Gå til innhold

Anbefalte innlegg

Kan noen paste teksten? Er login og krøll, kommer ikke inn

 

Hvordan folder 980-kortene deres?

Har 2 hjemme og har 160.000 PPD

Hatt det slik i 2 uker. Dårlig snitt.

 

Folder uten parametere, prøvd med advanced og beta også.

Klipp fra foldingforumet, tok med alt som er skrevet om tema, skulle brukt spoiler her, men husker ikke / finner ikke!

New Core 18(Windows) Beta Projects: 9104 and 9114

Moderators: Site Moderators, PandeGroup
27 posts • Page 1 of 2 • 12
New Core 18(Windows) Beta Projects: 9104 and 9114

Postby msultan » 17 Nov 2014, 22:42
Hello everyone,
These are the first two of 12 projects on Her2 Kinase. Her2 is a member of the EGFR family of kinases whose amplification has been linked to several types of cances(in particular breast cancer). The aim of p9104 to 9114 is to simulate Her2 its various mutants in order to understand its dynamics. We are also hoping in finding potentially druggable states with in the system.

Details:
Core:18
OS:Win
GPU:Nvidia
PPD:5000
Deadline:13.2
Timeout:10.2
k-factor:0.75
SystemSize: ~47000


Let me know if there are any issues with either of the projects.


Happy folding,
Muneeb

ps. These are Core18 projects for Windows+NVIDIA only.

User avatar
msultan
    Pande Group Member
     
    Posts: 30
    Joined: 25 Jun 2013, 00:27

Top
9104 ?

Postby Grandpa_01 » 18 Nov 2014, 04:51
I see we have a new beta project but e apparentley announced it in the announcement section rather than the beta section any way I have one of these running

P9104 (R10, C0, G0) tpf 00:03:29 PPD ? it is not showing up in the psummary yet to HFM has no Idea 

Fah controll has evened out now to report the same tpf as HFM and is showing 131,421 which is about 20% above a 13000 on a GTX 770 with 327.23 drivers slightley OCed 1293 Mhz

Edit by Mod: Announcement merged into this topic. -B
Image
4 - SuperMicro H8QGi-F AMD 6xxx=192 cores @ 2.8Ghz
3 - SM X9QRI-f+ Intel 4650 = 192 cores/threads @ 3Ghz
2 - I7 980X 4.4Ghz 2 - GTX680
1 - 2700k 4.4Ghz GTX680
Total = 419 cores folding

User avatar
Grandpa_01
     
    Posts: 1780
    Joined: 04 Mar 2009, 09:36

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby msultan » 18 Nov 2014, 07:37
Mea culpa. Those are my projects. I was trying to get multiple projects off the ground today and I completely forgot about making the announcement in the beta forumns. I believe Bruce is going to be merging the topics soon so I will post updates in that thread in the future. I am also updating psummary either tonight or tomorrow morning at the earliest so please look out for that.

Note: I have scaled the PPD down to 4300 from 5000 to better bring them in line with other projects. If these are too low let me know, and I will go back and benchmark more.

User avatar
msultan
    Pande Group Member
     
    Posts: 30
    Joined: 25 Jun 2013, 00:27

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby Kjetil » 18 Nov 2014, 19:45
It is slow on 970, TPF 4m 10s, PPD 86866 P 9114. The same on 980, TPF 3m 42s, PPD 103485 p 9114.
http://folding.extremeoverclocking.com/user_summary.php?s=&u=94811

Kjetil
     
    Posts: 64
    Joined: 14 Apr 2012, 19:56
    Location: Stavanger Norway

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby msultan » 18 Nov 2014, 23:09
@Kjetil I was expecting some variation within the TPF due to slightly different system sizes. Could you please tell me the R/C/G and perhaps if you are averaging the right number of points.
Thanks,
Muneeb

User avatar
msultan
    Pande Group Member
     
    Posts: 30
    Joined: 25 Jun 2013, 00:27

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby Kjetil » 19 Nov 2014, 00:44
Project: 9114 (Run 17, Clone 0, Gen 1) tpf is 3m 48s ppd 96356, 980. Project: 9114 (Run 20, Clone 0, Gen 2) tpf is 3m 52s ppd 96068. 970.

It is slow like 104xx is 87000 - 95000PPD.

Is bonus on or off? 11.19, 3am 4,300 P 1wu from extremeoverclocking.

