Core i7 Hyperthreading on or off? 1 or 2 smp clients?

[Pootank]

Currently using core i7 920 at 2.8 (multi 21 x 133, qpi link at 2.4 ram at 534mhz)

and a 4780x2 at stock on vista 32 no affinity changer.

PPD of ~5000, 2.6k from gpu client and 2.3k from 1 smp client.

The smp client is running 4 processes at 50% cpu usage (~13% per process on taskmgr), and the gpu client using 1 process at 13%.

If i run a 2nd smp client the ppd for the gpu client drops.

This is my main pc, it folds while im browsing/tv/film watching/sleeping, I turn the clients off to play games.

Do I need to turn hyperthreading off and use one smp client, or use 2 clients but set each client to use 3 and 4 cores, or 7 and 8?

Thanks in advance

[WangFeiHong]

Hmm this is interesting, can SMP detect the number of *physical* cores rather than virtual cores?

If it detects 4 virtual cores (with another 4 unused), then it would assign quad-core WUs, which the 4 virtual/2physical cores might not be able to handle

[Pootank]

Code: Select all


Note: Please read the license agreement (fahfooksake.exe -license). Further
use of this software requires that you have read and accepted this agreement.

8 cores detected
If you see this twice, MPI is working
If you see this twice, MPI is working


--- Opening Log file [December 1 11:41:55 UTC]


# Windows SMP Console Edition #################################################
###############################################################################

                       Folding@Home Client Version 6.22 SMP Beta2

                          http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: C:\fah 1
Executable: C:\fah 1\fahfooksake.exe
Arguments: -smp


This is from the start of the cmd for the smp, so it detects 8 cores, and then runs 4 processes over the 8 virtual cores.

[toTOW]

Back in the old time where P4 Ht were common, the official recommendations were :

- turn HT off and run one Uniprocessor client
- leave HT on and run one Uniprocessor client

The story is a bit different when you have multiple physical core and HT : you have to make that each Fahcore uses a different physical core. If you don't do that, you'll more likely to fully load two physical cores (4 vituals), and leave the two other core idle.

The official recommendation will be to run a unique SMP client, and to manually assign each fahcore to a physical CPU core. You'll use the benefits of HT to feed the GPU

Here's what I'd do :

CPU0 (Physical core 0, logical core 0) Fahcore_a1
CPU1 (Physical core 0, logical core 1)
CPU2 (Physical core 1, logical core 0) Fahcore_a1
CPU3 (Physical core 1, logical core 1)
CPU4 (Physical core 2, logical core 0) Fahcore_a1
CPU5 (Physical core 2, logical core 1)
CPU6 (Physical core 3, logical core 0) Fahcore_a1
CPU7 (Physical core 3, logical core 1)

You can use any of the other CPU to run the Fahcore_11 for the GPU.

Unofficial setup with two SMP clients will use the 8 cores automatically, but I'd recommend to set affinities too :

CPU0 (Physical core 0, logical core 0) Fahcore_a1
CPU1 (Physical core 0, logical core 1) Fahcore_a1
CPU2 (Physical core 1, logical core 0) Fahcore_a1
CPU3 (Physical core 1, logical core 1) Fahcore_a1
CPU4 (Physical core 2, logical core 0) Fahcore_a1
CPU5 (Physical core 2, logical core 1) Fahcore_a1
CPU6 (Physical core 3, logical core 0) Fahcore_a1
CPU7 (Physical core 3, logical core 1) Fahcore_a1

A CPU will be shared between GPU core and SMP core.

Always make sure that your GPU client is set at higher priority (slightly higher/low) than your SMP client (idle).

Here are two tools you might need to set affinities, and to remember them :
Baoba's configurable affinity changer : viewtopic.php?f=14&t=5564
Bill2's Process Manager : http://www.bill2-software.com/processmanager/ (site is in French)

Also, if you choose to run two SMP client, you might have troubles closing one without loosing work in the other client ... a clean windows shutdown shouldn't kill work, so it is better if it is possible.

edit : here is a thread about dual quad cores and running two Windows SMP clients : viewtopic.php?f=46&t=7144

[Pootank]

Thanks for your help toTDW
I had a go with prifinitty but that would only let me change the affinity of 4 of the a1 cores.

