Posts by HAL9000


log in
21) Message boards : Number crunching : Windows 10 - Yea or Nay? (Message 1706456)
Posted 6 days ago by Profile HAL9000
I grabbed the "Windows 10 IoT Core for Raspberry Pi 2 (ARM)" version when I was getting the standard .isos at work. Now I just need a Raspberry Pi 2 to install it on.
22) Message boards : Number crunching : Panic Mode On (99) Server Problems? (Message 1705862)
Posted 8 days ago by Profile HAL9000
Upgraded one of my machines to 7.4.42.....miss the messages section. Notices is just annoying....

I'm still using the manager v6.10.48. So I don't have to see the annoying notices tab & get my message right where I like them. In plain view.I've considered downloading the source & building a manager with the current version but restoring the messages tab instead of the hidden event log, but I don't want to go through the effort of setting up the environment to just to compile the manager.
23) Message boards : Number crunching : Reddit thinks SETI@home is not useful (Message 1705774)
Posted 8 days ago by Profile HAL9000
Well your "used ID is - 226846
Hal's is - 792
And mine is - 812533

Quite why you have a lower id than me seems odd but there probably is a reason.

Off topic, but since the question came up - I got UserID 5509.

About a year before SETI Classic closed down, and when BOINC was getting ready to roll, they did a bulk import of active SETI Classic users to seed the new BOINC database. I don't know how far down the list they went, but those which were imported were done in Classic ranking order - so the lower your UID now, the more Classic work you returned.

I had always wondered why my UID was so low. I figured it had something to do with signing up with BOINC during the original alpha testing, but your explanation seems rather logical. When they did the import I might have been #792, but by the time Classic ended I dropped down to #1001.
24) Message boards : Number crunching : Reddit thinks SETI@home is not useful (Message 1705660)
Posted 8 days ago by Profile HAL9000
Quite a few of us would like to see him disappear for good and never come back, but even his little contribution does help and has to be seen as welcome. If he annoys you that much, which of course he deliberately does to get attention, the answer is to put him on ignore. As an international project we cannot stop the mentally challenged from signing up and posting, provided they don't break the rules.

I found out that BOINC has a limit of only 26 users on the ignore list. Trying to add more results in an error.


Actually when Zapped Sparky reported this a a problem this was David Anderson's answer

The database field that stores ignored user IDs is 254 chars long, and Zapped has reached this limit.

The code should handle this more gracefully, or remove the limit. For now, you can un-ignore some people.


So it is all to do with how long the users ID's are.

Well that explanation makes sense. However with the 26 users I have on ignore their total length comes to 161 characters. Likely closers to 186 with a comma separating IDs or 238 if the values are stored in quotes & separated by commas. It seems like I should be able to add at least 1 more, but maybe there is some other overhead in that stored value.
25) Message boards : Number crunching : Windows 10 - Yea or Nay? (Message 1705571)
Posted 8 days ago by Profile HAL9000
Further updates are in the works (already)

Before Windows 10 has even launched, Microsoft's upgrade plans for the OS have leaked. Most recently, Neowin reports that wave 2 of Threshold, the long-standing codename for Windows 10, will land this October. However, this update will reportedly focus on optimizing and stabilizing the operating system. If you're already holding out for some major changes, expect to hold out for quite some time.

According to an earlier Neowin report, the first major update will come June 2016 and the other October of next year, with Microsoft operating under the codename Redstone. That said, I would expect that timing to change. A whopping 11 months is quite a long time before issuing features that were cut from the initial Windows 10 release.

Microsoft recently clarified that statement with specific figures: Windows 10 will be supported with new features and updates until October 13, 2020. Extended support, e.g. security patches, will carry on until October 14, 2025. This seems to apply to both Windows 10 on desktop and Windows 10 Mobile.


Win 10

As of Friday our contact to MS indicated that MS still does not have a release date for Windows 10 Mobile. Nor did they indicate which devices will currently support updating to Windows 10 Mobile. Which was our actual inquiry.
26) Message boards : Number crunching : Reddit thinks SETI@home is not useful (Message 1705566)
Posted 8 days ago by Profile HAL9000
Quite a few of us would like to see him disappear for good and never come back, but even his little contribution does help and has to be seen as welcome. If he annoys you that much, which of course he deliberately does to get attention, the answer is to put him on ignore. As an international project we cannot stop the mentally challenged from signing up and posting, provided they don't break the rules.

I found out that BOINC has a limit of only 26 users on the ignore list. Trying to add more results in an error.
27) Message boards : Number crunching : Waiting time for WU download (Message 1705565)
Posted 8 days ago by Profile HAL9000
This might seem like quite a minor point to bring up but every time I download new WU's I have to wait 5 minutes between batches. Is there any way to reduce this waiting time? Thanks.

