Panic Mode On (101) Server Problems?

Message boards : Number crunching : Panic Mode On (101) Server Problems?
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 . . . 27 · Next

AuthorMessage
Profile Zalster Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 27 May 99
Posts: 5517
Credit: 528,817,460
RAC: 242
United States
Message 1736556 - Posted: 23 Oct 2015, 23:18:34 UTC - in response to Message 1736483.  
Last modified: 23 Oct 2015, 23:19:00 UTC

Well I know I got my combo nvidia/intel macpro book crunching away.

Too bad there isn't a way to list all the imacs/macbook that use the intel to crunch.
ID: 1736556 · Report as offensive
Profile HAL9000
Volunteer tester
Avatar

Send message
Joined: 11 Sep 99
Posts: 6534
Credit: 196,805,888
RAC: 57
United States
Message 1736559 - Posted: 23 Oct 2015, 23:33:04 UTC - in response to Message 1736483.  
Last modified: 23 Oct 2015, 23:33:42 UTC

Here's an interesting phenomena. If you look at the Apps page you see the amount of production for each App. Let's look at these two,
Linux/x86_64 	        7.08 (opencl_nvidia_sah) 	18 Sep 2015, 18:34:25 UTC 	3,143 GigaFLOPS
Mac OS X/64-bit Intel 	7.08 (opencl_nvidia_mac) 	16 Oct 2015, 18:45:36 UTC 	4,943 GigaFLOPS

OK, now let's look at the GPU page here, http://setiathome.berkeley.edu/gpu_list.php. If you look at the associated results you see the Mac results are listed as, No GPU tasks reported. But...How can that be? There are more GigaFLOPS for the Mac nVidia App than the Linux one. Surely some NV GPU is producing those results. Is this some sort of server error or is something else at play?

I would guess not enough results have been reported by those devices in the time period required so that they GPUs are listed.

It could be useful to add another coloumn on the app list after "average computing" with the number of hosts that were used to get that average. Like the "Number of computers" column on the CPU list.
SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[
ID: 1736559 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 1736570 - Posted: 24 Oct 2015, 0:09:44 UTC - in response to Message 1736559.  
Last modified: 24 Oct 2015, 0:13:36 UTC

Here's an interesting phenomena. If you look at the Apps page you see the amount of production for each App. Let's look at these two,
Linux/x86_64 	        7.08 (opencl_nvidia_sah) 	18 Sep 2015, 18:34:25 UTC 	3,143 GigaFLOPS
Mac OS X/64-bit Intel 	7.08 (opencl_nvidia_mac) 	16 Oct 2015, 18:45:36 UTC 	4,943 GigaFLOPS

OK, now let's look at the GPU page here, http://setiathome.berkeley.edu/gpu_list.php. If you look at the associated results you see the Mac results are listed as, No GPU tasks reported. But...How can that be? There are more GigaFLOPS for the Mac nVidia App than the Linux one. Surely some NV GPU is producing those results. Is this some sort of server error or is something else at play?

I would guess not enough results have been reported by those devices in the time period required so that they GPUs are listed.

It could be useful to add another coloumn on the app list after "average computing" with the number of hosts that were used to get that average. Like the "Number of computers" column on the CPU list.

Considering the nVidia Mac App was only released 1 week ago and it has already amassed 5000 GigaFLOPS I don't think that is an accurate guess. There had to have been Many more results submitted in the last 7 days to pull ahead of the Linux App so quickly. Yet there aren't any listings on the GPU page that appears to be updated a couple times a day.
Something else is at play. Looking at the most recent update, there Now appears to be a few listings under the Mac column.
ID: 1736570 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 5 Jul 99
Posts: 4654
Credit: 47,537,079
RAC: 4
United Kingdom
Message 1736647 - Posted: 24 Oct 2015, 10:36:11 UTC - in response to Message 1736559.  