Kjetil
     
    Posts: 64
    Joined: 14 Apr 2012, 19:56
    Location: Stavanger Norway

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby EXT64 » 19 Nov 2014, 14:24
Well it should be slow (low PPD) on Maxwell compared to 9201, shouldn't it? Core 18 has the software fix for the NVIDIA OpenCL atomics bug which hurts performance on Maxwell (but gets it to run, of course).
Image

EXT64
     
    Posts: 294
    Joined: 10 Apr 2012, 01:54

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby Grandpa_01 » 19 Nov 2014, 14:48
They now appear to be withinn the + - 15% window just a touch towards the - side compared to the 13000 series.

9114 (9, 0, 5), 3 mins 49 secs, 98885, GTX 680 327 drivers
Last edited by Grandpa_01 on 19 Nov 2014, 14:50, edited 1 time in total.

User avatar
Grandpa_01
     
    Posts: 1780
    Joined: 04 Mar 2009, 09:36

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby Kjetil » 19 Nov 2014, 14:49
No it has not the fix for core 18. But it will be better soon i hope.

Kjetil
     
    Posts: 64
    Joined: 14 Apr 2012, 19:56
    Location: Stavanger Norway

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby bruce » 19 Nov 2014, 19:50
EXT64 does h ave a point.

The ultimate question is whether either Core_17 or Core_18 will be more productive once the atomic variable issue is fixed by new drivers. The work-around is known to reduce productivity, but that work-around will eventually be backed out of whatever cores it's actually in (and there's a big question about which versions of which cores contain that work-around).

Nobody can predict what will happen.
How to provide enough info to get helpful support.

bruce
    Site Admin
     
    Posts: 17771
    Joined: 30 Nov 2007, 00:13
    Location: So. Cal.

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby Kjetil » 19 Nov 2014, 21:19
Okay, not agreed. Problems started whit as upgrade. So no is not 100% nvidia, i will 50/50. The driver error did not start whit maxwell.

Kjetil
     
    Posts: 64
    Joined: 14 Apr 2012, 19:56
    Location: Stavanger Norway

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby Kjetil » 19 Nov 2014, 21:42

    msultan wrote:Mea culpa.. I am also updating psummary either tonight or tomorrow morning at the earliest so please look out for that.


Have you update psummary?

Kjetil
     
    Posts: 64
    Joined: 14 Apr 2012, 19:56
    Location: Stavanger Norway

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby HayesK » 20 Nov 2014, 06:39
completed project:9114 run:29 clone:0 gen:3
in progress project:9114 run:3 clone:0 gen:6
windows client 7.4.4, XP64, nvidia driver 327.23, GTX650Ti boost. no problem. ppd in middle of typical range for these GPU
HFM benchmark data below (still no psummary info for this project, had to manually edit the parameters into the HFM projectinfo.tab file)

Code: Select all
     Project ID: 9114
     Core: ZETA
     Credit: 5000
     Frames: 100
    KFactor:  0.75
    Pref Days:  10.2
    Max Days:  13.2

     Name: F60-X58pro-i930-6C-3.8+3x650TiB-XP64-V744 Slot 01
     Number of Frames Observed: 75
     Min. Time / Frame : 00:07:04 - 45,762 PPD
     Avg. Time / Frame : 00:07:09 - 44,965 PPD
     Cur. Time / Frame : 00:07:05 - 45,362 PPD
     R3F. Time / Frame : 00:07:11 - 44,652 PPD
     All  Time / Frame : 00:07:09 - 44,887 PPD
     Eff. Time / Frame : 00:07:10 - 44,769 PPD

     Name: F60-X58pro-i930-6C-3.8+3x650TiB-XP64-V744 Slot 02
     Number of Frames Observed: 100
     Min. Time / Frame : 00:07:07 - 45,281 PPD
     Avg. Time / Frame : 00:07:13 - 44,343 PPD

Image
<- 7-2600K, 6-i7, 23-GPUs -> folding for OCF T32 as HayesK

User avatar
HayesK
     
    Posts: 271
    Joined: 22 Feb 2009, 18:23
    Location: LaPorte, Texas

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby msultan » 20 Nov 2014, 08:33
Thanks for all the reports everyone! It is very strange that the psummary page is not getting updated since I already made the updates from my end and I can see the project summaries when I search for them individually.

http://fah-web.stanford.edu/cgi-bin/fah ... ned?p=9104

I will look into this in more detail tomorrow morning. As for now,it seems to me that the PPD are about right/slightly lower than average. I will keep an eye out for more reports and maybe bump the points up a bit in the next few days if the trend continues.