The interval limit was put in place to prevent the high end machines from pounding the servers with requests.
28) Message boards : Number crunching : Reddit thinks SETI@home is not useful (Message 1705319)
Posted 9 days ago by Profile HAL9000
When sometimes says "SETI hasn't found anything & isn't useful" I would ask.
If an x-ray is taken looking for broken bones & no broken bones are found. Does that mean that taking x-rays are not useful? It only indicates what was being looked for in that area at that time was not present. In science a negative result is still a result.
29) Message boards : Number crunching : Why this CPU task invalid so soon? (Message 1705314)
Posted 9 days ago by Profile HAL9000
It looks like you had a CPU AP task also go invalid. With your result containing 30/30 pulses vs the two valid results 2/0 pulses.

That system isn't starting to overheat or anything is it?
30) Message boards : Number crunching : Anything relating to AstroPulse tasks (Message 1704980)
Posted 10 days ago by Profile HAL9000
This is the current status As of 22 Jul 2015, 21:10:04 UTC. When I did the original post on this, the "total available channels on disk" was either 503 or 504, I can't remember which one. I may be wrong, but the total available channels on a disk does not change, while splitting files on that disk. If that is indeed the case, then the disk currently being worked is in fact a new disk. If my perspective on this is incorrect, then this senile old man deeply apologies.

total available channels on disk: 420 0
channels in progress: 7 0
channels done: 39 0
channels ended in error: 0 0
total channels to do: 374 0

'On disk' means "in currently accessible storage" - the server disk array where data is stored while being work on.

So, there's no "new disk" involved, just a number of image files stored 'on disk'. When an image file is completely split, the working copy is deleted from local storage, and the total number of channels reduces by 14.


Is there any way to make the "total available channels on disk' static as to when the complete initial loading of the array occurs, that way we see a more accurate accounting of the "channels done" & "total channels to do", also changing the wording from "disk" to maybe "disk array"? Instead of reducing the total available, we increase the number of channels done by x amount, and the channels to do is decreased by 14, which you stated is the amount that is currently being reduced from the total available. This way when one checks the status, we see get a more accurate accounting of what was initially loaded, the total number of channels already split, and the total number of channels remain to split.

I think what you basically want would be for completed data sets to not be removed automatically once they are finished being split. If you wanted the completed data sets to only be removed once new data was added that wouldn't serve your purpose either. Some data sets can be completed while new ones are still being loaded by the system. It is more of a continuous stream of data & more is added when it starts to get low. In the past data was added more often than once a week or two. I seem to recall that new data was added for 3 or 4 days in a row previously, but that may have been when they had more time on their hands.
They could change the display name, but in computer terms "disk", or sometimes disc, is a generic term to refer to any storage medium. In a RAID array it is common to divide the volume into separate "disks". Depending on the complexity of the RAID controller it may also be also to define which individual drives are assigned to which virtual disk.
31) Message boards : Number crunching : Monster GPU Cruncher Build (Message 1704702)
Posted 11 days ago by Profile HAL9000
I asked at the manfacturer (for some weeks, because I had the same idea ;-) , if I could remove the backplate of the HD7990, the support said it's urgently recommended to let installed the backplate.
It cool the PCB and the VRAM.
The 1st VGA card, 1st GPU, just in BIOS, the backplate was so hot that I couldn't touch it.

HD7990 at PCIe slot #1 and #5 (of 7) and it could work.
I didn't tested 3 VGA cards, it wouldn't go because not enough fresh air for cooling.
I have a Super Flower 2000W PSU.

The warning appeared just one time until now.
I guess because - maybe - I changed the jumper of the onboard VGA output 3 times (off, on, off - for tests) and the last time the jumper was very wobbly (and this is a €500 ASUS board).
I think I must repar it with a pliers and press together the contacts.


I'm still confused that the "Windows 8.1 Pro x64" DVD will be not recognized as 'bootable' source.
If I test the 'old' 'Windows 8.1 x64' DVD from my other J1900 PC - it will be recognized as 'bootable' and will install the OS.
What is the problem that the 'Pro' version of Windows 8.1 x64 don't want to install the OS?

I had contact with the online shop - they said opened software they don't take back (they said also, they registered the OS as I bought it. (If I buy an OS, they send Microsoft my address?) So they can't take back also. Is this all true?).
What? I payed €150 for a damaged DVD and that's all?

EDIT: I need the 'Pro' version because it have dual CPU support.

