
Through a process of elimination, I've worked out that it is the manual setting of the "threads" value which is causing the WUs to error.

If you remove the threads line the .config will work and the WUs will process as expected.
8) Message boards : Number crunching : Error: GPU steps do not match CPU steps. (Message 2957)
Posted 29 Oct 2020 by Profile IDEA
Post:
I've seen this error mentioned in other forum posts, but no answers were provided as to what was causing the problem.

I have a 1080ti card running on a Windows 10 box.

It will happily crunch WUs without any issues *until* I edit the .config file. Then it errors every WU with a "computation error".

The config settings worked fine with the same 1080ti card in another Windows 10 box prior to me swapping the GPUs. (I can't remember which driver the previous PC was using though).

If I delete the text in the .config file BOINC will resume crunching WUs as normal.

Collatz Config Settings:

verbose             1 (yes)
kernels/reduction   48
threads             2^9 (512)
lut_size            18 (2097152 bytes)
sieve_size          2^30 (51085096 bytes)
sleep               1
cache_sieve         1 (yes)
reducecpu           0 (no)
Processor Type      NVIDIA
Max Dimensions      3
Max Work Items      1024 1024 64
Max Work Groups     1024
Max Kernel Threads  256
Device Vendor       NVIDIA Corporation
Name                GeForce GTX 1080 Ti
Driver Version      456.71
OpenCL Version      OpenCL 1.2 CUDA
Device Vendor       NVIDIA Corporation
Name                GeForce GTX 1080 Ti
Driver Version      456.71
OpenCL Version      OpenCL 1.2 CUDA
actual threads      512
Error: GPU steps do not match CPU steps.  Workunit processing aborted.
15:02:23 (10332): called boinc_finish(9999)


Any suggestions?
9) Questions and Answers : Web site : DNS isn't working (Message 2953)
Posted 16 Oct 2020 by Profile IDEA
Post:
; <<>> DiG 9.10.6 <<>> SOA boinc.thesonntags.com
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: SERVFAIL, id: 21927
;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 1
10) Questions and Answers : Macintosh : Insights on crunching using a MacMini 8,1 (Message 2952)
Posted 16 Oct 2020 by Profile IDEA
Post:
Please keep us updated on your efforts to connect a TB3 eGPU to a MacPro 5,1 using a TB3 card. Given that my two primary "work" macs (one at home and one at work) are MacPro 5,1s, I am very interested in strategies to maintain the useful lifespan of these rigs. I can't afford an iMac Pro or a MacPro 7,1, and my MacBookPro is too limited to use as a primary work computer.


I'm using a Sonnet Allegro 2-port USB-C PCIe card.

It works in the 5,1 Mac Pros and I can also use it on Windows boxes too.

Neither need any special drivers as both the Mac OS and Windows 10 recognise the card and you can just plug your breakout box in and start playing (although, with your Mac you need to do the usual preliminary work to get it to recognise Nvidia cards).
11) Questions and Answers : Macintosh : Insights on crunching using a MacMini 8,1 (Message 2949)
Posted 14 Oct 2020 by Profile IDEA
Post:
What happened to your Mac setup?

Have you retired from BOINC now or just taking a rest?
12) Questions and Answers : Macintosh : Insights on crunching using a MacMini 8,1 (Message 2937)
Posted 6 Oct 2020 by Profile IDEA
Post:
Thanks for the info David.

Have you tried eGPUs with your older Macs?

I'm currently using a cheap Sonnet Breakout Box containing an Nvidia 1080ti attached to a Mac Pro 6,1 (running High Sierra).

The setup was pretty easy - you just need Apple's TB3 to TB2 adaptor and the purge-wrangler script and instructions that you can find here:

https://github.com/mayankk2308/purge-wrangler/wiki/Beginner's-Guide

Lack of a Mac version of Afterburner means that you can't get the same performance is you'd get running the Nvidia cards in Windows, but it's 10x faster than the fastest FirePros that the 6,1 is limited to and faster than any of AMD's cards too -- with an average time of 265s per WU.

Next move is adding a TB3 card to one of my old 5,1 Macs and running the Sonnet without Apple's TB3 to TB2 adaptor.