Here's an interesting phenomena. If you look at the Apps page you see the amount of production for each App. Let's look at these two,
Linux/x86_64 	        7.08 (opencl_nvidia_sah) 	18 Sep 2015, 18:34:25 UTC 	3,143 GigaFLOPS
Mac OS X/64-bit Intel 	7.08 (opencl_nvidia_mac) 	16 Oct 2015, 18:45:36 UTC 	4,943 GigaFLOPS

OK, now let's look at the GPU page here, http://setiathome.berkeley.edu/gpu_list.php. If you look at the associated results you see the Mac results are listed as, No GPU tasks reported. But...How can that be? There are more GigaFLOPS for the Mac nVidia App than the Linux one. Surely some NV GPU is producing those results. Is this some sort of server error or is something else at play?

I would guess not enough results have been reported by those devices in the time period required so that they GPUs are listed.

Agree,

Over a week has gone by now, and now they are displayed:

Mac

   1. (1.000) GeForce GTX 680
   2. (0.916) GeForce GTX 780M
   3. (0.760) GeForce GTX 675MX
   4. (0.216) GeForce GTX 680MX
   5. (0.009) GeForce 9400
   6. (0.009) GeForce GT 120 


Claggy
ID: 1736647 · Report as offensive
Profile HAL9000
Volunteer tester
Avatar

Send message
Joined: 11 Sep 99
Posts: 6534
Credit: 196,805,888
RAC: 57
United States
Message 1736669 - Posted: 24 Oct 2015, 13:21:04 UTC - in response to Message 1736647.  
Last modified: 24 Oct 2015, 13:27:56 UTC

Here's an interesting phenomena. If you look at the Apps page you see the amount of production for each App. Let's look at these two,
Linux/x86_64 	        7.08 (opencl_nvidia_sah) 	18 Sep 2015, 18:34:25 UTC 	3,143 GigaFLOPS
Mac OS X/64-bit Intel 	7.08 (opencl_nvidia_mac) 	16 Oct 2015, 18:45:36 UTC 	4,943 GigaFLOPS

OK, now let's look at the GPU page here, http://setiathome.berkeley.edu/gpu_list.php. If you look at the associated results you see the Mac results are listed as, No GPU tasks reported. But...How can that be? There are more GigaFLOPS for the Mac nVidia App than the Linux one. Surely some NV GPU is producing those results. Is this some sort of server error or is something else at play?

I would guess not enough results have been reported by those devices in the time period required so that they GPUs are listed.

Agree,

Over a week has gone by now, and now they are displayed:

Mac

   1. (1.000) GeForce GTX 680
   2. (0.916) GeForce GTX 780M
   3. (0.760) GeForce GTX 675MX
   4. (0.216) GeForce GTX 680MX
   5. (0.009) GeForce 9400
   6. (0.009) GeForce GT 120 


Claggy

The most recent polling at 24 Oct 2015, 0:10:45 looks to have picked them up. The last change to the Top GPUs list on github was: Oct 20, 2014 - @davidpanderson GPU list script: fix bug, and include anonymous platform hosts. So it just took some time for the app to go out & hosts to start returning enough work to be added to the list.
SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[
ID: 1736669 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 1736671 - Posted: 24 Oct 2015, 13:37:56 UTC - in response to Message 1736647.  
Last modified: 24 Oct 2015, 13:41:44 UTC

Here's an interesting phenomena. If you look at the Apps page you see the amount of production for each App. Let's look at these two,
Linux/x86_64 	        7.08 (opencl_nvidia_sah) 	18 Sep 2015, 18:34:25 UTC 	3,143 GigaFLOPS
Mac OS X/64-bit Intel 	7.08 (opencl_nvidia_mac) 	16 Oct 2015, 18:45:36 UTC 	4,943 GigaFLOPS

OK, now let's look at the GPU page here, http://setiathome.berkeley.edu/gpu_list.php. If you look at the associated results you see the Mac results are listed as, No GPU tasks reported. But...How can that be? There are more GigaFLOPS for the Mac nVidia App than the Linux one. Surely some NV GPU is producing those results. Is this some sort of server error or is something else at play?

I would guess not enough results have been reported by those devices in the time period required so that they GPUs are listed.