I had a go with Boaba's program but its over my head in term of how to set it up. I need a nice left click one number 0-7 type program im afraid :S. I've sent a log of the auto cpu detect to Boaba at his request, hope it helps

I've now got 2 smp clients running with 1gpu client
gpu has been set to slightly higher priority and is running at 2600 ppd now so that seems to have worked.
fah1 is at 1333 ppd and fah2 has completed the 2nd % mark yet so its not reporting tho I'll update this post when it does

One more question, should I be running the clients with the cpu affinity lock on or off? I tried both ways and it didnt help with the error i was getting using prifinitty

EDIT: fah1 @1218
fah2 @ 754
GPU @ 2660

Total 4630 which is worse that just running one smp client hmmmmmmmmmm

2nd Edit: first log I sent to Boaba with the pair'ing showing

Code: Select all

01/12/2008 16:41:25: WARNING elevatePrivileges: Cannot elevate privileges to handle affinity for processes running under other user accounts.
Auto detecting CPU configuration. Please wait.
01/12/2008 16:41:29: Found 8 CPUs:
    CPU0 has mask 00000001
    CPU1 has mask 00000002
    CPU2 has mask 00000004
    CPU3 has mask 00000008
    CPU4 has mask 00000010
    CPU5 has mask 00000020
    CPU6 has mask 00000040
    CPU7 has mask 00000080
01/12/2008 16:41:29: Found 4 PAIRs:
    PAIR0<CPU0,CPU4> has mask 00000011
    PAIR1<CPU1,CPU5> has mask 00000022
    PAIR2<CPU2,CPU6> has mask 00000044
    PAIR3<CPU3,CPU7> has mask 00000088
01/12/2008 16:41:29: WARNING CWinAFCDlg::TraverseProcessList: There are no valid application profiles. The program will auto-pause.

01/12/2008 17:04:07: WARNING elevatePrivileges: Cannot elevate privileges to handle affinity for processes running under other user accounts.
01/12/2008 17:04:07: DEBUG CWinAFCApp::InitInstance: Input file is "C:\Users\me\Desktop\WinAFC-0.9.1\affinityinput.txt"
01/12/2008 17:04:07: DEBUG CWinAFCDlg::CWinAFCDlg: Input file: "C:\Users\me\Desktop\WinAFC-0.9.1\affinityinput.txt", Path: "C:\Users\me\Desktop\WinAFC-0.9.1\"
01/12/2008 17:04:07: DEBUG CWinAFCDlg::initializeSystemConfiguration: System has 8 CPUs, activeProcessorMask=0x000000FF
01/12/2008 17:04:07: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU0 has affinity mask 0x00000001
01/12/2008 17:04:07: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU1 has affinity mask 0x00000002
01/12/2008 17:04:07: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU2 has affinity mask 0x00000004
01/12/2008 17:04:07: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU3 has affinity mask 0x00000008
01/12/2008 17:04:07: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU4 has affinity mask 0x00000010
01/12/2008 17:04:07: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU5 has affinity mask 0x00000020
01/12/2008 17:04:07: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU6 has affinity mask 0x00000040
01/12/2008 17:04:07: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU7 has affinity mask 0x00000080
01/12/2008 17:04:09: DEBUG first_thread_bdwth: Precision timer has 14318180 ticks/second
Auto detecting CPU configuration. Please wait.
01/12/2008 17:04:09: DEBUG maketest_bdwth: CPU detection finished? 1
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU5 is 6848.51 MB/s    (0.509148)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU3 is 6974.7 MB/s    (0.518529)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU2 is 7264.38 MB/s    (0.540065)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU7 is 7357.84 MB/s    (0.547014)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU6 is 7693.99 MB/s    (0.572004)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU6 and CPU7 is 7954.54 MB/s    (0.591375)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU6 is 7970.35 MB/s    (0.59255)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU4 is 8170.2 MB/s    (0.607408)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU5 and CPU6 is 8445.44 MB/s    (0.62787)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU6 is 8658.83 MB/s    (0.643734)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU5 is 8860.98 MB/s    (0.658763)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU3 is 8978.89 MB/s    (0.667529)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU5 and CPU7 is 9054.55 MB/s    (0.673154)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU6 is 9070.43 MB/s    (0.674335)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU4 and CPU6 is 9083.99 MB/s    (0.675343)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU4 is 9380.82 MB/s    (0.69741)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU7 is 9417.3 MB/s    (0.700122)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU4 and CPU5 is 9598.61 MB/s    (0.713602)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU4 and CPU7 is 9640.33 MB/s    (0.716704)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU4 is 9776.38 MB/s    (0.726818)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU5 is 9898.54 MB/s    (0.7359)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU3 is 10004 MB/s    (0.743741)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU7 is 10018 MB/s    (0.744782)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU1 is 10039.2 MB/s    (0.746355)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU2 is 10054 MB/s    (0.747456)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU7 is 11620.2 MB/s    (0.863899)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU5 is 12231.1 MB/s    (0.909312)
01/12/2008 17:04:09: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU4 is 13450.9 MB/s    (1)
01/12/2008 17:04:09: DEBUG build_cpu_sets: Min number visited=0, Max number visited=1
01/12/2008 17:04:09: ERROR build_cpu_sets: found sets of different sizes, 1-2.
01/12/2008 17:04:09: ERROR build_cpu_sets: Please send the above error message and the numbers listed below to: baobab13@users.sourceforge.net
01/12/2008 17:04:09: ERROR build_cpu_sets: Unable to determine the logical CPU sets. Use only physical CPU IDs.
01/12/2008 17:04:11: DEBUG CWinAFCDlg::OnFinalizeCPUDetection: finalizing CPU detection. pendingTraverseProcesses = 0, afcState=1, oldState=0, savedOldState=0
01/12/2008 17:04:11: DEBUG CWinAFCDlg::parse_input_file: TestMode set to 1.
01/12/2008 17:04:11: DEBUG CWinAFCDlg::parse_input_file: ConsoleLogLevel set to 4.
01/12/2008 17:04:11: WARNING CWinAFCDlg::TraverseProcessList: There are no valid application profiles. The program will auto-pause.