Worth noting that Apple & Nvidia's refusal to play nicely together is bad for end users like us as it means that we have no way of using the newer, energy efficient cards like the 1660 which are as fast as the 1080ti but use about a third of the power :(

Well done with getting Darwin to Number 13 in the Top Hosts chart ;-)
13) Message boards : Number crunching : Server Error/Not Running (Message 1719)
Posted 1 May 2019 by Profile IDEA
Post:
Still seeing http errors at this end. Web site is glitchy and the Participants page has PHP errors:

Notice: unserialize(): Error at offset 27970 of 28672 bytes in /home/boincadm/projects/collatz/html/user/top_users.php on line 95
Warning: Invalid argument supplied for foreach() in /home/boincadm/projects/collatz/html/user/top_users.php on line 113
14) Message boards : Number crunching : Server Error/Not Running (Message 1716)
Posted 30 Apr 2019 by Profile IDEA
Post:
Anybody got direct contact with Slicker?

We don't want a repeat of last year's Thanksgiving fubar :(
15) Message boards : Number crunching : Optimizing the apps (Message 1223)
Posted 5 Nov 2018 by Profile IDEA
Post:
This NVIDIA GTX 960 config works OK:

verbose 1 (yes)
kernels/reduction 48
threads 2^8 (256)
lut_size 16 (524288 bytes)
sieve_size 2^30 (51085096 bytes)
sleep 1
cache_sieve 1 (yes)
reducecpu 0 (no)
16) Message boards : News : Use at your own risk (Message 1204)
Posted 27 Oct 2018 by Profile IDEA
Post:
This info from Primegrid may be of interest to Jon?

Appears that Gridcoin have been inadvertently DDOSing project servers by having a huge number of hosts demand project stats at exactly the same time:

In early September I noticed that PrimeGrid was virtually unreachable for a few minutes. It turns out there was a DDOS attack against us, but it was a short lived attack. It only lasted a few minutes. Some investigation determined that this had been happening about once a day for a while, but since it was short in duration nobody really noticed.

By now you probably have guessed where this story is going.

Gridcoin needs our statistics, and collects them by having a large number of host computers each collect our stats files. At pretty much the exact same time. And these are really large files. While we have a fairly hefty server, and a decent pipe to the Internet, Gridcoin *is* a significant player at PrimeGrid, and all those hosts (over a thousand) downloading those large files at the same time was completely saturating our Internet connection.

PrimeGrid isn't the largest BOINC project, but we're up there, and our servers are more powerful than most. I'm not saying that to brag. The point is that if Gridcoin can effectively, albeit unintentionally, mount a successful DDOS attack against PrimeGrid's servers, imagine what Gridcoin is doing to all the other BOINC sites with much smaller hardware. Their only saving grace is that their stats files are almost certainly smaller than ours, and they may have fewer Gridcoiners (is that a word?) on their system. Nevertheless, this method of gathering the statistics is undoubtedly affecting many BOINC sites adversely. What is certain is that it has been making PrimeGrid's server unreachable for periods of time. That's unacceptable. Except for the fact that this is clearly unintentional, it would literally be criminal.

Back in September I reached out to Gridcoin. There's a fix in the pipeline. Things will hopefully be back to normal soon, minus the DDOS part.

We're not the bad guys here. Nor is Gridcoin. Everyone is working together to get this working.


Message link: http://www.primegrid.com/forum_thread.php?id=8229&nowrap=true#121175
17) Message boards : News : Use at your own risk (Message 1164)
Posted 15 Oct 2018 by Profile IDEA
Post:
If you take part in the project for the project great but then don't mention gridcoin.
If you take part for the gridcoin, move on to a other project. Goodby.
And if you don't want to move on, at least make a separate thread for your complaints so that I can ignore your whinging.

Thanks for your interest.

Our 1783 team members will be delighted to hear the news that Collatz is included again.
18) Message boards : News : Use at your own risk (Message 1160)
Posted 15 Oct 2018 by Profile IDEA
Post:
That isn't true -- new Bitcoins are mined everyday.


I said "minted" not "mined".

More GRC are minted every two years than the total number (21 million) of BTC that will *ever* exist.
19) Message boards : News : Use at your own risk (Message 1159)
Posted 15 Oct 2018 by Profile IDEA
Post:
Who cares. If you want to complain do it at that gridcoin site.


Who cares?

The largest and most productive team on this project.

And why complain to Gridcoin when the exclusions are usually caused by flaky hardware and administration on the project side?
20) Message boards : News : Use at your own risk (Message 1155)
Posted 14 Oct 2018 by Profile IDEA
Post:
Collatz has been excluded from Gridcoin again :(

Along with Primegrid and SETI too.

https://gridcoinstats.eu/project


Next 20


©2021 Jon Sonntag; All rights reserved