Page 1 of 2
SMP Frame Times Data Base
Posted: Sun Jan 15, 2012 6:24 am
by Grandpa_01
Since there is such a large variation in PPD on SMP projects I though I would start a data base so the problem might be fixed. Post your frame times and PPD here and any thoughts you may have. You can click on the Goggle docks link for a better view and more complete data list.
https://docs.google.com/spreadsheet/ccc ... W1Gak8tbmc was getting to long you will need to use the links to see the complete list. I will only be posting screen shot of individual user computer high and lows from now on to conserve space.
Project: ____
Core:
Average time/frame: ________ {in hh:mm:ss}
CPU: _______ @ ____ GHz
# of CPU sockets:
# of Physical cores:
# of FAH CPU processes:
# of FAH GPU Clients:
RAM GB installed:
RAM Type: {DDR/DDR2/DDR3}
RAM Speed:
OS name/kernel version
Client:
Running in VM: {Yes/No}
Uploaded with
ImageShack.us
https://docs.google.com/spreadsheet/ccc ... W1Gak8tbmc
Re: SMP Frame Times Data Base
Posted: Sun Jan 15, 2012 7:55 am
by Zagen30
CPU: i7-930 @ 3.71 GHz
# of CPU sockets: 1
# of Physical cores: 4
# of FAH CPU processes: 1
# of FAH GPU Clients: 2
RAM GB installed: 6
RAM Type: DDR3
RAM Speed: 1384 MHz
OS name/kernel version: Win 7 64-bit
Client: v7
Running in VM: No
I'm posting the composite data that FAHWatch7 compiles since I don't particularly want to post the information of over 200 separate WUs:
Code: Select all
Project number: 7000
Count: 2
Failed: 0
Computation time: 950:00
Succes rate: 100%
Avg ppd: 18591.29
Total credit: 7612.00
Avg tpf: 00:02:55
Project number: 7001
Count: 2
Failed: 0
Computation time: 07:28:58
Succes rate: 100%
Avg ppd: 27943.20
Total credit: 8723.00
Avg tpf: 00:02:12
Project number: 7004
Count: 1
Failed: 0
Computation time: 03:51:52
Succes rate: 100%
Avg ppd: 26589.38
Total credit: 4290.00
Avg tpf: 00:02:17
Project number: 7005
Count: 1
Failed: 0
Computation time: 03:44:12
Succes rate: 100%
Avg ppd: 28002.38
Total credit: 4364.00
Avg tpf: 00:02:12
Project number: 7007
Count: 1
Failed: 0
Computation time: 05:48:53
Succes rate: 100%
Avg ppd: 14432.99
Total credit: 3499.00
Avg tpf: 00:03:27
Project number: 7008
Count: 1
Failed: 0
Computation time: 03:49:28
Succes rate: 100%
Avg ppd: 27048.97
Total credit: 4314.00
Avg tpf: 00:02:16
Project number: 7012
Count: 1
Failed: 0
Computation time: 05:00:02
Succes rate: 100%
Avg ppd: 18098.11
Total credit: 3773.00
Avg tpf: 00:02:58
Project number: 7013
Count: 1
Failed: 0
Computation time: 03:37:50
Succes rate: 100%
Avg ppd: 29246.84
Total credit: 4428.00
Avg tpf: 00:02:09
Project number: 7015
Count: 2
Failed: 0
Computation time: 07:37:55
Succes rate: 100%
Avg ppd: 27134.275
Total credit: 8637.00
Avg tpf: 00:02:15
Project number: 7017
Count: 1
Failed: 0
Computation time: 03:50:46
Succes rate: 100%
Avg ppd: 26820.03
Total credit: 4302.00
Avg tpf: 00:02:16
Project number: 7018
Count: 3
Failed: 0
Computation time: 1826:49
Succes rate: 100%
Avg ppd: 17379.56
Total credit: 10771.00
Avg tpf: 00:03:38
[This one was all over the place, with TPFs of 5:18, 2:04, and 3:34]
Project number: 7019
Count: 2
Failed: 0
Computation time: 07:43:13
Succes rate: 100%
Avg ppd: 27087.135
Total credit: 8614.00
Avg tpf: 00:02:17
Project number: 7021
Count: 1
Failed: 0
Computation time: 04:04:36
Succes rate: 100%
Avg ppd: 24575.66
Total credit: 4179.00
Avg tpf: 00:02:25
Project number: 7023
Count: 1
Failed: 0
Computation time: 03:26:45
Succes rate: 100%
Avg ppd: 31604.67
Total credit: 4544.00
Avg tpf: 00:02:02
Project number: 7024
Count: 1
Failed: 0
Computation time: 03:31:20
Succes rate: 100%
Avg ppd: 30596.60