[Sahkuhnder]

I suspect this won't be the last time these questions arise.

Perhaps this information would make a good addition to the official FAQs?

[baobab]

Hi Pootank, thanks for the log.

If you use two SMP clients, you can set your affinity profiles as follows:

config 1) first SMP uses the first virtual cores of all physical cores, with the most expensive process on the first physical core and the least expensive process on the fourth physical core.
The second SMP uses the second virtual cores of all physical cores, but in reverse order. The most expensive process on the fourth physical core and the least expensive on the first physical core.

Add these two lines to affinityinput.txt

Code: Select all

PathToSMP1\FahCore_*.exe := PAIR0::CPU0+PAIR1::CPU0+PAIR2::CPU0+PAIR3::CPU0 [assign=1,resource=cpuuse,policy=pseudobalanced,priority=idle]

PathToSMP2\FahCore_*.exe := PAIR3::CPU1+PAIR2::CPU1+PAIR1::CPU1+PAIR0::CPU1 [assign=1,resource=cpuuse,policy=pseudobalanced,priority=idle]


Replace PathToSMP1 and PathToSMP2 with the actual paths to your SMP clients. Also, make sure each rule is on a single line (two lines in total).

config 2) first SMP uses both virtual cores of the first two physical cores. The second SMP uses the last two physical cores:

Code: Select all

PathToSMP1\FahCore_*.exe := PAIR0+PAIR1 [assign=1,resource=cpuuse,policy=pseudobalanced,priority=idle]

PathToSMP2\FahCore_*.exe := PAIR2+PAIR3 [assign=1,resource=cpuuse,policy=pseudobalanced,priority=idle]


Replace PathToSMP1 and PathToSMP2 with the actual paths to your SMP clients. Also, make sure each rule is on a single line (two lines in total).

You can experiment with these two configs to see if one of them is more productive. It is hard to tell which will do better without an actual experiment.

I guess you can let the gpu float all over the place. You can add a rule just to set its priority to something higher:

Code: Select all

PathToGPUClient\FahCore_*.exe := SKIP [priority=belownormal]


This rule does not change its affinity but only its priority. Optionally you can set the affinity mask for the GPU if you know where you would like to have it run.