Agree,

Over a week has gone by now, and now they are displayed:

Mac

   1. (1.000) GeForce GTX 680
   2. (0.916) GeForce GTX 780M
   3. (0.760) GeForce GTX 675MX
   4. (0.216) GeForce GTX 680MX
   5. (0.009) GeForce 9400
   6. (0.009) GeForce GT 120 


Claggy

Strange that when the ATI Apps were released it didn't take a week before some were displayed. I've launched specific hardware and seen it show up within a day back when there were few Linux GPUs being displayed. Not to mention the Mac nVidia AP App has been around for quite a while and recently has had constant results returned from caches yet nothing was being displayed...until I mentioned it. Look at the Mac nVidia AP number, it's quite high comparatively and the slower cards have been working continuously. Those of us that pay attention aren't easily fooled, believe what you wish, within hours of me mentioning it they appeared.
ID: 1736671 · Report as offensive
Profile HAL9000
Volunteer tester
Avatar

Send message
Joined: 11 Sep 99
Posts: 6534
Credit: 196,805,888
RAC: 57
United States
Message 1736677 - Posted: 24 Oct 2015, 14:07:52 UTC - in response to Message 1736671.  
Last modified: 24 Oct 2015, 14:08:59 UTC

Strange that when the ATI Apps were released it didn't take a week before some were displayed. I've launched specific hardware and seen it show up within a day back when there were few Linux GPUs being displayed. Not to mention the Mac nVidia AP App has been around for quite a while and recently has had constant results returned from caches yet nothing was being displayed...until I mentioned it. Look at the Mac nVidia AP number, it's quite high comparatively and the slower cards have been working continuously. Those of us that pay attention aren't easily fooled, believe what you wish, within hours of me mentioning it they appeared.


I'm not sure what the threshold value for a specific GPU to be displayed might be. The CPU list requires 10 hosts before a specific CPU is listed.
I'm not sure that same limit is used on the GPU list. As I have from time to time seem my custom GPU name of "ATI Radeon HD 6870 (Barts XT)" be displayed instead of the BOINC stock "AMD Radeon HD 6790/6850/6870 series (Barts)"

From what I can tell in the code it requires there be at least 1 device in the past 30 days. There is a section where it is getting the results that states $results = BoincResult::enum( "app_version_id in ($av_ids) and create_time > $t and elapsed_time>100 limit 500" );. So it looks like tasks under 100 seconds don't get counted. Which would prevent those that GPU's that only produce invalids ever 5 seconds from getting mixed in.
However I'm not seeing any other limiting factor.
SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[
ID: 1736677 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 5 Jul 99
Posts: 4654
Credit: 47,537,079
RAC: 4
United Kingdom
Message 1736891 - Posted: 25 Oct 2015, 11:55:13 UTC - in response to Message 1736677.  

Strange that when the ATI Apps were released it didn't take a week before some were displayed. I've launched specific hardware and seen it show up within a day back when there were few Linux GPUs being displayed. Not to mention the Mac nVidia AP App has been around for quite a while and recently has had constant results returned from caches yet nothing was being displayed...until I mentioned it. Look at the Mac nVidia AP number, it's quite high comparatively and the slower cards have been working continuously. Those of us that pay attention aren't easily fooled, believe what you wish, within hours of me mentioning it they appeared.


I'm not sure what the threshold value for a specific GPU to be displayed might be. The CPU list requires 10 hosts before a specific CPU is listed.
I'm not sure that same limit is used on the GPU list. As I have from time to time seem my custom GPU name of "ATI Radeon HD 6870 (Barts XT)" be displayed instead of the BOINC stock "AMD Radeon HD 6790/6850/6870 series (Barts)"

From what I can tell in the code it requires there be at least 1 device in the past 30 days. There is a section where it is getting the results that states $results = BoincResult::enum( "app_version_id in ($av_ids) and create_time > $t and elapsed_time>100 limit 500" );. So it looks like tasks under 100 seconds don't get counted. Which would prevent those that GPU's that only produce invalids ever 5 seconds from getting mixed in.
However I'm not seeing any other limiting factor.

And back to:

Mac
No GPU tasks reported

For both Nvidia and Intel GPUs, perhaps there weren't enough hosts/samples in this latest snapshot.

Claggy
ID: 1736891 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 1736894 - Posted: 25 Oct 2015, 12:27:33 UTC - in response to Message 1736891.  