Total credit: 4495.00
Avg tpf: 00:02:05
Project number: 7025
Count: 1
Failed: 0
Computation time: 04:56:45
Succes rate: 100%
Avg ppd: 18401.03
Total credit: 3793.00
Avg tpf: 00:02:55
Project number: 7026
Count: 1
Failed: 0
Computation time: 03:32:23
Succes rate: 100%
Avg ppd: 30359.50
Total credit: 4484.00
Avg tpf: 00:02:06
Project number: 7600
Count: 3
Failed: 0
Computation time: 1 Days, 5 hours, 11 minutes and 4 seconds
Succes rate: 100%
Avg ppd: 15139.7366666667
Total credit: 18413.00
Avg tpf: 00:05:46
Project number: 7610
Count: 29
Failed: 0
Computation time: 10 Days, 4 hours, 57 minutes and 8 seconds
Succes rate: 100%
Avg ppd: 14519.7634482759
Total credit: 147520.00
Avg tpf: 00:05:00
Project number: 7611
Count: 18
Failed: 0
Computation time: 9 Days, 5 hours, 11 minutes and 59 seconds
Succes rate: 100%
Avg ppd: 8505.07111111111
Total credit: 76302.00
Avg tpf: 00:07:17
Project number: 7808
Count: 22
Failed: 0
Computation time: 14 Days, 4 hours, 54 minutes and 40 seconds
Succes rate: 100%
Avg ppd: 18467.0540909091
Total credit: 261851.00
Avg tpf: 00:09:11
Project number: 7809
Count: 20
Failed: 0
Computation time: 13 Days, 8 hours, 36 minutes and 48 seconds
Succes rate: 100%
Avg ppd: 18135.9475
Total credit: 239747.00
Avg tpf: 00:09:30
Project number: 7903
Count: 17
Failed: 1
Computation time: 6 Days, 20 hours, 33 minutes and 7 seconds
Succes rate: 94.12%
Avg ppd: 11751.5970588235
Total credit: 83613.00
Avg tpf: 00:05:55
Project number: 7904
Count: 2
Failed: 0
Computation time: 1545:25
Succes rate: 100%
Avg ppd: 12656.735
Total credit: 8312.00
Avg tpf: 00:04:40
Project number: 7905
Count: 4
Failed: 0
Computation time: 2345:29
Succes rate: 100%
Avg ppd: 13057.695
Total credit: 12972.00
Avg tpf: 00:03:31
Project number: 8001
Count: 22
Failed: 1
Computation time: 1 Days, 5 hours, 41 minutes and 50 seconds
Succes rate: 95.45%
Avg ppd: 25843.505
Total credit: 27859.00
Avg tpf: 00:00:45
Project number: 8004
Count: 9
Failed: 0
Computation time: 954:50
Succes rate: 100%
Avg ppd: 29175.0866666667
Total credit: 12098.00
Avg tpf: 00:00:38
Project number: 8011
Count: 9
Failed: 0
Computation time: 1127:49
Succes rate: 100%
Avg ppd: 35340.7355555556
Total credit: 16812.00
Avg tpf: 00:00:44
Project number: 10083
Count: 1
Failed: 0
Computation time: 03:47:41
Succes rate: 100%
Avg ppd: 27335.38
Total credit: 4329.00
Avg tpf: 00:02:15
Project number: 10084
Count: 3
Failed: 0
Computation time: 1043:07
Succes rate: 100%
Avg ppd: 29980.26
Total credit: 13392.00
Avg tpf: 00:02:06
Project number: 10090
Count: 9
Failed: 0
Computation time: 1 Days, 10 hours, 16 minutes and 56 seconds
Succes rate: 100%
Avg ppd: 27249.5355555556
Total credit: 38915.00
Avg tpf: 00:02:15
Project number: 10126
Count: 2
Failed: 0
Computation time: 1 Days, 9 hours, 6 minutes and 29 seconds
Succes rate: 100%
Avg ppd: 11162.25
Total credit: 15364.00
Avg tpf: 00:09:50
Project number: 10127
Count: 1
Failed: 0
Computation time: 1 Days, 14 hours, 50 minutes and 5 seconds
Succes rate: 100%
Avg ppd: 13746.14
Total credit: 22246.00
Avg tpf: 00:23:04
Project number: 10130
Count: 1
Failed: 0
Computation time: 1741:38
Succes rate: 100%
Avg ppd: 12923.05
Total credit: 9530.00
Avg tpf: 00:10:30
Re: SMP Frame Times Data Base
Posted: Sun Jan 15, 2012 6:51 pm
by Hisuichan
CPU: i5-2500K @ 4326 MHz
GPU: MSI R6870 Hawk
# of CPU sockets: 1
# of Physical cores: 4
# of FAH CPU processes: 1
# of FAH GPU Clients: 1
Note: data might be partially inaccurate as I did run smp:4 in the first few weeks, but then changed to smp:3 as I realised this would work better with the AMD GPU. Take the fluctuations with a pinch of salt.