User avatar
msultan
    Pande Group Member
     
    Posts: 30
    Joined: 25 Jun 2013, 00:27

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby msultan » 20 Nov 2014, 08:37
I am not too familiar with the details about the cores.But there are a few people in lab working on the new core 18 and I can ask them what is causing the speed differential.

User avatar
msultan
    Pande Group Member
     
    Posts: 30
    Joined: 25 Jun 2013, 00:27
New Core 18(Windows) Beta Projects: 9104 and 9114

Moderators: Site Moderators, PandeGroup
27 posts • Page 2 of 2 • 12
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby HayesK » 20 Nov 2014, 15:12
in progress project:9114 run:20 clone:0 gen:6 Name: F59-X58pro-i930-6C-3.84+1x660Ti+1x650TiB-XP64-V744 Slot 01
windows client 7.4.4, XP64, nvidia driver 327.23, GTX660Ti. no problem. ppd ~30% above middle of typical range for this GPU

HFM benchmark data below (still no psummary info for this project, had to manually edit the parameters into the HFM projectinfo.tab file)

Code: Select all
     Project ID: 9114
     Core: ZETA
     Credit: 5000
     Frames: 100


     Name: F59-X58pro-i930-6C-3.84+1x660Ti+1x650TiB-XP64-V744 Slot 01
     Number of Frames Observed: 55
     Min. Time / Frame : 00:04:11 - 100,473 PPD
     Avg. Time / Frame : 00:04:16 - 97,543 PPD
     Cur. Time / Frame : 00:04:11 - 99,738 PPD
     R3F. Time / Frame : 00:04:18 - 96,437 PPD
     All  Time / Frame : 00:04:16 - 97,361 PPD
     Eff. Time / Frame : 00:04:17 - 96,897 PPD

     Name: F60-X58pro-i930-6C-3.8+3x650TiB-XP64-V744 Slot 01
     Number of Frames Observed: 100
     Min. Time / Frame : 00:07:04 - 45,762 PPD
     Avg. Time / Frame : 00:07:09 - 44,965 PPD

     Name: F60-X58pro-i930-6C-3.8+3x650TiB-XP64-V744 Slot 02
     Number of Frames Observed: 100
     Min. Time / Frame : 00:07:07 - 45,281 PPD
     Avg. Time / Frame : 00:07:13 - 44,343 PPD

Image
<- 7-2600K, 6-i7, 23-GPUs -> folding for OCF T32 as HayesK

User avatar
HayesK
     
    Posts: 271
    Joined: 22 Feb 2009, 18:23
    Location: LaPorte, Texas

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby Grandpa_01 » 21 Nov 2014, 09:44
Well you were not kidding about variations, I have 2 - 9114 running right now on 2 different rigs both on GTX 680's both with the same OC on both the cards and CPU's and there is a little more than 1/2 minute difference in TPF both are running the same drivers 327.23, these may be a little on the tough side to get to what may be considered average PPD.

9114 (2, 0, 7), 4 mins 10 secs, PPD 86,852 about 10% below normal for a 13000

9114 (15, 0, 9), 3 mins 34 secs, PPD 109,617 about 10% above normal for a 13000
Image
4 - SuperMicro H8QGi-F AMD 6xxx=192 cores @ 2.8Ghz
3 - SM X9QRI-f+ Intel 4650 = 192 cores/threads @ 3Ghz
2 - I7 980X 4.4Ghz 2 - GTX680
1 - 2700k 4.4Ghz GTX680
Total = 419 cores folding

User avatar
Grandpa_01
     
    Posts: 1780
    Joined: 04 Mar 2009, 09:36

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby msultan » 21 Nov 2014, 21:56
Ya I took a look at the structures for these two runs and the issue is that they have a different number of water molecules(hence different TPF). I will talk to Vijay about what might be a good way to get around this issue since we both believe including such a variety of structures reduces the amount of simulation time needed on F@H but it comes at the expense of slightly uneven point allocation. Sorry about the variation guys, I will think about the best way to reduce it.

