Posts by HAL9000


log in
1) Message boards : Number crunching : Monitoring inconclusive GBT validations and harvesting data for testing (Message 1820567)
Posted 10 hours ago by Profile HAL9000
And don't we also have to consider the threshhold calculation when considering whether the outcome of those averaging/blocksum calculations is reportable?


Yes. For the fixed thresholds such as spikes there is an issue whereby a spike happening to fall just above threshold will be reported, and in another system may not. That uncertainty is the 'threshold problem' Eric has alluded to in the past, whereby there is no hysterisis band on reportability. [I have offered forward suggestions for solutions for the future, and Eric has expressed interest, though they require detailed modelling, and integration of error measurements in the apps (partially present in stock CPU bench part)]. Unlikely to go in anytime soon since v8 was pushed out before these moves could germinate fully.

For the changing thresholds that are calculated, similar constraints apply for the levels or fit, with the complication that there will be some small amount of unavoidable error in the threshold value calculation itself.

[Edit:] I should point out that we're talking tiny fractions of very small numbers, that, where would apply, would stand out as being close to some threshold in one or another result file)

Interestingly techniques for this sortof stuff were invented at Berkeley decades ago, but mostly ignored in the US. The Japanese took up a lot of the work in that area and made autofocus, and ricecookers into things.

It's good to know that the maths behind looking for ETI signals & perfectly cooked rice are related.

Also at the rate that processors are being stuffed into everything I expect my next rice cooker will be able to run some of the SETI@home apps in a timely manor. http://www.geek.com/android/android-makes-the-transition-to-rice-cookers-and-beyond-1535582/
2) Message boards : Number crunching : BOINC client isn't downloading new S@H workunits on S6 Android (Message 1820163)
Posted 2 days ago by Profile HAL9000
Can you upload or send it ?

https://boinc.berkeley.edu/dl/boinc_7.4.43.apk
3) Message boards : Number crunching : Reload S@H to get Astropulse? (Message 1818785)
Posted 8 days ago by Profile HAL9000
I must not have been as specific as I intended.

My plan is to download SETI@Home from the BOINC site; not to download BOINC again.

I understand that SETI files come from SETI.

No, now that I re-read my original question, I was more than specific enough in describing what I would like to do. The problem seems to be that you had difficulties understanding my description.

In looking at the S@H files in the projects subdirectory, it looks like I'm missing some files. My plan is/was to finish up all my current S@H tasks, completely unload S@H, reboot everything and reinstall S@H from the BOINC site on the assumption that I will get a complete fileset.

The main reason your machines are not getting Astropulse work is because the project has been issuing very little of it.

There isn't really any "install" for the SETI@home project files. Unless you are referring the using the Lunatics optimized application installer.

If you feel your local project files are damaged in some way you can simply hit the "Reset Project" button in BOINC. Then the application files will be downloaded as needed when your machine is issued new work.
4) Message boards : Number crunching : RX 480 - A serious contender? (Message 1818564)
Posted 8 days ago by Profile HAL9000
Actually the RX480 is in my graphs under "Ellesmere" -- it's infuriating but it seems like the CUDA device identifier insists on naming the family not the model which tragically will mean this includes RX470's etc -- however if you go back in that thread you'll find my earliest data for Ellesmere was consistently mediocre even with the RX480 was the only part in circulation.

BOINC 7.6.23+ uses a new name detection method for Radeon GPUs so that the BOINC developers don't have to hard code the GPU names. I don't know if you are taking that into account when searching for GPUs to chart.

