Posts by arcturus

1) Message boards : Number crunching : i7-8700K performance (Message 2027)
Posted 7 Mar 2023 by arcturus
Post:
The i7 hasn’t completed anywhere near as many tasks as the Ryzens, so its measured processing rate for the new “long tasks” application version is still lagging (4.1 GFLOPS vs. 7.3, at the time of writing). The moving average from which credit is calculated is quite heavily filtered, so it will take several weeks to stabilise.

As opposed to stabilization which took under 5 days in your disingenuous apples vs oranges comparison versus the roughly 35 days which has already elapsed (add 2 more weeks according to you) running the newer units due to inherent optimizations either missing or not utilized which reduces the data set?

Don't quote me about patience.
2) Message boards : Number crunching : i7-8700K performance (Message 2024)
Posted 4 Mar 2023 by arcturus
Post:
40 credits an hour seems to be normal across the projects I work. Sometimes better when a new type of task is introduced but there or thereabouts.

In looking back at previous work for this project .. before the long work units .. the production with this hardware was considerably more. I would like to ask the developer what has changed, possibly with optimizations, to explain the reduced point awards. Also, I encourage others that have similar hardware to look back using something like boincstats to either confirm or deny what I'm seeing.
3) Message boards : Number crunching : i7-8700K performance (Message 2022)
Posted 3 Mar 2023 by arcturus
Post:
My R9 3900 averages about 40 so that sounds about right.

Right, within project. But with TN-Grid or Universe the difference is at most 10% to 15%. What makes SiDock so different?
4) Message boards : Number crunching : i7-8700K performance (Message 2020)
Posted 3 Mar 2023 by arcturus
Post:
I'm averaging just around 29 points per hour with this cpu. Full resources committed .. running Linux .. at 4.5Ghz. My newer Ryzen 5700's produce considerably more, over 52. Are there optimizations (SSE,AVX, etc) missing with the i7 which explains the considerable difference?

The differences aren't nearly as pronounced in other projects like TN-Grid or Universe.
5) Message boards : Number crunching : Credit Award Consistency? (Message 1981)
Posted 4 Feb 2023 by arcturus
Post:
Yes, the absence of any concerns or complaints lately speaks volumes. People just need to exercise a bit of patience.
6) Message boards : Number crunching : Credit Award Consistency? (Message 1978)
Posted 30 Jan 2023 by arcturus
Post:
For granting a fair credit to a computer, BOINC server must have a fair benchmark and have gathered appropriate statistics based on the set of tasks completed on this computer.
You can look, for instance, at my computer (# 4). Credits awarded per hour do not vary much because it has relatively large statistics of computing with cmdock-l 2.02 application version. This illustrates what value of granted credit is fair, too.

The Boinc benchmark is not fair/correct.
With Linux 20.x and higher you get the best benchmark value.
With Linux 18.x it is lower.
With Windows it is much lower.


This is common for all projects. Some are optimized for Linux, some aren't.
7) Message boards : Number crunching : Credit Award Consistency? (Message 1976)
Posted 30 Jan 2023 by arcturus
Post:
I unfortunately have to share this observations/concerns.
There is no systematic behind the credits granted - at least it is not visible for me.

Some examples below (identical Hardware):

99,486.52 - 99,478.48 - 1,332.88 - CurieMarieDock 0.2.0 long tasks v2.02 - windows_x86_64
84,178.01 - 84,168.72 - 1,074.65 - CurieMarieDock 0.2.0 long tasks v2.02 - windows_x86_64

54,468.31 - 54,464.46 - 1,259.42 - CurieMarieDock 0.2.0 long tasks v2.02 - x86_64-pc-linux-gnu
51,716.79 - 51,689.46 - 660.01 - CurieMarieDock 0.2.0 long tasks v2.02 - x86_64-pc-linux-gnu

Any serious ideas?

Chris


You haven't allowed enough time. Message 1973.
8) Message boards : Number crunching : Credit Award Consistency? (Message 1967)
Posted 29 Jan 2023 by arcturus
Post:
Adding some more examples of inconsistencies:

77689125 vs. 77689111 on one pair of identical machines.

77691276 vs. 77691076 on another pair. This machine seems consistently to get significantly more credit than this identical one.

These are all dedicated heaters running nothing but SiDock@home.


This happened to me also when I first started running the latest revised work but normalized quickly. Give it a few days and see if the same happens.
9) Message boards : Number crunching : Credit Award Consistency? (Message 1966)
Posted 29 Jan 2023 by arcturus
Post:
On one of my hosts Credit dropped from 955.43Cr to 390.61Cr , same runtime :
https://www.sidock.si/sidock/result.php?resultid=77686084
https://www.sidock.si/sidock/result.php?resultid=77686084

Is there a reason for that.

