Posts by HAL9000


log in
21) Message boards : News : Arecibo still threatened with closure. (Message 1795851)
Posted 13 Jun 2016 by Profile HAL9000
Arecibo is what in Carso (Karst) highland is called a dolina. It cannot be dismantled. Also the China FAST is in a dolina.
Tullio

The landform it's in can't be dismantled (at least not without lots of explosives), however all of the structures & equipment can be removed. Arecibo the place will still be there, but not the observatory.

I'm sure we could blow the whole area well below sea level if needed. I mean have you seen our military spending?

I'd like to see the Arecibo Observatory stay funded & running, but on the other hand I hate already seeing 30% of my pay taken in taxes.
22) Message boards : Number crunching : Welcome to the 17 Year Club! (Message 1795541)
Posted 12 Jun 2016 by Profile HAL9000
Well, are here only BOINC addicts welcome or does the "old" Seti@Home also count?
Cause if it counts, i'm in... ;)

BOINC was released 14 years ago. 2002-04-10
SETI@home officially started using BOINC 12 years ago. 2004-06-22

Some of us have been going for 17 years since the project started in 1999.
23) Message boards : Number crunching : LotzaCores and a GTX 1080 FTW (Message 1795376)
Posted 11 Jun 2016 by Profile HAL9000
Thanks! My RAC in the software shows that it has rocketed from 0 to 4600 in those 2 short days. Who knows how high it just might go? :-)


Don't know for sure, but can do some ballparking.

IIRC 8 Xeon cores doing MB, back in cobblestone scale days, used to get about 20K RAC on PreAVX AKv8 code. Since then there's been two main credit drops amounting to x ~30%. You claw back a little for increased throughput with AVX (about 1.5x), so my guess with 48 CPU cores alone (AVX capable + fast memory), would be 20K*6*0.3*1.5 ~= 50K (1 significant digit). Lots of variability, especially if adding AP and weird work mixes and GPUs into the picture.



. . Should not that formula be 20K*6*0.7*1.5 ??

. . Or is that drop "to" 30% not "of" 30%.

. . I am curious to know if the drop has been that large?

The change from MBv6->MBv7 was a drop of 40-50% for some.

I prefer to just use the actual run times to calculate the number of tasks a day & then guesstimate the credit. It looks like their normal AR tasks are running ~2.5 hours. So that gives us ~450 tasks/day. I like to figure 100 credit per task to give a max theoretical RAC value of 45,000. Then I figure 80% of the max for the low end of 36,000. Which would put a RAC of ~40.5K in the middle.
The daily credit values for the past few days on their 48 core host are: 34,506 36,640 46,096 62,035 35,202 39,718 35,010 35,898 44,786. Which averages out to 41099 for the past 9 days.
24) Message boards : Number crunching : CPU time difference (Message 1795368)
Posted 11 Jun 2016 by Profile HAL9000

It was recommended they suspend GPU work several days ago to see how it effected their CPU processing times. No feedback was provided as to the results or if it was tried.
Ivy Bridge & Haswell based CPUs have shown using SETI@home apps on the CPU & iGPU at the same time would cause CPU apps to run more slowly. From one host with a Skylake i5 it was thought that the CPU slowdown may be greater on the newest CPUs. If that is true I would speculate that the increased iGPU speed in Skylake may be causing even more "cache thrashing". It has been speculated that running an app from a less cache heavy app on either the CPU or iGPU may be the solution, but I'm not sure anyone has set down and tested that yet.


. .

. . I5-6400 (2.7GHz) HD530 GPU.

. . Runtime for CPU non VLAR tasks with iGPU crunching approx 3.5 hours

. . Runtime same tasks with GPU use discontinued approx 2 hours

. . Runtime with Lunatics 0.44 running AVX approx 1 hour 15 mins.

. . And since 2 cores are hyperthreading does that not mean there is competition for the associated maths unit? I would think it would be best to try just 2 basic cores under lunatics and compare his productivity then.