This is how my R9 390X is displayed with the new code vs the old.
BOINC 7.6.23+: AMD AMD Radeon (TM) R9 390 Series (8192MB) OpenCL: 2.0
BOINC 7.6.22: AMD Grenada (8192MB) OpenCL: 2.0

For clarity I should have stated that the new detection method is only for non-CAL GPU/drivers. Older drivers that still report a CAL version will continue to display the hard coded names from BOINC like:
AMD AMD Radeon HD 7870/7950/7970/R9 280/R9 280X series (Tahiti) (3072MB) driver: 1.4.1848 OpenCL: 1.2

Also it has been found, at least in Linux, that no board name is displayed for RX 400 GPUs. Do we have any Windows users with a RX 400's using a BOINC 7.6.23 or newer?
5) Message boards : Number crunching : Open Beta test: SoG for NVidia, Lunatics v0.45 - Beta4 update (Message 1818387)
Posted 9 days ago by Profile HAL9000
There's a suitable power plug or two. I just ordered a multimeter.

Will be very interesting to see how the 5/12v moves when BOINC is and is not running... can't say I've tested that, either ...

If you really want to see something scary you should use an oscilloscope to view your power rails at idle & with BOINC running apps.
I did that a few years ago on one of my machines.
6) Message boards : Number crunching : RX 480 - A serious contender? (Message 1818231)
Posted 10 days ago by Profile HAL9000
Actually the RX480 is in my graphs under "Ellesmere" -- it's infuriating but it seems like the CUDA device identifier insists on naming the family not the model which tragically will mean this includes RX470's etc -- however if you go back in that thread you'll find my earliest data for Ellesmere was consistently mediocre even with the RX480 was the only part in circulation.

BOINC 7.6.23+ uses a new name detection method for Radeon GPUs so that the BOINC developers don't have to hard code the GPU names. I don't know if you are taking that into account when searching for GPUs to chart.

This is how my R9 390X is displayed with the new code vs the old.
BOINC 7.6.23+: AMD AMD Radeon (TM) R9 390 Series (8192MB) OpenCL: 2.0
BOINC 7.6.22: AMD Grenada (8192MB) OpenCL: 2.0
7) Message boards : Number crunching : SETI@home on a Samsung Galaxy Tab 2 (Message 1817993)
Posted 11 days ago by Profile HAL9000
Not sure if this question should go here or over on the BOINC website.

Purely for kicks, I installed BOINC and S@H on my Galaxy Tab 2. Everything runs ok - I've even gotten a few credits.

I added an SD micro chip, but can't get either BOINC or S@H to use it.

Is there a setting or something somewhere to tell S@H to move to the SD chip?

The Android version is 4.2.2, which is current for this device.

On all of my Android devices I don't get the option of "Move to SD Card" for BOINC. I imagine this was a choice they made for the app. As many devices unmount the SD card when connected to a PC. Some devices also unmount the SD card in low power states.
8) Message boards : Number crunching : MB v8: CPU vs GPU (in terms of efficiency) (Message 1816031)
Posted 20 days ago by Profile HAL9000

This type of calculation does not take into account the whole power usage of the system the device is in. To do that you would really need to use a power meter, or UPS with a power usage display, and measure each system at idle and then when processing tasks. The delta could then be used to calculate Wh/Task.

And that's quite important part. If some device completes task much faster - it needs correspondingly less amount of time when whole system powered ON. If very low-power device takes much longer to complete the same task - it requires full system support (with all its energy consuming overhead) through all this long time.

Hence, w/o accounting for whole system energy consumption overhead such CPU vs GPU efficiency comparisons are quite biased IMO and don't show real benefits of fast GPU computing.

I find it most useful to calculate the CPU & GPU in the same system for each app or type of work from a project. Then I can use the information to determine which device in the system is most suited to running a given type of work.
9) Message boards : Number crunching : MB v8: CPU vs GPU (in terms of efficiency) (Message 1815815)
Posted 21 days ago by Profile HAL9000
I normally calculate the Watt Hours per task for my devices to rank them by their efficiency. I have done a few posts about the method I use here & also here.
Using the TDP, run time, & # of tasks per device for your CPU & GPU I get.
Note: I rounded the task times to the nearest minute.
Also this calculation expects the run times to be when the specified number of tasks were running.
Device Watts # Tasks Run Time Task/hr Task/day Wh/Day Wh/Task GTX 750 55 2 42 2.857142 68.571428 1320 19.25 N3520 7.5 4 305 0.786885 18.885245 180 9.53

Which would make the N3520 ~2.02 times as efficient as the GTX 750, but it completes less than a third as much work in a given day.