RAM GB installed: 8 GB
RAM Type: DDR3-1600
RAM Timings: 9-9-9-24-2T DC
OS name/kernel version: Win 7 x64 Ultimate
Client: latest v7
Running in VM: Nope
Took the info from FAHWatch7 as well, stripped it down to the juicy bits and added core versions.
Code: Select all
Project number: 7000
Core: a4
Avg ppd: 24638.7266666667
Avg tpf: 00:02:24
---------------------------
Project number: 7002
Core: a4
Avg ppd: 26479.01
Avg tpf: 00:02:17
---------------------------
Project number: 7003
Core: a4
Avg ppd: 26066.33
Avg tpf: 00:02:19
---------------------------
Project number: 7004
Core: a4
Avg ppd: 26000.015
Avg tpf: 00:02:19
---------------------------
Project number: 7005
Core: a4
Avg ppd: 25923.45
Avg tpf: 00:02:19
---------------------------
Project number: 7006
Core: a4
Avg ppd: 25885.4725
Avg tpf: 00:02:19
---------------------------
Project number: 7007
Core: a4
Avg ppd: 23138.375
Avg tpf: 00:02:30
---------------------------
Project number: 7008
Core: a4
Avg ppd: 27803.395
Avg tpf: 00:02:13
---------------------------
Project number: 7009
Core: a4
Avg ppd: 24373.52
Avg tpf: 00:02:25
---------------------------
Project number: 7010
Core: a4
Avg ppd: 23428.168
Avg tpf: 00:02:30
---------------------------
Project number: 7011
Core: a4
Avg ppd: 22970.29
Avg tpf: 00:02:31
---------------------------
Project number: 7012
Core: a4
Avg ppd: 22583.07
Avg tpf: 00:02:33
---------------------------
Project number: 7013
Core: a4
Avg ppd: 27627.695
Avg tpf: 00:02:13
---------------------------
Project number: 7014
Core: a4
Avg ppd: 33614.00
Avg tpf: 00:01:59
---------------------------
Project number: 7015
Core: a4
Avg ppd: 28048.8075
Avg tpf: 00:02:12
---------------------------
Project number: 7016
Core: a4
Avg ppd: 28102.2725
Avg tpf: 00:02:11
---------------------------
Project number: 7017
Core: a4
Avg ppd: 26123.825
Avg tpf: 00:02:18
---------------------------
Project number: 7018
Core: a4
Avg ppd: 25683.27
Avg tpf: 00:02:20
---------------------------
Project number: 7019
Core: a4
Avg ppd: 29722.9575
Avg tpf: 00:02:07
---------------------------
Project number: 7020
Core: a4
Avg ppd: 27103.17
Avg tpf: 00:02:14
---------------------------
Project number: 7021
Core: a4
Avg ppd: 26247.2875
Avg tpf: 00:02:18
---------------------------
Project number: 7022
Core: a4
Avg ppd: 26444.16
Avg tpf: 00:02:17
---------------------------
Project number: 7023
Core: a4
Avg ppd: 29613.965
Avg tpf: 00:02:07
---------------------------
Project number: 7024
Core: a4
Avg ppd: 27501.2166666667
Avg tpf: 00:02:14
---------------------------
Project number: 7025
Core: a4
Avg ppd: 29256.99
Avg tpf: 00:02:07
---------------------------
Project number: 7026
Core: a4
Avg ppd: 30270.92
Avg tpf: 00:02:05
---------------------------
Project number: 7028
Core: a4
Avg ppd: 26813.75
Avg tpf: 00:02:16
---------------------------
Project number: 7200
Core: a4
Avg ppd: 9649.375
Avg tpf: 00:04:02
---------------------------
Project number: 7610
Core: a4
Avg ppd: 9111.40
Avg tpf: 00:07:03
---------------------------
Project number: 7611
Core: a4
Avg ppd: 4136.155
Avg tpf: 00:09:15
---------------------------
Project number: 7808
Core: a4
Avg ppd: 10101.595
Avg tpf: 00:12:37
---------------------------
Project number: 7809
Core: a4
Avg ppd: 11213.11
Avg tpf: 00:11:59
---------------------------
Project number: 7903
Core: a4
Avg ppd: 6524.19
Avg tpf: 00:09:10
---------------------------