Linux/x86_64 	        7.08 (opencl_nvidia_sah) 	18 Sep 2015, 18:34:25 UTC 	3,191 GigaFLOPS
Linux/x86_64 	        7.08 (opencl_nvidia_100) 	21 May 2015, 23:56:05 UTC 	328 GigaFLOPS
Linux/x86_64 	        7.08 (opencl_nvidia_cc1) 	21 May 2015, 23:56:05 UTC 	156 GigaFLOPS

Mac OS X/64-bit Intel 	7.08 (opencl_nvidia_mac) 	16 Oct 2015, 18:45:36 UTC 	5,806 GigaFLOPS
Mac OS X/64-bit Intel 	7.07 (opencl_nvidia_mac) 	23 Apr 2015, 18:50:41 UTC 	3,906 GigaFLOPS

http://setiathome.berkeley.edu/apps.php
Over Twice as many nVidia Mac Results as linux results, yet SETI says; Mac: No GPU tasks reported

Just as a I said, obviously a reoccurring glitch in the matrix.
A glitch that's been there quite a while.
ID: 1736894 · Report as offensive
Profile HAL9000
Volunteer tester
Avatar

Send message
Joined: 11 Sep 99
Posts: 6534
Credit: 196,805,888
RAC: 57
United States
Message 1736928 - Posted: 25 Oct 2015, 15:03:16 UTC - in response to Message 1736891.  

Strange that when the ATI Apps were released it didn't take a week before some were displayed. I've launched specific hardware and seen it show up within a day back when there were few Linux GPUs being displayed. Not to mention the Mac nVidia AP App has been around for quite a while and recently has had constant results returned from caches yet nothing was being displayed...until I mentioned it. Look at the Mac nVidia AP number, it's quite high comparatively and the slower cards have been working continuously. Those of us that pay attention aren't easily fooled, believe what you wish, within hours of me mentioning it they appeared.


I'm not sure what the threshold value for a specific GPU to be displayed might be. The CPU list requires 10 hosts before a specific CPU is listed.
I'm not sure that same limit is used on the GPU list. As I have from time to time seem my custom GPU name of "ATI Radeon HD 6870 (Barts XT)" be displayed instead of the BOINC stock "AMD Radeon HD 6790/6850/6870 series (Barts)"

From what I can tell in the code it requires there be at least 1 device in the past 30 days. There is a section where it is getting the results that states $results = BoincResult::enum( "app_version_id in ($av_ids) and create_time > $t and elapsed_time>100 limit 500" );. So it looks like tasks under 100 seconds don't get counted. Which would prevent those that GPU's that only produce invalids ever 5 seconds from getting mixed in.
However I'm not seeing any other limiting factor.

And back to:

Mac
No GPU tasks reported

For both Nvidia and Intel GPUs, perhaps there weren't enough hosts/samples in this latest snapshot.

Claggy

The Total is also missing under Intel. Maybe it doesn't show up when there is only one platform? That would make logical sense I suppose.
SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[
ID: 1736928 · Report as offensive
WezH
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 576
Credit: 67,033,957
RAC: 95
Finland
Message 1736934 - Posted: 25 Oct 2015, 15:13:22 UTC - in response to Message 1736391.  

Splitters are struggling to produce MB wu's, no Ready-to-send buffer for a day.


And now they are in Maximum Overdrive. Over 500K MB's ready to send. Haven't seen that before....
ID: 1736934 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14649
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1736937 - Posted: 25 Oct 2015, 15:28:55 UTC - in response to Message 1736934.  

Splitters are struggling to produce MB wu's, no Ready-to-send buffer for a day.

And now they are in Maximum Overdrive. Over 500K MB's ready to send. Haven't seen that before....

That happens when they miss the 'inhibit' signal at high water mark. Not a good sign.
ID: 1736937 · Report as offensive
Profile JaundicedEye
Avatar

Send message
Joined: 14 Mar 12
Posts: 5375
Credit: 30,870,693
RAC: 1
United States
Message 1736966 - Posted: 25 Oct 2015, 16:38:01 UTC

I recently (2 weeks ago) initiated using the Intel GPU portion of my I7 Haswell chips through lunatics with 3 threads processing (as do my nvidia GPUs). Are these GPU activities shown on the comparison chart?