    Grandpa_01 wrote:Well you were not kidding about variations, I have 2 - 9114 running right now on 2 different rigs both on GTX 680's both with the same OC on both the cards and CPU's and there is a little more than 1/2 minute difference in TPF both are running the same drivers 327.23, these may be a little on the tough side to get to what may be considered average PPD.

    9114 (2, 0, 7), 4 mins 10 secs, PPD 86,852 about 10% below normal for a 13000

    9114 (15, 0, 9), 3 mins 34 secs, PPD 109,617 about 10% above normal for a 13000

User avatar
msultan
    Pande Group Member
     
    Posts: 30
    Joined: 25 Jun 2013, 00:27

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby Grandpa_01 » 21 Nov 2014, 23:26
Do the water molecules change from generation to generation or just from run to run ?

If i just from run to run you could assign each run it's own point value something like.

9114 (R 2) Cx Gx) = 4300 base points
9114 (R 15) Cx Gx) = 5000 base points

Or you could raise the base value to a point where the avg of the different (R C G) = an equal avg. of the other GPU WU's. This method may promote dumping of the lesser value WU's though. But I do not know that is / would be all that big of a problem since many of the current WU's already promote that ie core 15 vs core 17 and core 18, and 8101 vs 8102, 8103, 8104 and 8105. So in reality it is not a new problem it is just a continuation of an old problem. 

User avatar
Grandpa_01
     
    Posts: 1780
    Joined: 04 Mar 2009, 09:36

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby msultan » 22 Nov 2014, 18:49
Just run to run. I was trying go for the average across runs and its seems like that I at least got the average right. Let me see if I cant move the most egregious offenders either off the WS or put them into a separate project.

Thanks for all the reports so far!

Muneeb

    Grandpa_01 wrote:Do the water molecules change from generation to generation or just from run to run ?

    If i just from run to run you could assign each run it's own point value something like.

    9114 (R 2) Cx Gx) = 4300 base points
    9114 (R 15) Cx Gx) = 5000 base points

    Or you could raise the base value to a point where the avg of the different (R C G) = an equal avg. of the other GPU WU's. This method may promote dumping of the lesser value WU's though. But I do not know that is / would be all that big of a problem since many of the current WU's already promote that ie core 15 vs core 17 and core 18, and 8101 vs 8102, 8103, 8104 and 8105. So in reality it is not a new problem it is just a continuation of an old problem. 

User avatar
msultan
    Pande Group Member
     
    Posts: 30
    Joined: 25 Jun 2013, 00:27

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby 7im » 22 Nov 2014, 19:05
Base points are set by benchmarking and should not change unless there was a mistake made in that process.

Any adjustment has to come from other parts of the bonus system as per standard procedure.
Please do not misinterpret my brevity as dispassion or condescension. I recognize the time you spend reading this forum is time you could use elsewhere, so my short responses save you time.
Please ask for clarification if I was too terse.

User avatar
7im
     
    Posts: 13935
    Joined: 29 Nov 2007, 18:30
    Location: Arizona

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby bruce » 24 Nov 2014, 18:37

    msultan wrote:Just run to run. I was trying go for the average across runs and its seems like that I at least got the average right. Let me see if I cant move the most egregious offenders either off the WS or put them into a separate project.

    Thanks for all the reports so far!

    Muneeb


If a protein undergoes a significant fold, does the box grow/shrink changing the number of water molecules
How to provide enough info to get helpful support.

bruce
    Site Admin
     
    Posts: 17771
    Joined: 30 Nov 2007, 00:13
    Location: So. Cal.

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby msultan » 24 Nov 2014, 20:55

    7im wrote:Base points are set by benchmarking and should not change unless there was a mistake made in that process.

    Any adjustment has to come from other parts of the bonus system as per standard procedure.



Agreed, I think the base points are fine for the number of water molecules in the average R/C/G. However, the problem of +/-10% is coming from the variation across runs of each of which was not separately benchmarked.



    bruce wrote:

        msultan wrote:Just run to run. I was trying go for the average across runs and its seems like that I at least got the average right. Let me see if I cant move the most egregious offenders either off the WS or put them into a separate project.

        Thanks for all the reports so far!