Project number: 7904
Core: a4
Avg ppd: 6474.55
Avg tpf: 00:07:14
---------------------------
Project number: 7905
Core: a4
Avg ppd: 3454.08
Avg tpf: 00:08:35
---------------------------
Project number: 8001
Core: a4
Avg ppd: 17135.3385
Avg tpf: 00:00:55
---------------------------
Project number: 8011
Core: a4
Avg ppd: 21434.295
Avg tpf: 00:01:02
---------------------------
Project number: 10083
Core: a4
Avg ppd: 24916.31
Avg tpf: 00:02:23
---------------------------
Project number: 10084
Core: a4
Avg ppd: 23138.2133333333
Avg tpf: 00:02:34
---------------------------
Project number: 10085
Core: a4
Avg ppd: 27490.33
Avg tpf: 00:02:14
---------------------------
Project number: 10133
Core: a3
Avg ppd: 8010.82
Avg tpf: 00:13:45
---------------------------
Project number: 11020
Core: a3
Avg ppd: 11850.275
Avg tpf: 00:03:18
---------------------------
Project number: 11021
Core: a3
Avg ppd: 10098.344
Avg tpf: 00:03:52
---------------------------
Project number: 11060
Core: a3
Avg ppd: 8411.915
Avg tpf: 00:03:25
---------------------------
Project number: 11061
Core: a3
Avg ppd: 7409.065
Avg tpf: 00:04:12
---------------------------
GPU cores:
Project number: 5736
Core: GPU2
Avg ppd: 3395.40
Avg tpf: 00:02:06
---------------------------
Project number: 11293
Core: OPENMM_OPENCL
Avg ppd: 7529.10735632183
Avg tpf: 00:03:25
---------------------------
Project number: 11294
Core: OPENMM_OPENCL
Avg ppd: 6225.78884615385
Avg tpf: 00:04:33
---------------------------
Re: SMP Frame Times Data Base
Posted: Sun Jan 15, 2012 7:19 pm
by orion
Project: 8001
Core: a4
Average time/frame: 00:00:43
CPU: AMD 8150 @ 4.2 GHz
# of CPU sockets: 1
# of Physical cores: 8
# of FAH CPU processes: 8
# of FAH GPU Clients: 0
RAM GB installed: 4
RAM Type: DDR3
RAM Speed: 1600
OS name/kernel: version w7 Home 64bit
Client: v7.1.43
Running in VM: No
Nice to see you taking your time to do this again grandpa.
Re: SMP Frame Times Data Base
Posted: Sun Jan 15, 2012 9:30 pm
by Grandpa_01
To big of variation in fame times and PPD being ignored, so somebody had to do it.
Re: SMP Frame Times Data Base
Posted: Sun Jan 15, 2012 10:06 pm
by Spongebob25
Project: 7018
Core: 0xa4
Average time/frame: 5 mins 64 secs
CPU: 980 @4.2GHz
# of CPU sockets: 1
# of Physical cores: 6
# of FAH CPU processes: 12
# of FAH GPU Clients: 2 x 480GTXs
RAM GB installed: 12
RAM Type: DDR3 1600
RAM Speed: 8-8-8-24-1T
OS name/kernel version
Client: Win 7 64bit/ v7.1.43
Running in VM: NO
P.S. Your link to the database doesn't work. I get a 404 error when I click on it.
Re: SMP Frame Times Data Base
Posted: Sun Jan 15, 2012 11:09 pm
by 7im
Unless you also include a column for base PPD, showing the bonus PPD only serves to exagerate any normal variations in WU performance, in addition to whatever point your'e trying to make. The normal +/- %age between WUs, even in the same project, get blown up much more when you include a large bonus. +/- 5000 points turns out to be little more than +/- a percent or two.
For example, lines 30 and 32 in the picture above. The two similar projects are within the standard +/- 10% deviation allowed before Pande Group would even consider rebenchmarking them. And yet the difference is 10,000 PPD . Looks like a huge difference, and really is not.
Let's also see where 8xxx WUs land when the incorrect PPD is fixed next week. Then you can remove those mistaken entries that falsely inflate the problem...