"Sour Grapes make a bitter Whine." <(0)>
ID: 1736966 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 1737033 - Posted: 25 Oct 2015, 22:34:03 UTC - in response to Message 1736928.  
Last modified: 25 Oct 2015, 22:39:37 UTC

Strange that when the ATI Apps were released it didn't take a week before some were displayed. I've launched specific hardware and seen it show up within a day back when there were few Linux GPUs being displayed. Not to mention the Mac nVidia AP App has been around for quite a while and recently has had constant results returned from caches yet nothing was being displayed...until I mentioned it. Look at the Mac nVidia AP number, it's quite high comparatively and the slower cards have been working continuously. Those of us that pay attention aren't easily fooled, believe what you wish, within hours of me mentioning it they appeared.


I'm not sure what the threshold value for a specific GPU to be displayed might be. The CPU list requires 10 hosts before a specific CPU is listed.
I'm not sure that same limit is used on the GPU list. As I have from time to time seem my custom GPU name of "ATI Radeon HD 6870 (Barts XT)" be displayed instead of the BOINC stock "AMD Radeon HD 6790/6850/6870 series (Barts)"

From what I can tell in the code it requires there be at least 1 device in the past 30 days. There is a section where it is getting the results that states $results = BoincResult::enum( "app_version_id in ($av_ids) and create_time > $t and elapsed_time>100 limit 500" );. So it looks like tasks under 100 seconds don't get counted. Which would prevent those that GPU's that only produce invalids ever 5 seconds from getting mixed in.
However I'm not seeing any other limiting factor.

And back to:

Mac
No GPU tasks reported

For both Nvidia and Intel GPUs, perhaps there weren't enough hosts/samples in this latest snapshot.

Claggy

The Total is also missing under Intel. Maybe it doesn't show up when there is only one platform? That would make logical sense I suppose.

Maybe you can elaborate on what you mean by only one platform? Most people consider a single platform OS being one where you can't mix AMD/ATI cards with nVidia cards, such as Linux or Vista. Certainly not OSX where AMD & nVidia drivers are built into the System. Currently you can even download nVidia drivers for the Macs with nVidia cards, I've been able to download NV Webdrivers for my MacPro ever since I can remember.

Let's look at another comparison, this time comparing the results from both platforms. Note the current OSX AMD MBv7 Apps were released the same date as the nVidia MBv7 App and the counters began from Zero for both;

Mac OS X/64-bit Intel 	7.08 (opencl_ati5zc_mac) 	16 Oct 2015, 18:45:36 UTC 	2,788 GigaFLOPS
Mac OS X/64-bit Intel 	7.08 (opencl_ati5_mac) 	        16 Oct 2015, 18:45:36 UTC 	6,073 GigaFLOPS
Mac OS X/64-bit Intel 	7.08 (opencl_ati_mac) 	        16 Oct 2015, 18:45:36 UTC 	2,117 GigaFLOPS
Mac OS X/64-bit Intel 	7.07 (opencl_ati_mac) 	        23 Apr 2015, 18:50:41 UTC 	  552 GigaFLOPS
                                                                                       11,530

Mac OS X/64-bit Intel 	7.08 (opencl_nvidia_mac) 	16 Oct 2015, 18:45:36 UTC 	6,043 GigaFLOPS
Mac OS X/64-bit Intel 	7.07 (opencl_nvidia_mac) 	23 Apr 2015, 18:50:41 UTC 	3,820 GigaFLOPS
                                                                                        9,863

Not much difference there, note the nVidia MBv7 App is still increasing at a good pace.
Not a problem with the AMD Results, the nVidia results Still say, No GPU tasks reported
Strange seeing as how there isn't that much difference between the AMD & NV numbers.
Now here is the difference, note the Total Linux GPU number, 4,312. I'm afraid there are Over Twice as many OSX nVidia results as ALL the Linux GPU results combined. Why some people continue to deny the Obvious is a mystery.
It's Broken.
ID: 1737033 · Report as offensive
Profile HAL9000
Volunteer tester
Avatar

Send message
Joined: 11 Sep 99
Posts: 6534
Credit: 196,805,888
RAC: 57
United States
Message 1737054 - Posted: 26 Oct 2015, 0:12:41 UTC - in response to Message 1737033.  