[Pootank]

I've tried the first setup as test only

Code: Select all

C:\fah1\FahCore_*.exe := PAIR0::CPU0+PAIR1::CPU0+PAIR2::CPU0+PAIR3::CPU0 [assign=1,resource=cpuuse,policy=pseudobalanced,priority=idle]
C:\fah2\FahCore_*.exe := PAIR3::CPU1+PAIR2::CPU1+PAIR1::CPU1+PAIR0::CPU1 [assign=1,resource=cpuuse,policy=pseudobalanced,priority=idle]
C:\Users\me\AppData\Roaming\Folding@home-gpu\FahCore_*.exe := SKIP [priority=belownormal]

and got this back

Code: Select all

WinAFC 0.9.1, (c) 2008 <baobab13@users.sourceforge.net>

DEBUG Args::parseCommandLine: using debug level 3
DEBUG CWinAFCApp::InitInstance: Log file is "C:\Users\me\Desktop\WinAFC-0.9.1\affinitylog.txt"
WARNING elevatePrivileges: Cannot elevate privileges to handle affinity for processes running under other user accounts.
DEBUG CWinAFCApp::InitInstance: Input file is "C:\Users\me\Desktop\WinAFC-0.9.1\affinityinput.txt"
DEBUG CWinAFCDlg::CWinAFCDlg: Input file: "C:\Users\me\Desktop\WinAFC-0.9.1\affinityinput.txt", Path: "C:\Users\me\Desktop\WinAFC-0.9.1\"
DEBUG CWinAFCDlg::initializeSystemConfiguration: System has 8 CPUs, activeProcessorMask=0x000000FF
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU0 has affinity mask 0x00000001
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU1 has affinity mask 0x00000002
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU2 has affinity mask 0x00000004
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU3 has affinity mask 0x00000008
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU4 has affinity mask 0x00000010
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU5 has affinity mask 0x00000020
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU6 has affinity mask 0x00000040
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU7 has affinity mask 0x00000080
DEBUG first_thread_bdwth: Precision timer has 14318180 ticks/second
Auto detecting CPU configuration. Please wait.
|==================================================| 100.0%
DEBUG maketest_bdwth: CPU detection finished? 1
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU2 is 7206.4 MB/s  (0.530775)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU5 is 7226.57 MB/s         (0.532261)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU3 is 7295.33 MB/s         (0.537325)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU7 is 7470.25 MB/s         (0.550209)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU6 is 8394.71 MB/s         (0.618298)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU5 is 8552.7 MB/s  (0.629935)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU3 is 8927.38 MB/s         (0.657531)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU4 is 9244.62 MB/s         (0.680897)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU6 is 9257.32 MB/s         (0.681832)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU5 and CPU7 is 9315.29 MB/s         (0.686102)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU6 and CPU7 is 9418.5 MB/s  (0.693704)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU4 and CPU5 is 9428.46 MB/s         (0.694437)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU5 and CPU6 is 9497.53 MB/s         (0.699524)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU4 is 9498.4 MB/s  (0.699588)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU7 is 9503.76 MB/s         (0.699983)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU6 is 9649.71 MB/s         (0.710733)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU4 and CPU6 is 9761.67 MB/s         (0.71898)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU4 is 9805.59 MB/s         (0.722214)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU4 and CPU7 is 9899.14 MB/s         (0.729104)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU5 is 9989.31 MB/s         (0.735746)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU7 is 10121.1 MB/s         (0.745452)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU6 is 10143.6 MB/s         (0.747109)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU1 is 10184.2 MB/s         (0.750098)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU2 is 10184.3 MB/s         (0.750104)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU3 is 10194.2 MB/s         (0.750835)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU7 is 11819.2 MB/s         (0.870521)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU5 is 12735.5 MB/s         (0.938015)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU4 is 13577.1 MB/s         (1)