Nearly double CPU run time when using the iGPU is similar to previous observations.
My initial posts when I noticed it occurring
A journey: iGPU slowing CPU processing
iGPU tuning
Raistmer's research thread
Loading APU to the limit: performance considerations - ongoing research
It would be interesting to see the results of a CPU where the iGPUs has a dedicated cache to use. Typically only the Iris Pro iGPUs have the addition cache, but the current generation Iris 540 & 550 also list a dedicated iGPU cache. I suspect that little to no CPU slowdown will occur. Similar to the results I found with my Celeron J1900, Bay Trail, CPU when using the iGPU.

To run HT or not is an often debated topic here. I have always seen an increase in work output when using HT.
There are a few different configurations to test how well using HT works in a specific configuration after a baseline using all cores & threads has been found.
1) Disable HT in their BIOS.
2) Leave HT enabled & reduce the number of CPU threads BOINC is allowed to run.
3) Leave HT enabled, reduce the number of CPU threads BOINC is allowed to run, & start BOINC with affinity settings only allowing the physical cores to be used.
Running 4 CPU tasks at once on my i7-860 I found no real noticeable difference using config #3 over #2, but config #1 was slightly worse. Running 8 CPU tasks at once on my i7-860 I found an 11.1% increase in power consumption & a 27.7% increase in work output.

I have not run similar test on my i3-390M system. I will typically just use its Radeon GPU. As it is a notebook it doesn't really have the thermal capacity to run the GPU+CPU full tilt & the GPU can do about as much work as the CPU alone.
25) Message boards : Number crunching : Are some gpu tasks longer now? (Message 1795190)
Posted 10 Jun 2016 by Profile HAL9000

MB8_win_x86_SSE3_OpenCL_NV_r3430_SoG.exe, I thought that was clear from my post, where I mention SoG several times. MB8_win_x86_SSE3_OpenCL_NV_r3430.exe is not a SoG version.
MB8_win_x86_SSE3_OpenCL_NV_r3430_SoG.exe also have the -use_sleep option if one wants to use it.


I guess I am very confused then. So you are saying that MB8_win_x86_SSE3_OpenCL_NV_r3430_SoG.exe IS NOT a SoG app, EVEN THOUGH it ships with the <plan_class>opencl_nvidia_SoG</plan_class> in its aistub file???

I think you might have misread their post.
26) Message boards : Number crunching : Burning rubber... err, CPU's... on an Android phone. (Message 1795098)
Posted 10 Jun 2016 by Profile HAL9000
The default configuration for BOINC on Android stops applications from running when the device battery temp reaches 40ºC(104ºF). You can set it to a lower value by going to Preferences, checking Show advanced preferences and controls... and then scrolling down to MAX battery temperature.
27) Message boards : Number crunching : CPU time difference (Message 1794872)
Posted 10 Jun 2016 by Profile HAL9000
It was also recommended that I try 50% CPUs. I didn't see any point in trying both that and GPU-suspended at the same time. I wasn't seeing anything clearly apples-to-apples in the immediate results, so I was just waiting. Since host RAC was in a slow but steady decline with 50% CPUs and GPU enabled, I have now switched to 100% CPUs and GPU suspended.

An easy-to-use way to do before-and-after testing using the same real-world data would be very useful. If maximizing total global throughput is a goal, seems like that would be something worth working on at Berkeley, but that's just me.


I would probably say that their primary goal is stability of applications & being able to support as many types of hardware as possible. Instead of maximum raw throughput of data processing.
Much of the application development is done by volunteers rather than by the SETI@home project. As the admins just don't have the time, funding, or resources. It is left to the users if they wish to get go deeper & get the absolute most out of their systems.
The way BOINC works does not lend itself to being easily manipulated to throw in work for benchmarking purposes. So most of the tools to do that will operate outside of BOINC, but, for the most part, are fairly simple to use.
28) Message boards : Number crunching : Windows 10 - Yea or Nay? (Message 1794863)
Posted 9 Jun 2016 by Profile HAL9000
As I said (please re-read) update was installed in automatic mode w/o user consent. That's the issue.