However, actual GPU power consumption is typically lower than the rated TDP value when processing tasks. A good rule of thumb is to use ~80%.
Many have found that their GTX 750's run in the 40-45W range. If we figure 80% of 55W that does happen to be 44W. Which would give the GTX 750 a Wh/Task of 15.4 instead of 19.25 & make the N3520 only ~1.62 times as efficient.
I should also add that my GTX 750 Ti FTW would complete two 0.42AR tasks at once in ~25 min while using ~45w. Giving it a Wh/Task of 9.38.

This type of calculation does not take into account the whole power usage of the system the device is in. To do that you would really need to use a power meter, or UPS with a power usage display, and measure each system at idle and then when processing tasks. The delta could then be used to calculate Wh/Task.

Task/hr = (60/Run Time)/# Tasks
Task/day = Task/hr * 24
Wh/Day = Watts * 24
Wh/Task = (Wh/Day)/(Task/day)
10) Message boards : Number crunching : odd performance issues with two i7 4770K CPU's (Message 1815104)
Posted 25 days ago by Profile HAL9000
thanks guys, this page http://boinc.berkeley.edu/wiki/Computation_credit gives a description but I don't understand what it is meant by:

Recent average credit: The average number of Cobblestones per day granted recently. This average decreases by a factor of two every week.

especially the last sentence...

A simplified explanation could be something like.
RAC is an average of the past few weeks of daily credit. So the first day a host reports credit is has several days worth of 0's to overcome. If we go with 6 weeks & 1000 credit each day. Then the RAC would look something like:
Day 1: 1000/42 = 23.81 RAC
Day 7: 7000/42 = 166.66 RAC
Day 14: 14000/42 = 333.33 RAC
Day 21: 21000/42 = 500 RAC
Day 42: 42000/42 = 1000 RAC

Since RAC is Recent Average Credit the goal is likely to keep the value recent. Host RAC is, or was at one time, updated each time host credit updated. If a host stopped reporting completed work the RAC value would become stale with no further updates. So the decay mechanism will very slowly reduce the RAC of an inactive host to 0.
11) Message boards : Number crunching : Thought(s) on changing S@h limit of: 100tasks/mobo ...to a: ##/CPUcore (Message 1813479)
Posted 29 Aug 2016 by Profile HAL9000
Some links on the WU limit:


I was thinking of posts a bit older.
1185411
1197674
1229214

The posts I can find where the staff told us the values of the task limits are from 2010. Other posts are just notes like "task limits were raised".
12) Message boards : Number crunching : Thought(s) on changing S@h limit of: 100tasks/mobo ...to a: ##/CPUcore (Message 1813062)
Posted 27 Aug 2016 by Profile HAL9000
The last time per processor CPU limits were used the value was 50. So perhaps half of that, at 25 per processor, would be sufficient.

I don't recall ever having per processor or per core WU limits before.
I do remember them making the GPU limit per GPU instead of for all GPUs.

I believe it was around the end of 2011 or start of 2012. I seem to recall when they first set the task limits they had accidentally set 50 total per host. Then changed it to per processor for CPU. Then after some time the limits were removed, db when splat again, & then the limits were implemented again.
13) Message boards : Number crunching : Thought(s) on changing S@h limit of: 100tasks/mobo ...to a: ##/CPUcore (Message 1812863)
Posted 27 Aug 2016 by Profile HAL9000
Most of my machines are i7's so they get 100 / 8 threads = 12.5 WU per thread.

If we expand on that by saying we get 12.5 x number of threads then it could work for the smaller (single thread) machines as well as the larger (56 thread) machines. I would suggest we make it something like 15 per thread. Its simplistic and achievable with the current infrastructure.