DEBUG build_cpu_sets: Min number visited=0, Max number visited=4
ERROR build_cpu_sets: found sets of different sizes, 1-5.
ERROR build_cpu_sets: Please send the above error message and the numbers listed below to: baobab13@users.sourceforge.net
1, 2, 7206.4
2, 5, 7226.57
2, 3, 7295.33
2, 7, 7470.25
2, 6, 8394.71
3, 5, 8552.7
1, 3, 8927.38
2, 4, 9244.62
3, 6, 9257.32
5, 7, 9315.29
6, 7, 9418.5
4, 5, 9428.46
5, 6, 9497.53
3, 4, 9498.4
1, 7, 9503.76
1, 6, 9649.71
4, 6, 9761.67
1, 4, 9805.59
4, 7, 9899.14
0, 5, 9989.31
0, 7, 10121.1
0, 6, 10143.6
0, 1, 10184.2
0, 2, 10184.3
0, 3, 10194.2
3, 7, 11819.2
1, 5, 12735.5
0, 4, 13577.1
ERROR build_cpu_sets: Unable to determine the logical CPU sets. Use only physical CPU IDs.
DEBUG CWinAFCDlg::OnFinalizeCPUDetection: finalizing CPU detection. pendingTraverseProcesses = 0, afcState=1, oldState=0, savedOldState=0
DEBUG CWinAFCDlg::parse_input_file: TestMode set to 1.
DEBUG CWinAFCDlg::parse_input_file: ConsoleLogLevel set to 4.
DEBUG CWinAFCDlg::parse_input_file: TimeInterval set to 3600.
DEBUG CWinAFCDlg::parse_input_file: Found profile line: >>C:\fah1\FahCore_*.exe := PAIR0::CPU0+PAIR1::CPU0+PAIR2::CPU0+PAIR3::CPU0 [assign=1,resource=cpuuse,pol
icy=pseudobalanced,priority=idle]<<
WARNING process_nested_cpus::operator (): no logical CPU set of type PAIR was detected on this machine.
WARNING process_nested_cpus::operator (): no logical CPU set of type PAIR was detected on this machine.
WARNING process_nested_cpus::operator (): no logical CPU set of type PAIR was detected on this machine.
WARNING process_nested_cpus::operator (): no logical CPU set of type PAIR was detected on this machine.
WARNING CWinAFCDlg::parse_input_file: Ignoring profile line: C:\fah1\FahCore_*.exe := PAIR0::CPU0+PAIR1::CPU0+PAIR2::CPU0+PAIR3::CPU0 [assign=1,resource=cpuuse,
policy=pseudobalanced,priority=idle]
DEBUG CWinAFCDlg::parse_input_file: Found profile line: >>C:\fah2\FahCore_*.exe := PAIR3::CPU1+PAIR2::CPU1+PAIR1::CPU1+PAIR0::CPU1 [assign=1,resource=cpuuse,pol
icy=pseudobalanced,priority=idle]<<
WARNING process_nested_cpus::operator (): no logical CPU set of type PAIR was detected on this machine.
WARNING process_nested_cpus::operator (): no logical CPU set of type PAIR was detected on this machine.
WARNING process_nested_cpus::operator (): no logical CPU set of type PAIR was detected on this machine.
WARNING process_nested_cpus::operator (): no logical CPU set of type PAIR was detected on this machine.
WARNING CWinAFCDlg::parse_input_file: Ignoring profile line: C:\fah2\FahCore_*.exe := PAIR3::CPU1+PAIR2::CPU1+PAIR1::CPU1+PAIR0::CPU1 [assign=1,resource=cpuuse,
policy=pseudobalanced,priority=idle]
DEBUG CWinAFCDlg::parse_input_file: Found profile line: >>C:\Users\me\AppData\Roaming\Folding@home-gpu\FahCore_*.exe := SKIP [priority=belownormal]<<
DEBUG process_nested_attributes::operator (): Priority attribute set to belownormal.
AppName: >>C:\Users\me\AppData\Roaming\Folding@home-gpu\FahCore_*.exe<<
Skip=1, assign=0, policy=0, resource=1, poolSize=0, proc_mask=00000000, threads=0, force=0, CPUs:
01/12/2008 18:55:16: C:\Users\me\AppData\Roaming\Folding@home-gpu\FahCore_11.exe with PID 5884, set priority to Below Normal (TestMode is ENABLED; priority was
not modified)
DEBUG CWinAFCDlg::TraverseProcessList: Process >>C:\Users\me\AppData\Roaming\Folding@home-gpu\FahCore_11.exe<< matches a profile!
AppName: >>C:\Users\me\AppData\Roaming\Folding@home-gpu\FahCore_*.exe<<
Skip=1, assign=0, policy=0, resource=1, poolSize=0, proc_mask=00000000, threads=0, force=0, CPUs:

It doesnt seem to have found the paired ones this time

Ill try with the 2nd setup now

[Pootank]