Ah, then they must have consented then prior. The EULA is presented during the install process. It's the only way to back out at that point without it installing and requiring reversion.

As I already mentioned in this thread. I left a Windows 7 VM running with auto update enabled. Then I came back it to to find it running Windows 10. The only purpose of that VM was to see if an auto install of Windows 10 would happen.
It would seem that if auto update is enabled it is taken as consent to install Windows 10.
29) Message boards : Number crunching : CPU time difference (Message 1794766)
Posted 9 Jun 2016 by Profile HAL9000
This WU is about as close to apples-and-apples as I think I'll ever get. The computers are similar and even their benchmarks are close. Yet there's a 3.5X difference in CPU time. This is with "50% of CPUs" and GPU tasks enabled.

FTR when the task drops off the list:
Computer 7914479: 3,810 secs
Computer 7973906 (mine): 13,495 secs
Name: blc6_2bit_guppi_57403_69499_HIP11048_OFF_0005.5998.0.22.45.118.vlar
Application (both computers): SETI@home v8 v8.00 windows_intelx86

The difference between the 2 systems is that you are using your integrated Intel GPU for crunching, the other person isn't.
EDIT- also their CPU has 4 physical cores, yours would be 2 physical, 2 Hyperthreaded.
Your CPU is capable of very high clock speeds, but not at the same time as doing heavy work on the internal GPU.


I'd suggest re-enabling all CPU cores, and disable GPU crunching for a while & see what times result.

It was recommended they suspend GPU work several days ago to see how it effected their CPU processing times. No feedback was provided as to the results or if it was tried.
Ivy Bridge & Haswell based CPUs have shown using SETI@home apps on the CPU & iGPU at the same time would cause CPU apps to run more slowly. From one host with a Skylake i5 it was thought that the CPU slowdown may be greater on the newest CPUs. If that is true I would speculate that the increased iGPU speed in Skylake may be causing even more "cache thrashing". It has been speculated that running an app from a less cache heavy app on either the CPU or iGPU may be the solution, but I'm not sure anyone has set down and tested that yet.
30) Message boards : Number crunching : GPU Wars 2016: GTX 1060 July 19th (Message 1794517)
Posted 8 Jun 2016 by Profile HAL9000
Is there a difference between NV & Radeon GPU detection? I noticed my R9 390X has always been detected with 8GB.

Yes. Detection of each card type depends on the (limited) amount of hardware information which each manufacturer makes available through their interface API -

HMODULE cudalib = LoadLibrary("nvcuda.dll");

for NVidia. I think it was Claggy who first noticed that calls to that DLL returned negative values for cards close to 4 GB... Replacing them with 32-bit calls solved the problem.

AMD's equivalent is obviously better at 64-bit maths, but less good at English - BOINC has to look up the card names from a list of ID numbers.

Perhaps OpenCL detection could be used universally. Similar to how it was implemented in BOINC 7.6.23 and later builds for Radeon GPU name detection. Where it reads the OpenCL BoardName value instead of hard coded names for non-CAL GPUs.

But it may be a non issue until there are apps that use a few GB of GPU memory. Like there are at other projects with their CPU apps.
31) Message boards : Number crunching : GPU Wars 2016: GTX 1060 July 19th (Message 1794474)
Posted 8 Jun 2016 by Profile HAL9000
Yes, seen it, thanks. I'll pass all that back to Ray - should be all he needs.

I assume your card is a standard 8 GB memory? BOINC is only seeing 4 GB, but that's normal because it's using a 32-bit subsystem for detection.