        Muneeb


    If a protein undergoes a significant fold, does the box grow/shrink changing the number of water molecules



Not in these systems since it is computationally hard to define when its "alright" to change the box dimensions. However, we have people in the lab working on a slight variation of this method which could lead to a significant speed(provided it works)

User avatar
msultan
    Pande Group Member
     
    Posts: 30
    Joined: 25 Jun 2013, 00:27

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby bruce » 24 Nov 2014, 21:27
The box is reasonable when a protein grows into the next bos except for the viewer _ except if one molecule gets too near its neighbors. Knowing when the box can shrink would be challenging but might save som time __ and then mess up the benchmark data.

bruce
    Site Admin
     
    Posts: 17771
    Joined: 30 Nov 2007, 00:13
    Location: So. Cal.

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby PinHead » 25 Nov 2014, 03:38
19:55:05:WU01:FS00:Sending unit results: id:01 state:SEND error:NO_ERROR project:9104 run:6 clone:0 gen:12 core:0x18 unit:0x0000000f0a3b1e78546a55f02f1f0c0b
.
.
19:58:40:WU01:FS00:Final credit estimate, 22499.00 points
19:58:40:WU01:FS00:Cleaning up

GTX 570 TPF AVG 5:00

PSummary still not updated and only receiving base points instead of the estimated QRB 22499.00 points.

PinHead
     
    Posts: 368
    Joined: 24 Jan 2012, 05:43

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby HayesK » 30 Nov 2014, 08:07
EOC stats only reported 4300 credit for project:9104 run:10 clone:0 gen:25. The client log Final credit estimate, 23672.00 points.

Noticed that EOC stats has been reporting a few 100K less ppd for my clients than HFM over past few days. Have run a number of p9104-9114 during that same time period and not aware of any problems on my end. Suspect that there may be issue with the QRB for some or all of the series, but my EOC stats had too many wu completed each update to isolate individual wu credit. Might be worth checking a few recently submitted wu to confirm affected projects.

project:9106 run:4 clone:2 gen:0
project:9108 run:0 clone:3 gen:27
project:9106 run:3 clone:6 gen:4
project:9108 run:1 clone:3 gen:22
project:9106 run:0 clone:9 gen:27
project:9108 run:2 clone:4 gen:11
project:9113 run:0 clone:8 gen:14
project:9106 run:3 clone:9 gen:1
project:9110 run:2 clone:3 gen:16
project:9105 run:1 clone:0 gen:25
project:9104 run:18 clone:0 gen:21
project:9114 run:28 clone:0 gen:20
project:9109 run:0 clone:2 gen:26
project:9104 run:22 clone:0 gen:18
project:9110 run:2 clone:4 gen:18
project:9114 run:8 clone:0 gen:31
project:9112 run:0 clone:4 gen:6
project:9107 run:0 clone:6 gen:22
project:9104 run:38 clone:0 gen:6
project:9110 run:1 clone:0 gen:18

portion of the client log for project:9104 run:10 clone:0 gen:25.