Code: Select all

C:\fah1\FahCore_*.exe := PAIR0+PAIR1 [assign=1,resource=cpuuse,policy=pseudobalanced,priority=idle]
C:\fah2\FahCore_*.exe := PAIR2+PAIR3 [assign=1,resource=cpuuse,policy=pseudobalanced,priority=idle]
C:\Users\me\AppData\Roaming\Folding@home-gpu\FahCore_*.exe := SKIP [priority=belownormal]


Code: Select all

WinAFC 0.9.1, (c) 2008 <baobab13@users.sourceforge.net>

DEBUG Args::parseCommandLine: using debug level 3
DEBUG CWinAFCApp::InitInstance: Log file is "C:\Users\me\Desktop\WinAFC-0.9.1\affinitylog.txt"
WARNING elevatePrivileges: Cannot elevate privileges to handle affinity for processes running under other user accounts.
DEBUG CWinAFCApp::InitInstance: Input file is "C:\Users\me\Desktop\WinAFC-0.9.1\affinityinput.txt"
DEBUG CWinAFCDlg::CWinAFCDlg: Input file: "C:\Users\me\Desktop\WinAFC-0.9.1\affinityinput.txt", Path: "C:\Users\me\Desktop\WinAFC-0.9.1\"
DEBUG CWinAFCDlg::initializeSystemConfiguration: System has 8 CPUs, activeProcessorMask=0x000000FF
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU0 has affinity mask 0x00000001
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU1 has affinity mask 0x00000002
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU2 has affinity mask 0x00000004
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU3 has affinity mask 0x00000008
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU4 has affinity mask 0x00000010
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU5 has affinity mask 0x00000020
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU6 has affinity mask 0x00000040
DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU7 has affinity mask 0x00000080
DEBUG first_thread_bdwth: Precision timer has 14318180 ticks/second
Auto detecting CPU configuration. Please wait.
|==================================================| 100.0%
DEBUG maketest_bdwth: CPU detection finished? 1
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU3 is 6886.74 MB/s         (0.523753)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU5 is 6932.84 MB/s         (0.527259)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU2 is 7150.13 MB/s         (0.543784)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU7 is 7343.27 MB/s         (0.558473)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU6 is 7483.01 MB/s         (0.569101)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU6 is 7601.37 MB/s         (0.578102)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU5 and CPU6 is 7748.25 MB/s         (0.589272)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU5 is 8280.65 MB/s         (0.629763)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU4 is 8424.56 MB/s         (0.640708)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU6 and CPU7 is 8431.88 MB/s         (0.641265)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU4 and CPU5 is 8607.75 MB/s         (0.654639)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU4 and CPU6 is 8675.29 MB/s         (0.659776)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU6 is 8767.25 MB/s         (0.66677)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU5 and CPU7 is 8827.55 MB/s         (0.671356)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU3 is 9098.24 MB/s         (0.691942)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU4 is 9190.99 MB/s         (0.698996)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU4 and CPU7 is 9217.91 MB/s         (0.701044)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU6 is 9295.9 MB/s  (0.706975)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU7 is 9536.27 MB/s         (0.725256)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU4 is 9747.64 MB/s         (0.741331)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU5 is 9807.98 MB/s         (0.74592)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU7 is 9886.03 MB/s         (0.751856)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU1 is 9996.94 MB/s         (0.760291)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU2 is 10121.3 MB/s         (0.76975)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU3 is 10133 MB/s   (0.770638)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU7 is 11560.8 MB/s         (0.879225)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU5 is 12492.8 MB/s         (0.950103)
DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU4 is 13148.8 MB/s         (1)
DEBUG build_cpu_sets: Min number visited=0, Max number visited=5
ERROR build_cpu_sets: found sets of different sizes, 1-6.
ERROR build_cpu_sets: Please send the above error message and the numbers listed below to: baobab13@users.sourceforge.net
2, 3, 6886.74
2, 5, 6932.84
1, 2, 7150.13
2, 7, 7343.27
3, 6, 7483.01
2, 6, 7601.37
5, 6, 7748.25
3, 5, 8280.65
2, 4, 8424.56
6, 7, 8431.88
4, 5, 8607.75
4, 6, 8675.29
1, 6, 8767.25
5, 7, 8827.55
1, 3, 9098.24
3, 4, 9190.99
4, 7, 9217.91
0, 6, 9295.9
1, 7, 9536.27
1, 4, 9747.64
0, 5, 9807.98
0, 7, 9886.03
0, 1, 9996.94
0, 2, 10121.3
0, 3, 10133
3, 7, 11560.8
1, 5, 12492.8
0, 4, 13148.8
ERROR build_cpu_sets: Unable to determine the logical CPU sets. Use only physical CPU IDs.
DEBUG CWinAFCDlg::OnFinalizeCPUDetection: finalizing CPU detection. pendingTraverseProcesses = 0, afcState=1, oldState=0, savedOldState=0
DEBUG CWinAFCDlg::parse_input_file: TestMode set to 1.
DEBUG CWinAFCDlg::parse_input_file: ConsoleLogLevel set to 4.
DEBUG CWinAFCDlg::parse_input_file: TimeInterval set to 3600.
DEBUG CWinAFCDlg::parse_input_file: Found profile line: >>C:\fah1\FahCore_*.exe := PAIR0+PAIR1 [assign=1,resource=cpuuse,policy=pseudobalanced,priority=idle]<<
WARNING process_nested_cpus::operator (): no logical CPU set of type PAIR was detected on this machine.
WARNING process_nested_cpus::operator (): no logical CPU set of type PAIR was detected on this machine.
WARNING CWinAFCDlg::parse_input_file: Ignoring profile line: C:\fah1\FahCore_*.exe := PAIR0+PAIR1 [assign=1,resource=cpuuse,policy=pseudobalanced,priority=idle]