Is there a difference between NV & Radeon GPU detection? I noticed my R9 390X has always been detected with 8GB.
08-Jun-2016 11:05:36 [---] Starting BOINC client version 7.4.42 for windows_x86_64
08-Jun-2016 11:05:36 [---] OpenCL: AMD/ATI GPU 0: Hawaii (driver version 1912.5 (VM), device version OpenCL 2.0 AMD-APP (1912.5), 8192MB, 8192MB available, 3802 GFLOPS peak)
08-Jun-2016 11:05:36 [---] OpenCL CPU: Intel(R) Core(TM) i5-4670K CPU @ 3.40GHz (OpenCL driver vendor: Intel(R) Corporation, driver version 4.2.0.148, device version OpenCL 1.2 (Build 148))

08-Jun-2016 11:11:22 [---] Starting BOINC client version 7.6.22 for windows_x86_64
08-Jun-2016 11:11:22 [---] OpenCL: AMD/ATI GPU 0: Hawaii (driver version 1912.5 (VM), device version OpenCL 2.0 AMD-APP (1912.5), 8192MB, 8192MB available, 3802 GFLOPS peak)
08-Jun-2016 11:11:22 [---] OpenCL CPU: Intel(R) Core(TM) i5-4670K CPU @ 3.40GHz (OpenCL driver vendor: Intel(R) Corporation, driver version 4.2.0.148, device version OpenCL 1.2 (Build 148))
32) Message boards : Number crunching : CPU time difference (Message 1793678)
Posted 5 Jun 2016 by Profile HAL9000
Yep, you can't compare i3 CPU times to i5 times as the i3 is 4 virtual cores made out of 2 physical cores whereas the i5 is just 4 pure cores, no comparison at all.

Cheers.

Their i3 host is clearly having some other issue. My old i3-390M can complete work several times faster than what they are presently seeing.
33) Message boards : Number crunching : Open Beta test: SoG for NVidia, Lunatics v0.45 (Message 1793677)
Posted 5 Jun 2016 by Profile HAL9000
HKEY_LOCAL_MACHINE\Software\Wow6432Node

That sounds like a goer. I can ask for something universally present, like

HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Microsoft\Windows NT\CurrentVersion\CurrentVersion

and if I get an answer, I'm on a 64-bit machine? Silence means 32-bit? I'll give it a try. Thanks.

I've found these keys helpful for determining the OS architecture.
HKEY_LOCAL_MACHINE\HARDWARE\RESOURCEMAP\Hardware Abstraction Layer\ACPI x64 platform
HKEY_LOCAL_MACHINE\HARDWARE\RESOURCEMAP\Hardware Abstraction Layer\ACPI x86 platform
34) Message boards : Number crunching : CPU time difference (Message 1793442)
Posted 4 Jun 2016 by Profile HAL9000
Right click on the BOINC icon in the task tray down by the clock, and choose Snooze GPU tasks. That should do it.

That will only cause GPU processing to stop for 1 hr. Using the menu option to select suspend GPU processing will remain active until changed again by the user.
35) Message boards : Number crunching : CPU time difference (Message 1793423)
Posted 4 Jun 2016 by Profile HAL9000
Re: http://setiathome.berkeley.edu/workunit.php?wuid=2173961623

How can two computers rated at roughly the same speed produce the same result, one consuming 8X the CPU time as the other?

Hm... very interesting case.
What about cache (L2 cache) size of your i3 and wingman's i5?

i3-6100 3MB
i5-2500 6MB
http://ark.intel.com/compare/52209,90729
36) Message boards : Number crunching : CPU time difference (Message 1793422)
Posted 4 Jun 2016 by Profile HAL9000
Re: http://setiathome.berkeley.edu/workunit.php?wuid=2173961623

How can two computers rated at roughly the same speed produce the same result, one consuming 8X the CPU time as the other?