Strange that when the ATI Apps were released it didn't take a week before some were displayed. I've launched specific hardware and seen it show up within a day back when there were few Linux GPUs being displayed. Not to mention the Mac nVidia AP App has been around for quite a while and recently has had constant results returned from caches yet nothing was being displayed...until I mentioned it. Look at the Mac nVidia AP number, it's quite high comparatively and the slower cards have been working continuously. Those of us that pay attention aren't easily fooled, believe what you wish, within hours of me mentioning it they appeared.


I'm not sure what the threshold value for a specific GPU to be displayed might be. The CPU list requires 10 hosts before a specific CPU is listed.
I'm not sure that same limit is used on the GPU list. As I have from time to time seem my custom GPU name of "ATI Radeon HD 6870 (Barts XT)" be displayed instead of the BOINC stock "AMD Radeon HD 6790/6850/6870 series (Barts)"

From what I can tell in the code it requires there be at least 1 device in the past 30 days. There is a section where it is getting the results that states $results = BoincResult::enum( "app_version_id in ($av_ids) and create_time > $t and elapsed_time>100 limit 500" );. So it looks like tasks under 100 seconds don't get counted. Which would prevent those that GPU's that only produce invalids ever 5 seconds from getting mixed in.
However I'm not seeing any other limiting factor.

And back to:

Mac
No GPU tasks reported

For both Nvidia and Intel GPUs, perhaps there weren't enough hosts/samples in this latest snapshot.

Claggy

The Total is also missing under Intel. Maybe it doesn't show up when there is only one platform? That would make logical sense I suppose.

Maybe you can elaborate on what you mean by only one platform?

I'd be happy to do so. The Top GPU models list, which are talking about, is separated by platform, Windows, Linux, & Mac. Additionally a total column ranking the GPUs across all three platforms. As of the 25 Oct 2015, 16:06:29 UTC update the Total column under Intel is missing. There are only GPUs listed under Windows with Linux & Mac indicating No GPU tasks reported.
I am guessing as there is only data for the one platform the total column is not being included as it would be redundant. At this point all we can do is guess.
SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[
ID: 1737054 · Report as offensive
Profile Sutaru Tsureku
Volunteer tester

Send message
Joined: 6 Apr 07
Posts: 7105
Credit: 147,663,825
RAC: 5
Germany
Message 1737058 - Posted: 26 Oct 2015, 0:23:34 UTC - in response to Message 1736966.  
Last modified: 26 Oct 2015, 0:27:57 UTC

JaundicedEye wrote:
I recently (2 weeks ago) initiated using the Intel GPU portion of my I7 Haswell chips through lunatics with 3 threads processing (as do my nvidia GPUs). Are these GPU activities shown on the comparison chart?


If you use opti apps with app_info.xml file (anonymous platform) then your GPUs are not added to the list.

AFAIK, just stock SETI BOINCs are shown.


Because of this my AMD Radeon R9 Fury X's ('Fiji') are not shown in this list.


Because of: Top GPU models. But I guess it's also because of the Applications overview.
ID: 1737058 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 1737067 - Posted: 26 Oct 2015, 1:33:34 UTC - in response to Message 1737054.  

Strange that when the ATI Apps were released it didn't take a week before some were displayed. I've launched specific hardware and seen it show up within a day back when there were few Linux GPUs being displayed. Not to mention the Mac nVidia AP App has been around for quite a while and recently has had constant results returned from caches yet nothing was being displayed...until I mentioned it. Look at the Mac nVidia AP number, it's quite high comparatively and the slower cards have been working continuously. Those of us that pay attention aren't easily fooled, believe what you wish, within hours of me mentioning it they appeared.


I'm not sure what the threshold value for a specific GPU to be displayed might be. The CPU list requires 10 hosts before a specific CPU is listed.
I'm not sure that same limit is used on the GPU list. As I have from time to time seem my custom GPU name of "ATI Radeon HD 6870 (Barts XT)" be displayed instead of the BOINC stock "AMD Radeon HD 6790/6850/6870 series (Barts)"

From what I can tell in the code it requires there be at least 1 device in the past 30 days. There is a section where it is getting the results that states $results = BoincResult::enum( "app_version_id in ($av_ids) and create_time > $t and elapsed_time>100 limit 500" );. So it looks like tasks under 100 seconds don't get counted. Which would prevent those that GPU's that only produce invalids ever 5 seconds from getting mixed in.
However I'm not seeing any other limiting factor.