DEBUG CWinAFCDlg::parse_input_file: Found profile line: >>C:\fah2\FahCore_*.exe := PAIR2+PAIR3 [assign=1,resource=cpuuse,policy=pseudobalanced,priority=idle]<<
WARNING process_nested_cpus::operator (): no logical CPU set of type PAIR was detected on this machine.
WARNING process_nested_cpus::operator (): no logical CPU set of type PAIR was detected on this machine.
WARNING CWinAFCDlg::parse_input_file: Ignoring profile line: C:\fah2\FahCore_*.exe := PAIR2+PAIR3 [assign=1,resource=cpuuse,policy=pseudobalanced,priority=idle]

DEBUG CWinAFCDlg::parse_input_file: Found profile line: >>C:\Users\me\AppData\Roaming\Folding@home-gpu\FahCore_*.exe := SKIP [priority=belownormal]<<
DEBUG process_nested_attributes::operator (): Priority attribute set to belownormal.
AppName: >>C:\Users\me\AppData\Roaming\Folding@home-gpu\FahCore_*.exe<<
Skip=1, assign=0, policy=0, resource=1, poolSize=0, proc_mask=00000000, threads=0, force=0, CPUs:
01/12/2008 19:21:29: C:\Users\me\AppData\Roaming\Folding@home-gpu\FahCore_11.exe with PID 5884, set priority to Below Normal (TestMode is ENABLED; priority was
not modified)
DEBUG CWinAFCDlg::TraverseProcessList: Process >>C:\Users\me\AppData\Roaming\Folding@home-gpu\FahCore_11.exe<< matches a profile!
AppName: >>C:\Users\me\AppData\Roaming\Folding@home-gpu\FahCore_*.exe<<
Skip=1, assign=0, policy=0, resource=1, poolSize=0, proc_mask=00000000, threads=0, force=0, CPUs:


So I seem to have broken it :S

I assume nothing i could be running would alter the result>? I shall restart my machine and see if it can find the pairs again

[MtM]

Subscribed, we have some members on XS popping up with I7's and so far I'm insure on what to say if anything, as I wasn't even sure afc supported nehalems ( when it popped up the thread about afc didn't have any posts concerning nehalems ).