Code: Select all
    00:39:38:WU01:FS01:0x18:Completed 2475000 out of 2500000 steps (99%)
    00:44:14:WU01:FS01:0x18:Completed 2500000 out of 2500000 steps (100%)
    00:44:14:WU02:FS01:Connecting to 171.67.108.200:80
    00:44:15:WU02:FS01:Assigned to work server 171.64.65.93
    00:44:15:WU02:FS01:Requesting new work unit for slot 01: RUNNING gpu:0:GK104 [GeForce GTX 660 Ti] from 171.64.65.93
    00:44:15:WU02:FS01:Connecting to 171.64.65.93:8080
    00:44:15:WU02:FS01:Downloading 3.46MiB
    00:44:20:WU02:FS01:Download complete
    00:44:20:WU02:FS01:Received Unit: id:02 state:DOWNLOAD error:NO_ERROR project:9107 run:0 clone:2 gen:34 core:0x18 unit:0x000000230a3b1e81546bd40762ef6f19
    00:44:35:WU01:FS01:0x18:Saving result file logfile_01.txt
    00:44:36:WU01:FS01:0x18:Saving result file checkpointState.xml
    00:44:38:WU01:FS01:0x18:Saving result file checkpt.crc
    00:44:38:WU01:FS01:0x18:Saving result file log.txt
    00:44:38:WU01:FS01:0x18:Saving result file positions.xtc
    00:44:40:WU01:FS01:0x18:Folding@home Core Shutdown: FINISHED_UNIT
    00:44:40:WU01:FS01:FahCore returned: FINISHED_UNIT (100 = 0x64)
    00:44:40:WU01:FS01:Sending unit results: id:01 state:SEND error:NO_ERROR project:9104 run:10 clone:0 gen:25 core:0x18 unit:0x0000001a0a3b1e78546a55f0b42692cf
    00:44:40:WU01:FS01:Uploading 6.64MiB to 171.64.65.84
    00:44:40:WU01:FS01:Connecting to 171.64.65.84:8080
    00:44:40:WU02:FS01:Starting
    00:44:40:WU02:FS01:Running FahCore: C:\\folding\\FAH-V7\\FAHClient/FAHCoreWrapper.exe C:/folding/FAH-V7/cores/web.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/beta/Core_18.fah/FahCore_18.exe -dir 02 -suffix 01 -version 704 -lifeline 1532 -checkpoint 15 -gpu 0 -gpu-vendor nvidia -gpu-vendor=nvidia -noclean -tmax=80 -twait=900
    00:44:40:WU02:FS01:Started FahCore on PID 2436
    00:44:40:WU02:FS01:Core PID:3036
    00:44:40:WU02:FS01:FahCore 0x18 started
    00:44:41:WU02:FS01:0x18:*********************** Log Started 2014-11-30T00:44:41Z ***********************
    00:44:41:WU02:FS01:0x18:Project: 9107 (Run 0, Clone 2, Gen 34)
    00:44:41:WU02:FS01:0x18:Unit: 0x000000230a3b1e81546bd40762ef6f19
    00:44:41:WU02:FS01:0x18:CPU: 0x00000000000000000000000000000000
    00:44:41:WU02:FS01:0x18:Machine: 1
    00:44:41:WU02:FS01:0x18:Reading tar file state.xml
    00:44:42:WU02:FS01:0x18:Reading tar file system.xml
    00:44:43:WU02:FS01:0x18:Reading tar file integrator.xml
    00:44:43:WU02:FS01:0x18:Reading tar file core.xml
    00:44:43:WU02:FS01:0x18:Digital signatures verified
    00:44:43:WU02:FS01:0x18:Folding@home GPU core18
    00:44:43:WU02:FS01:0x18:Version 0.0.3
    00:44:46:WU01:FS01:Upload 14.13%
    00:44:52:WU01:FS01:Upload 29.20%
    00:44:58:WU01:FS01:Upload 45.21%
    00:45:04:WU01:FS01:Upload 61.22%
    00:45:10:WU01:FS01:Upload 77.23%
    00:45:16:WU01:FS01:Upload 92.30%
    00:45:25:WU01:FS01:Upload complete
    00:45:25:WU01:FS01:Server responded WORK_ACK (400)
    00:45:25:WU01:FS01:Final credit estimate, 23672.00 points
    00:45:25:WU01:FS01:Cleaning up
    00:45:41:WU02:FS01:0x18:Completed 0 out of 2500000 steps (0%)
    00:45:41:WU02:FS01:0x18:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
    00:50:31:WU02:FS01:0x18:Completed 25000 out of 2500000 steps (1%)
    00:55:06:WU02:FS01:0x18:Completed 50000 out of 2500000 steps (2%)

User avatar
HayesK
     
    Posts: 271
    Joined: 22 Feb 2009, 18:23
    Location: LaPorte, Texas

Top
Re: New Core 18(Windows) Beta Projects: 9104 and 9114

Postby bollix47 » 30 Nov 2014, 13:48
Summary:

9104 - no bonus
9105/6/7 - bonus
9108/09/10 - not updating DB
9112 - no bonus
9113 - bonus
9114 - no bonus

Details:

project:9106 run:4 clone:2 gen:0 - 23697.3 points
project:9108 run:0 clone:3 gen:27 - no entry in DB
project:9106 run:3 clone:6 gen:4 - 18945.8 points
project:9108 run:1 clone:3 gen:22 - no entry in DB
project:9106 run:0 clone:9 gen:27 - 23779 points
project:9108 run:2 clone:4 gen:11 - no entry in DB
project:9113 run:0 clone:8 gen:14 - 19087.7 points
project:9106 run:3 clone:9 gen:1 - 23637.1 points
project:9110 run:2 clone:3 gen:16 - no entry in DB
project:9105 run:1 clone:0 gen:25 - 18954.4 points
project:9104 run:18 clone:0 gen:21 - 4300 points
project:9114 run:28 clone:0 gen:20 - 4300 points
project:9109 run:0 clone:2 gen:26 - no entry in DB
project:9104 run:22 clone:0 gen:18 - 4300 points
project:9110 run:2 clone:4 gen:18 - no entry in DB
project:9114 run:8 clone:0 gen:31 - 4300 points
project:9112 run:0 clone:4 gen:6 - 4300 points
project:9107 run:0 clone:6 gen:22 - 19362.9 points
project:9104 run:38 clone:0 gen:6 - 4300 points
project:9110 run:1 clone:0 gen:18 - no entry in DB
project:9104 run:10 clone:0 gen:25 - 4300 points
Endret av -alias-
Lenke til kommentar
Videoannonse
Annonse