All CPU tasks on your i3-6100 are taking a very long time. You may want to use an app like CPUz to confirm that your CPU is running at its rated clock speed. If it isn't perhaps there could be a thermal issue.
If everything looks OK I would suggest suspending GPU processing and see if the CPU times improve.
37) Message boards : Number crunching : No more guppi's=vlars on the gpu please (Message 1793419)
Posted 4 Jun 2016 by Profile HAL9000
I'm firmly in the "If they are sent to my crunchers I'll let them crunch" camp. I would rather one of mine didn't get so many guppi, but it does, so it carries on crunching.

Indeed. The original reason for not sending VLARs to GPUs was that, at the time, they could cause a host to lockup or crash.
I don't see the argument that they run less efficiently on GPUs now as valid. If we are going that route we might as well disallow NV GPUs from downloading AP tasks. Since they are less efficient than Radeon GPUs at processing the tasks.

Having options from the project to tailor which apps and what kind of work run on specific hardware would be nice. Something similar to what Collatz or PrimeGrid have would probably be ideal, but require a lot of work.

Out of curiosity do we know how iGPUs handle these tasks?
38) Message boards : Number crunching : GPU Wars 2016: GTX 1060 July 19th (Message 1793046)
Posted 3 Jun 2016 by Profile HAL9000
I dunno, one part of me feels NV has held back, just because p100 supposedly is a thing, and 1080 is 'only' ~2x970. Sure, bad yields + HBM2 unavailability, but never stopped NV prematurely launching unobtainium before.

Another part is that traditionally AMD has been in bed with m$ via XBox, and now has ties with Samsung and China, so low to medium end mass production seems like a probable advantage there for them.

Last part that bothers me, is that the topend consumer/gaming cards usually require more power. Of course using less power is good for preserving fossil fuels, but there are those with hydro power, nuclear and renewable sources of energy these days. Pushing the 300W PCIe spec limits probably isn't a big deal for some number of customers

So my feeling is the 1080 is mid-high-end, rx480 upper mid, and the true mainstream and ultra-enthusiast platforms are yet to appear.

Looks to me like they're just avoiding competing for the time being, at least until they can build the emerging VR and 4k gaming markets which Intel and m$ have no hope of surviving in. [Consider: Samsung makes 4k screens I believe]

Looking at the trend for the R9 x80 cards the 400 series looks like a heck of a bump in oomph. While being more affordable and efficient.
R9 280 2964 GFLOPs SP 250w $249
R9 380 3476 GFLOPs SP 190w $329
RX 480 5000 GFLOPs SP 150w $199

My R9 390X is only ~6000 GFLOPs. So I look forward to seeing what they do with the rest of the range. If the high end Fury GPUs get a similar treatment I might even end up going that route for one of my system.
39) Message boards : Number crunching : Thoughts on this card, Deal or more likely Crapshoot? (Message 1792881)
Posted 2 Jun 2016 by Profile HAL9000
This seems kind of like a frankenstein card that someone has whipped together? Wierd stuff, for certain, and the performance isn't all that great either. Def seems like kind of a kludge card.

Since everything fits the 760 192-bit GPU specs except the cores. It is likely these cards were built from chips with a large number of dead/non working cores. It looks like you got a few of these. Do they all show 192 shaders?
Using silicone that doesn't meet all the specs for the intended product in a lower end product is a legitimate business practice. It is why you will often see different level products with the same base silicone.
40) Message boards : Number crunching : GPU Wars 2016: GTX 1060 July 19th (Message 1792879)
Posted 2 Jun 2016 by Profile HAL9000
$200, only 150W TDP, and greater than 5 TFLOPS of performance. This might be one cheap beast for us crunchers!

The 1070 is also only 150W TDP, so I would guesstimate the 1060 when it appears, in 3 months, might be as low as 100W TDP and cheaper.

The NV 10 series looks to have a $50 bump in price over the 900 series. So the 1060 will likely be $250 & only ~3500 GFLOPs.


Previous 20 · Next 20

Copyright © 2016 University of California