A more long term approach might be to increase that number based upon the average turnaround if the host is considered reliable. It could also be applied the other way to reduce the number if the host is unreliable.

The last time per processor CPU limits were used the value was 50. So perhaps half of that, at 25 per processor, would be sufficient.
14) Message boards : Number crunching : Thought(s) on changing S@h limit of: 100tasks/mobo ...to a: ##/CPUcore (Message 1812650)
Posted 26 Aug 2016 by Profile HAL9000
A few options come to mind.

1.
The CPU limit could be applied based on processors and determined value. Something along the lines of (host # of processors/4)*task limit.

2.
Perhaps modifying the JobLimits options to allow for specified host # of processor range limits.

Something along the lines of:
<project> <cpu_limit> if set, limit is applied to all hosts unless another values applies to the host <jobs>N</jobs> <cpu_limit_16> if set, limit is applied to host with 16+ processors <jobs>N</jobs> <cpu_limit_32> if set, limit is applied to host with 32+ processors <jobs>N</jobs> <cpu_limit_64> if set, limit is applied to host with 64+ processors <jobs>N</jobs> <cpu_limit_128> if set, limit is applied to host with 128+ processors <jobs>N</jobs> </project>


3.
A graduated max CPU tasks in progress could be derived using Number of tasks today from the CPU apps. It might be necessary to take the Number of tasks today and then create an Average daily Number of tasks to use. Then a using the specified value set by the project a dynamic limit could be applied based on how productive the machine is rather than by the indicated number of processors.
I think this might be the most complicated method & with each app version change the average would be reset.
15) Message boards : Number crunching : Thought(s) on changing S@h limit of: 100tasks/mobo ...to a: ##/CPUcore (Message 1812197)
Posted 24 Aug 2016 by Profile HAL9000
I don't know the method that is currently being used to detect the # of processors, but I'm aware that there are cpuid functions that provide core/thread counts. That may also provide the socket count for the system, or could be somewhat wonky. Providing two lists of core 0,1,2,3 for a dual 4 core MB.

If the CPU socket count can be determined. Then having a limit of tasks per CPU socket would be helpful for multi CPU systems.
16) Message boards : Number crunching : Windows 10 - Yea or Nay? (Message 1810603)
Posted 20 Aug 2016 by Profile HAL9000
My computers just went into no more updates mode.

I installed Sophos to handle my AV and malware needs.

Most of my machines never do updates.
http://i.imgur.com/TH5OjMS.png
But I did have to install an update on my HTPC once to get something to work.
http://i.imgur.com/WJ9Ggf8.png
17) Message boards : Number crunching : Windows 10 - Yea or Nay? (Message 1809912)
Posted 17 Aug 2016 by Profile HAL9000
[IMG ]http://i715.photobucket.com/albums/ww153/Jimbocous/the-true-origin-of-the-tin-foil-hat.jpg[/IMG][/URL] Here we go again ...

I don't know why anyone would wear a foil hat like that to try to block signals in/out of their head. Unless it is properly grounded it would only serve as antenna.

With the free upgrade period offer over I've been considering setting up another Window 7 VM to see what updates they push out to it. Last time I did that I came back to the VM running Windows 10 with no human intervention.
18) Message boards : Number crunching : 20-core/40-thread Xeon E5-2630 v4 duo (Message 1809080)
Posted 14 Aug 2016 by Profile HAL9000
I'm pretty sure there are ways to split PCIe slots into multiple x4s and so forth. Theoretically you should be able to turn one x16 into 16 x1's.

I don't see how.

Each slot has it's own device number. The PCIe*1 section includes power, data, signalling connections. The other sections just include data & signalling to indicate they are being used.
The slot, and the software drivers, are designed for a single device- not for running multiple different devices in the one slot.
That's why there are multiple slots.