Hvordan folder 980-kortene deres?

Har 2 hjemme og har 160.000 PPD

Hatt det slik i 2 uker. Dårlig snitt.

 

Folder uten parametere, prøvd med advanced og beta også.

Nå har jeg 970, men de oppfører seg likt. På core_15/18 ligger ppd samlet rundt 150k, med core_17 ser jeg ppd på 500+k.

 

Bruker ingen parametre og får ned en salig blanding. Core_15/18 slettes med mindre de har gått nesten ferdig.

Lenke til kommentar

Ahh mye klarere nå. Noen av 91xx jobbene gir bonus andre ikke. Dette bør de vel fikse opp i. Jeg har noen 670 kort og de leverer omtrent likt PPD på core 17 og 18. Men uten bonus blir det jo ingenting :(

 

Den andre siden av denne saken, er at verdien av forskningen er like verdifull, selv om QRB ikke blir registrert for alle prosjekter (for meg er det hovedårsaken med folding).

 

Er jo en mager trøst. Det er rettet mot bryst kreft spesielt hvis jeg ikke tar feil (de nye 91xx jobbene).

 

Her har du en link til en artikkel som faktisk viser til at F@H har målbare resultater av sin forskning:

 

http://www.geek.com/science/foldinghome-actually-solves-something-1587368/

Endret av Sn1ken
Lenke til kommentar

Det har du helt rett i Sn1ken, men jeg ser jo at det er mange som sletter core 18 jobber fordi de gir liten return. Det er en uheldig utvikling da jeg regner med at disse jobbene er like viktig for forskningen som andre.

Videre er det jo en fordel at jobbene blir fulført fort(man har jo 8.8 dager på å levere en 91xx jobb for bace credit) og jeg har til tider droppet spillingen for å få en jobb fulført fortest mulig. Det drivet mister jeg uten QBR. Videre tror jeg det er flere som tenker som meg på dette området og antall core 18 jobber som blir fullført reduseres av denne uheldige no-QRB'en.

Lenke til kommentar

En del som tyder på at msultan mangler en del erfaring. Mulig det kan være en årsakssammenheng når det kommer til disse nye 91xx jobbene.

 

Men problemet med QRB (da tenker jeg på 0 QRB) er vell i hovedsak gjort rede for ganske nylig? Det er ikke noe folk på jobb i helgene der heller?

 

Regner med at problemet blir adressert og prioritert når folk er tilbake på jobb.

Lenke til kommentar

Leste du linken min, sa det til dem den 19 nov. Så hva skjer, ikke en dritt.

De har enda ikke fått de inn i hfm.net. P91xx jobbene. Så de bare drikker kaffe på jobb.

Så sier de ikke hva som er problemet, sier bare venter på nvidia? Vi vet nvidia og openCL sliter.

Kan de heller ikke si til oss hva som er det virkelige problemet.

Og en annen ting, er andre jeg kan kjøre om de ikke får opp dampen, er møkk lei dem nå.

Lenke til kommentar

Leste du linken min, sa det til dem den 19 nov. Så hva skjer, ikke en dritt.

De har enda ikke fått de inn i hfm.net. P91xx jobbene. Så de bare drikker kaffe på jobb.

 

Jeg så at du etterlyste psummary i HFM (det er jo en annen sak).

 

Men forøvrig må jeg ærlig innrømme at jeg ikke kan oppfatte budskapet om null QRB før mot slutten av tråden. Som i tid vil si ca nå i helgen. Men mulig jeg har oversett noe.

 

Ser det ble diskutert en del om PPD variasjoner i prosjektet osv (men ikke noe om QRB som ikke ble registrert på permanent basis).