But if we do come across real problems, I'll be first in line behind you to support getting the project or projects changed to the right points level.
Re: SMP Frame Times Data Base
Posted: Mon Jan 16, 2012 12:34 am
by Grandpa_01
I am not sure you understand the point I am making. I do not care about and I would venture to say most would not care about the + / - 10% deviation. What matters is the end result which is 50% to 70%, I do not think that is acceptable. In just a few days the new bigadv changes are going into effect and allot of the faster rigs out there will be switching over to smp and they are going to see a very large spread in PPD production when it comes to smp folding because of this 10% + / - allowance = (20% in actuality) in bench variation. I do not know if you looked at the spreadsheet or not but the PPD is all over the place, and all classes or rigs that have reported so far show the same problem. No consistence what so ever in the different SMP projects.
I am sorry I just do not think that this is acceptable.
Re: SMP Frame Times Data Base
Posted: Mon Jan 16, 2012 12:39 am
by Spongebob25
Project: 8001
Core: 0xa4
Average time/frame: 36.38 secs
CPU: 980 @4.2GHz
# of CPU sockets: 1
# of Physical cores: 6
# of FAH CPU processes: 12
# of FAH GPU Clients: 2 x 480GTXs
RAM GB installed: 12
RAM Type: DDR3 1600
RAM Speed: 8-8-8-24-1T
OS name/kernel version
Client: Win 7 64bit/ v7.1.43
Running in VM: NO
Re: SMP Frame Times Data Base
Posted: Mon Jan 16, 2012 3:51 am
by 7im
Grandpa_01 wrote:...
I do not know if you looked at the spreadsheet or not but the PPD is all over the place, and all classes or rigs that have reported so far show the same problem. No consistence what so ever in the different SMP projects.
I am sorry I just do not think that this is acceptable.
Yes, I looked at the chart above, and more than half the WUs are from the 8xxx serious of projects with the screwed up benchmark and abnormally high PPD. That problem is already going to be fixed. Rip those out and things start to look much more normal. The chart is a sham until you pull out that known anomalous data.
Re: SMP Frame Times Data Base
Posted: Mon Jan 16, 2012 4:13 am
by Grandpa_01
7im wrote:Grandpa_01 wrote:...
I do not know if you looked at the spreadsheet or not but the PPD is all over the place, and all classes or rigs that have reported so far show the same problem. No consistence what so ever in the different SMP projects.
I am sorry I just do not think that this is acceptable.
Yes, I looked at the chart, and more than half the WUs are from the 8xxx serious of projects with the screwed up benchmark and abnormally high PPD. That problem is already going to be fixed. Rip those out and things start to look much more normal. The chart is a sham until you pull out that known anomalous data.
7im are you even able to acces the spreadsheet page from the link's there are 113 entries on the spread sheet and only 17 of them are the 8xxx project.
From what you are saying I do not think you are able to or have not looked. There are several projects that give as many PPD as the 8xxx Wu's do the only 8xxx WU that is out of line is 8011. While there are several projects that are way low.
Re: SMP Frame Times Data Base
Posted: Mon Jan 16, 2012 7:38 am
by 7im
The link to the google doc was 404, so no, I couldn't review it. It seems to be working now, but I was just leaving. Maybe tomorrow... Thanks.
Re: SMP Frame Times Data Base
Posted: Mon Jan 16, 2012 8:55 am
by darkbasic
On a 2x uniprocessor dual core machine:
Project 7705 PPD 3101
Project 7025 PPD 857
Project 7025 has a 4 days deadline and an 1.5 days ETA. WTF!? What's the point running uniprocessor instead of SMP? This is a 3GHz Athlon64 X2!
Re: SMP Frame Times Data Base
Posted: Mon Jan 16, 2012 12:02 pm
by bollix47
@darkbasic
Please check your PPD again. I think you'll find that you've switched your figures around and that it's the P7705 that's getting 857 and P7025 is getting the 3101.
Just going by what I see on my uniprocessor PPDs.
Re: SMP Frame Times Data Base
Posted: Mon Jan 16, 2012 1:25 pm
by ChasR
V7 is telling me that 80xx is fixed now. Performance on p8004 a 2600K @ 4.8 is about 24.5 sec/frame and ppd is reported to be ~29,000, down from ~59,600. So this WU went from producing 65% more than the average a3 WU to 18% less. Kfactor changed from 3 to .75. TIme to switch back to v6.34 to avoid 80xx WUs.