I think I had it backwards before when you stated Os would not boot. I thought you said that the disc would boot on your J1900 system, but not the new one. I missed the fact that you were listing two different disc.
If the only problem is a damaged disc then there are several options. You could install Windows from the non pro disc without entering a key. Then once in Windows enter your Pro key to upgrade the OS & activate it.
Also you could burn a new disc from .iso. MS may allow you to download image to burn (They do this if you buy Windows from the MS Store), or maybe I could put an .iso on my server for you to try. However, I would need to know if you would want English or German language .iso.
32) Message boards : Number crunching : Anything relating to AstroPulse tasks (Message 1704692)
Posted 11 days ago by Profile HAL9000
Only a Star Trek thing

So option C. It was!
33) Message boards : Number crunching : Panic Mode On (99) Server Problems? (Message 1704659)
Posted 11 days ago by Profile HAL9000
Beta is down :( and just as I started to test something, lol

edit.. an no, I didn't break it thank you!!

The might be implementing the database change at Beta first to make sure it will work here next week.
34) Message boards : Number crunching : GPU performance problems (Message 1704657)
Posted 11 days ago by Profile HAL9000
Hi,

yes, Catalist driver is 15.7

Here you have more details...



https://goo.gl/photos/Ke3rU9KHRCq2aFqt9

I'm sorry, I'm not able to link the image properly.

By the way, Mike, why do you recommend me to be sure to free one CPU core?

I believe the recommendation originates from the app developer. However it might have been from those that did initial app testing. With OpenCL apps CPU resources are required.
35) Message boards : Number crunching : Anything relating to AstroPulse tasks (Message 1704624)
Posted 11 days ago by Profile HAL9000
yes, nobody complained about the nexus


My nexus7 and nexus9 both rock for sure. Nothing works on them yet for seti, but I still love them.

I could be mistaken, but I think woohoo was referring to the nexus of ones RAC, or perhaps they meant the apex?
36) Message boards : Number crunching : Monster GPU Cruncher Build (Message 1704619)
Posted 11 days ago by Profile HAL9000
AFAIK, each HD7990 have a PCIe 3.0 x16 plug.
So 2 HD7990 each CPU, 32 PCIe lanes used.

It looks like I have the v5304 BIOS installed.

Thanks for tips and hints (thanks also for PM :-) .
All the day I looked in BIOS what I could set, had set wrong, and so on...
I found no mistake.

Strange happened, I don't know maybe the BIOS is... I don't know...
Today, just 1 HD7990 installed, start, again a BIOS windows say something: 'fatal system error'.
Reboot and I see again the DOS screen.


What ever...

I got the idea, I test the 'old' Windows 8.1 x64 DVD of my J1900 PC.
And... - and... - it worked, Windows started to install. *confused*
So the Windows 8.1 Pro x64 DVD is broken?

BTW, I ordered 4* 'R9 FuryX' VGA cards.

If they are here, and I can't install the OS and/or something don't go like it should, I will never again build my own PC. *annoyed* ;-)

If you are getting random fatal system error messages then there is a problem with the system. Would be poorly seated CPU or RAM. Could also be a faulty component. If the system will not respond correctly with 1 GPU I would not expect >1 to ever work correctly. I would do further troubleshooting to get the system working normal. Then add additional GPUs.
37) Message boards : Number crunching : Using BOINCView. (Message 1704615)
Posted 11 days ago by Profile HAL9000
I prefer it over BOINCTasks as does continuously switch to the task view when I am looking at the hosts info & it is much more responsive in getting the number of tasks on each host.

I just use <allow_remote_gui_rpc>1</allow_remote_gui_rpc> in my cc_config.xml. Rather than enter every IP address on my network as remote hosts. Once a gui_rpc_auth.cfg, cc_config.xml settings & firewall settings are configured BOINCView is ready to go. Personally I don't bother with a highly secure password in my gui_rpc_auth.cfg. As I don't have the ports for BOINC open in my routers.

If you do that then you can select the 2nd radial button that says "Get data via network access" instead of "Get data via direct file access".
38) Message boards : Number crunching : Monster GPU Cruncher Build (Message 1704476)
Posted 12 days ago by Profile HAL9000
I didn't read all the BIOS release note, but this older one caught my eye. If your BIOS version is older I would probably just install the newest version.

BIOS 5304 for Z9PE-D8 WS released to public
1. Support Windows 2012 R2.
2. Improve compatibility for add-on VGAs while 2 CPUs installed.
3. Improve compatibility for add-on VGAs with Ivy-Bridge CPUs.
39) Message boards : Number crunching : Anything relating to AstroPulse tasks (Message 1704473)
Posted 12 days ago by Profile HAL9000
how come nobody complains when their rac shoots upward

When you are shot out of a cannon you don't worry about how high you are going until you start to come back down.
40) Message boards : Number crunching : New cruncher problems (Message 1704348)
Posted 12 days ago by Profile HAL9000
4.4 GHz is the regular turbo boost on this CPU.

Gigabyte motherboards often have an A.I. overclocking function. Which is on by default. I always turn it off immediately. So I didn't even consider that it was running on their system.


Previous 20 · Next 20

Copyright © 2015 University of California