Lenke til kommentar

 

Leste du linken min, sa det til dem den 19 nov. Så hva skjer, ikke en dritt.

De har enda ikke fått de inn i hfm.net. P91xx jobbene. Så de bare drikker kaffe på jobb.

 

Jeg så at du etterlyste psummary i HFM (det er jo en annen sak).

 

Men forøvrig må jeg ærlig innrømme at jeg ikke kan oppfatte budskapet om null QRB før mot slutten av tråden. Som i tid vil si ca nå i helgen. Men mulig jeg har oversett noe.

 

Ser det ble diskutert en del om PPD variasjoner i prosjektet osv (men ikke noe om QRB som ikke ble registrert på permanent basis).

 

Re: New Core 18(Windows) Beta Projects: 9104 and 9114

icon_post_target.gifby HayesK » Thu Nov 20, 2014 5:39 am

completed project:9114 run:29 clone:0 gen:3

in progress project:9114 run:3 clone:0 gen:6

windows client 7.4.4, XP64, nvidia driver 327.23, GTX650Ti boost. no problem. ppd in middle of typical range for these GPU

HFM benchmark data below (still no psummary info for this project, had to manually edit the parameters into the HFM projectinfo.tab file)

Står her, 20 november, så drikker bare kaffe :rofl:

Her er min:

Re: New Core 18(Windows) Beta Projects: 9104 and 9114

icon_post_target.gifby Kjetil » Tue Nov 18, 2014 11:44 pm

Project: 9114 (Run 17, Clone 0, Gen 1) tpf is 3m 48s ppd 96356, 980. Project: 9114 (Run 20, Clone 0, Gen 2) tpf is 3m 52s ppd 96068. 970.

 

It is slow like 104xx is 87000 - 95000PPD.

 

Is bonus on or off? 11.19, 3am 4,300 P 1wu from extremeoverclocking

Endret av Kjetil Lura
Lenke til kommentar

 

Re: New Core 18(Windows) Beta Projects: 9104 and 9114

icon_post_target.gifby HayesK » Thu Nov 20, 2014 5:39 am

completed project:9114 run:29 clone:0 gen:3

in progress project:9114 run:3 clone:0 gen:6

windows client 7.4.4, XP64, nvidia driver 327.23, GTX650Ti boost. no problem. ppd in middle of typical range for these GPU

HFM benchmark data below (still no psummary info for this project, had to manually edit the parameters into the HFM projectinfo.tab file)

Står her, 20 november, så drikker bare kaffe :rofl:

 

 

Hehe.

 

Sant nok. Men jeg husker også at msultan skriver i neste post at han har fulgt opp den forespørselen hvor han mener å ha lagt den inn manuelt, men ikke helt har kompetanse til å se hva han har gjort feil.

Han sier også at han skal forhøre seg med noen andre for bistand i saken.

 

Det er jo et ærlig svar slik jeg ser det.

 

Vi får gjøre som tidligere. Gi det tid, og være tollmodige.

 

Vi har nesten doblet lagets poeng siden februar. Se denne gml posten hvor jeg gratulerer laget med 7 milliarder Poeng :)

 

Så på tross av Core 15 og andre småting (som har ført til irritasjon hos enkelte) har vi hatt et jubelår med tanke på produksjon av nettopp poeng. Ligger for øyeblikket med den høyeste produksjonen av PPD noen gang.

 

Hurra for Team Hardware.no og takk så mye for alle bidrag :love:

Lenke til kommentar

Ja jøss laget er ikke mitt problem, men lå på ca 3.7m før as opptateringen. Jeg bryr meg om antall wu ikke poengene. Men vet at laget trenger/vil ha dem. Core 15 p8018(tror jeg det var) gir meg 5700P men bruker ca 2.5 timer, slike jobber liker jeg og der er 9201 glimrende, ca 2t 40m per wu. På 980 kjører på 1300 base.

Lenke til kommentar

Opprett en konto eller logg inn for å kommentere

Du må være et medlem for å kunne skrive en kommentar

Opprett konto

Det er enkelt å melde seg inn for å starte en ny konto!

Start en konto

Logg inn

Har du allerede en konto? Logg inn her.

Logg inn nå
  • Hvem er aktive   0 medlemmer

    • Ingen innloggede medlemmer aktive
×
×
  • Opprett ny...