Hi!
Very unusual results! How many cores of this CPU used for computing? Could you run benchmark over used cores and post the result? And what frequency of CPU cores during computing?
Thank you!

The host is running headless only Boinc controlled with BoincTasks. I have run the benchmark after installation.
All 8C/8T running Boinc. 2 x Sidock, 4 x NFS, 1 x iThenaComp and 1 x Climate , frequency 2,44GHz at full load.


You linked the same work unit. How is one to compare?
10) Message boards : News : Target # 22: corona_RdRp_v2 (Message 1951)
Posted 26 Jan 2023 by arcturus
Post:
This is why I currently quit this project. Waiting for an improvement to return here :-)


The improvement happened days ago with the release of version 2.02 units. I've completed over 150 without one issue.
11) Message boards : Number crunching : Tasks hanging - (Message 1945)
Posted 24 Jan 2023 by arcturus
Post:
The newer 2.02 units seem to be completing ok, one puzzle that remains is inconsistent granted credit, in some cases a variation of up to 25% for the same computer.

https://www.sidock.si/sidock/workunit.php?wuid=49667156
https://www.sidock.si/sidock/workunit.php?wuid=49667239
12) Message boards : Number crunching : how long can "long tasks" be? (Message 1917)
Posted 21 Jan 2023 by arcturus
Post:
I'm running another project that has tasks that are quite allergic to restarts, and I'm running SiDock on a Ubuntu 22 machine on which the Boinc Manager keeps crashing a second after being started, so I can't use it and instead rely on boinccmd which makes aborting tasks a PITA.
Because of that I told my machine to stop asking for more SiDock long tasks. If I don't get anything else in the meantime I will have to wait until the SiDock long tasks are working properly.
Thanks for reading.


Try this: https://gene.disi.unitn.it/test/
13) Message boards : News : СmDock "long" and "short" tasks applications (Message 1908)
Posted 20 Jan 2023 by arcturus
Post:
You have been very vocal in your disappointment as seen over a number of different threads.

I suggest you move on to a different project (like I and others) or simply stop until the issues are resolved. It's not worth getting bent out of shape over.
14) Message boards : Number crunching : how long can "long tasks" be? (Message 1906)
Posted 20 Jan 2023 by arcturus
Post:
Has no determination why certain units are hanging been made yet? Or is this going to morph into a 'please give me extra time' thread?
15) Message boards : News : Target # 22: corona_RdRp_v2 (Message 1853)
Posted 17 Jan 2023 by arcturus
Post:
I understand the developer is looking into this but these new units are muddying the situation as to whether these are new 'normal' run times or whether they're truly hanging.

Not everyone gets around to checking progress often so I'm guessing we're going to see a stream of 'these units have been hanging for days!' posts soon. Looks like it has begun.
16) Message boards : Number crunching : Tasks hanging - (Message 1824)
Posted 16 Jan 2023 by arcturus
Post:
Project reset with no new work units instructed based on the number of them hanging.

On to another project until this issue is resolved.
17) Message boards : Number crunching : how long can "long tasks" be? (Message 1819)
Posted 15 Jan 2023 by arcturus
Post:
The circumstances that led to the unit hanging hasn't changed so I just abort them to avoid the possibility the same thing happens again to the same task.


No, the circumstances that led to the unit hanging are intermittent, by going back to the last checkpoint the likelihood of it happening again is the same as the likelihood of it happening in the first place.


In other words, the chance in intermittent. I'll rather avoid it completely and await the determination of the developer.
18) Message boards : Number crunching : how long can "long tasks" be? (Message 1816)
Posted 15 Jan 2023 by arcturus
Post:
The circumstances that led to the unit hanging hasn't changed so I just abort them to avoid the possibility the same thing happens again to the same task.
19) Message boards : Number crunching : Tasks hanging - (Message 1808)
Posted 15 Jan 2023 by arcturus
Post:
I have had 12 tasks hang among 4 different computers over the past few days. The lost time represents roughly 116 tasks which could of been processed.
20) Message boards : Number crunching : Credit Award Consistency? (Message 1770)
Posted 28 Nov 2022 by arcturus
Post:
Work unit taking 3hr 7min awarded 23.52 points
Work unit taking 2hr 25 min awarded 151.94 points

Usual behavior?




©2024 SiDock@home Team