And back to:

Mac
No GPU tasks reported

For both Nvidia and Intel GPUs, perhaps there weren't enough hosts/samples in this latest snapshot.

Claggy

The Total is also missing under Intel. Maybe it doesn't show up when there is only one platform? That would make logical sense I suppose.

Maybe you can elaborate on what you mean by only one platform?

I'd be happy to do so. The Top GPU models list, which are talking about, is separated by platform, Windows, Linux, & Mac. Additionally a total column ranking the GPUs across all three platforms. As of the 25 Oct 2015, 16:06:29 UTC update the Total column under Intel is missing. There are only GPUs listed under Windows with Linux & Mac indicating No GPU tasks reported.
I am guessing as there is only data for the one platform the total column is not being included as it would be redundant. At this point all we can do is guess.

Yes, that part is also currently Broken. There are usually 3 entries under the Mac section, Iris Pro, Iris , and 1 other. Since someone kicked it the other day that part has also gone belly up. It has been working almost continuously since my Mac CPU apps were released and I started paying close attention to the page. You see, I have been looking at that page almost Everyday for sometime, so I can speak from observation. That why I get annoyed when I report yet Another Bug and get the same old song & dance about some river in Egypt.
ID: 1737067 · Report as offensive
Cosmic_Ocean
Avatar

Send message
Joined: 23 Dec 00
Posts: 3027
Credit: 13,516,867
RAC: 13
United States
Message 1737133 - Posted: 26 Oct 2015, 8:43:46 UTC - in response to Message 1736937.  

Splitters are struggling to produce MB wu's, no Ready-to-send buffer for a day.

And now they are in Maximum Overdrive. Over 500K MB's ready to send. Haven't seen that before....

That happens when they miss the 'inhibit' signal at high water mark. Not a good sign.

Judging by the pretty steady increase on the weekly graph, I would vote for something is amiss.

There have been a few times in the past couple of years where there was a DB crash that took a while to recover from, and when the DB came back online, there were over 2 million RTS, and it took a surprisingly short amount of time to get that back down to under 300K.. like.. two hours to assign that many WUs. I want to say having that high of an RTS was somewhat common before the move to the co-lo, but has happened once or twice since then.
Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)
ID: 1737133 · Report as offensive
WezH
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 576
Credit: 67,033,957
RAC: 95
Finland
Message 1737233 - Posted: 26 Oct 2015, 17:03:17 UTC - in response to Message 1737133.  

Splitters are struggling to produce MB wu's, no Ready-to-send buffer for a day.

And now they are in Maximum Overdrive. Over 500K MB's ready to send. Haven't seen that before....

That happens when they miss the 'inhibit' signal at high water mark. Not a good sign.

Judging by the pretty steady increase on the weekly graph, I would vote for something is amiss.

There have been a few times in the past couple of years where there was a DB crash that took a while to recover from, and when the DB came back online, there were over 2 million RTS, and it took a surprisingly short amount of time to get that back down to under 300K.. like.. two hours to assign that many WUs. I want to say having that high of an RTS was somewhat common before the move to the co-lo, but has happened once or twice since then.


Or maybe someone did increase RTS -buffer to about 600K? Just my guess...

We wait and see.
ID: 1737233 · Report as offensive
Swibby Bear

Send message
Joined: 1 Aug 01
Posts: 246
Credit: 7,945,093
RAC: 0
United States
Message 1737237 - Posted: 26 Oct 2015, 17:25:41 UTC - in response to Message 1737233.  

[/quote] Or maybe someone did increase RTS -buffer to about 600K? Just my guess... [/quote]

I was about to make the same guess, but add that maybe they are planning some operation during tomorrow's outage that might take a longer-than-usual time. With a bulked-up RTS queue, recovery would be easier.
ID: 1737237 · Report as offensive
Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 . . . 27 · Next

Message boards : Number crunching : Panic Mode On (101) Server Problems?


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.