Some slots share PCIe lanes, which is why a PCIe*16 slot can become a PCIe*8 slot when another slot on the motherboard is used.
One device per slot; the PCIe lanes can be shared between slots, but they are only ever used by one device at a time.

I believe the configuration of the PCIe slots on a MB may be limited by Intel or AMD. On Intel's site they often list the PCIe configuration options under the CPU specs.
Supported Processor PCI Express Port Configurations: 1x16, 2x8, 1x8 and 2x4

The use of PCIe switches or splitters can be utilized to allow more PCIe devices than originally intended. The easiest options is to use a PCIe expansion chassis.
19) Message boards : Number crunching : 20-core/40-thread Xeon E5-2630 v4 duo (Message 1808581)
Posted 11 Aug 2016 by Profile HAL9000
Given:

... a simple rig with twin 1070's should be more productive.

also

And when you consider how well Al's system performs, this would be a very capable cruncher.

and, although this remains unanswered

I'd hazard a guess that a pair of 1070's would be a bit better throughput.

For guppies or for traditional work?

My interest is long-term (ie. Efficiency).  For example, referencing Shaggie's graphs (most recent one here), we can see that when it comes to production (Avg Credit/Hour), the 1070 spanks the 750Ti.  However, when it comes to efficiency (Avg Credit/Watt-Hour), the 750 Ti more than holds it own, which means more bang for your buck (efficiency-wise).

Having said that, if we were to compare a rig only crunching on two 1070s (with 1, 2 or 3 WUs concurrently), with the topic rig of this thread crunching on all threads concurrently, which one would wear the efficiency crown?

Since Credit is variable. I find comparing watt hours per task to be much more helpful in finding how efficient a device is for crunching. With the Wh/task info charts could be made ranking the devices by efficiency and a column for the run time would be helpful to guesstimate credit.
20) Message boards : Number crunching : Connection/computing scheduler...anyone ? (Message 1806843)
Posted 3 Aug 2016 by Profile HAL9000
I tried manually setting multiple network time windows in the global_prefs_override.xml, but as I suspected BOINC only read the last entry. So it looks like BOINC itself is not designed for what you want to accomplish.

However as Shaggie suggested you can use the OS to schedule events. Then you can use boinccmd to give BOINC commands with a script along the lines of:
boinccmd --set_network_mode auto timeout 30 boinccmd --project http://setiathome.berkeley.edu/ update timeout 120 boinccmd --set_network_mode never

Note: You may need to adjust the timeout values depending on your system(s) and internet connection.


Thank you Shaggie & Hal, so I guess that the answer to my question about the single instance of boinc is implied in your answer : scheduling 5 times per day your script ( embedded in a .cmd file) and adjust especially the second timeout to upload the 100/120 worked wu should do the trick, right ?
One more question: is this command line parameter overriding any gui settings or should I set the network activities to " based on preferences" ?

Thank you..!


Yes, For your Windows hosts setting multiple triggers for the specific times you wish to connect is likely the route you will want to take. You could even factor in the schedule weekly server maintenance if you wanted to do so. I would guess that OS X on your MAC host would have something similar, but I don't have an OS X host to check at this time. The scripting would be similar, but I don't know what the OS X command for a timeout is without looking it up.
The first delay is to give BOINC a moment to switch modes. On a slower machine this can take a few seconds. A faster machine might be ready in ~5 seconds.
The second delay is how long to wait before telling BOINC to stop network activity again. A value of 600 to 900 sec, 10 to 15 min, might be more suitable. It just depends on your internet connection to the servers.

Boinccmd is simply another way to control the BOINC client. Changing your network connection to never will leave it set to never. Just as if you used the GUI to do so. For a full description of the command line tool you might want to check out http://boinc.berkeley.edu/wiki/Boinccmd_tool
A quick description of the options for the network, CPU, & CPU run commands.
always = always
never = never/suspend
auto = based on preferences


Next 20

Copyright © 2016 University of California