[Pootank]

Ok after restart I'm getting this

Code: Select all

01/12/2008 19:52:05: WARNING elevatePrivileges: Cannot elevate privileges to handle affinity for processes running under other user accounts.
01/12/2008 19:52:05: DEBUG CWinAFCApp::InitInstance: Input file is "C:\Users\me\Desktop\WinAFC-0.9.1\affinityinput.txt"
01/12/2008 19:52:05: DEBUG CWinAFCDlg::CWinAFCDlg: Input file: "C:\Users\me\Desktop\WinAFC-0.9.1\affinityinput.txt", Path: "C:\Users\me\Desktop\WinAFC-0.9.1\"
01/12/2008 19:52:05: DEBUG CWinAFCDlg::initializeSystemConfiguration: System has 8 CPUs, activeProcessorMask=0x000000FF
01/12/2008 19:52:05: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU0 has affinity mask 0x00000001
01/12/2008 19:52:05: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU1 has affinity mask 0x00000002
01/12/2008 19:52:05: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU2 has affinity mask 0x00000004
01/12/2008 19:52:05: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU3 has affinity mask 0x00000008
01/12/2008 19:52:05: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU4 has affinity mask 0x00000010
01/12/2008 19:52:05: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU5 has affinity mask 0x00000020
01/12/2008 19:52:05: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU6 has affinity mask 0x00000040
01/12/2008 19:52:05: DEBUG CWinAFCDlg::initializeSystemConfiguration: CPU7 has affinity mask 0x00000080
01/12/2008 19:52:07: DEBUG first_thread_bdwth: Precision timer has 14318180 ticks/second
Auto detecting CPU configuration. Please wait.
01/12/2008 19:52:07: DEBUG maketest_bdwth: CPU detection finished? 1
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU5 is 8616.59 MB/s    (0.618395)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU7 is 8635.85 MB/s    (0.619778)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU2 is 8795.09 MB/s    (0.631206)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU3 is 8864.58 MB/s    (0.636193)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU4 is 8969.25 MB/s    (0.643706)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU5 is 9426.48 MB/s    (0.67652)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU4 is 9580.38 MB/s    (0.687565)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU3 is 9596.39 MB/s    (0.688714)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU6 is 9653.7 MB/s    (0.692827)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU6 and CPU7 is 9700.09 MB/s    (0.696156)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU5 is 9726.07 MB/s    (0.698021)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU4 and CPU5 is 9746.18 MB/s    (0.699464)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU5 and CPU6 is 9759.8 MB/s    (0.700442)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU5 and CPU7 is 9769.21 MB/s    (0.701117)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU4 and CPU7 is 9793.35 MB/s    (0.702849)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU4 and CPU6 is 9801.27 MB/s    (0.703418)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU7 is 9802.27 MB/s    (0.70349)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU6 is 9846.89 MB/s    (0.706692)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU4 is 10113.7 MB/s    (0.725838)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU7 is 10130.9 MB/s    (0.727077)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU6 is 10190.1 MB/s    (0.73132)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU3 is 10192.1 MB/s    (0.731464)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU2 is 10211.8 MB/s    (0.732882)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU1 is 10232.6 MB/s    (0.734371)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU2 and CPU6 is 12141.8 MB/s    (0.871395)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU3 and CPU7 is 12956.7 MB/s    (0.92988)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU0 and CPU4 is 13761.8 MB/s    (0.987655)
01/12/2008 19:52:07: DEBUG CWinAFCDlg::detectCPUConfiguration: Communication bandwidth between CPU1 and CPU5 is 13933.8 MB/s    (1)
01/12/2008 19:52:07: DEBUG build_cpu_sets: Min number visited=1, Max number visited=1
01/12/2008 19:52:09: Found 8 CPUs:
    CPU0 has mask 00000001
    CPU1 has mask 00000002
    CPU2 has mask 00000004
    CPU3 has mask 00000008
    CPU4 has mask 00000010
    CPU5 has mask 00000020
    CPU6 has mask 00000040
    CPU7 has mask 00000080
01/12/2008 19:52:09: Found 4 PAIRs:
    PAIR0<CPU0,CPU4> has mask 00000011
    PAIR1<CPU1,CPU5> has mask 00000022
    PAIR2<CPU2,CPU6> has mask 00000044
    PAIR3<CPU3,CPU7> has mask 00000088
01/12/2008 19:52:09: DEBUG CWinAFCDlg::OnFinalizeCPUDetection: finalizing CPU detection. pendingTraverseProcesses = 0, afcState=1, oldState=0, savedOldState=0
01/12/2008 19:52:09: DEBUG CWinAFCDlg::parse_input_file: TestMode set to 1.
01/12/2008 19:52:09: DEBUG CWinAFCDlg::parse_input_file: ConsoleLogLevel set to 4.
01/12/2008 19:52:09: DEBUG CWinAFCDlg::parse_input_file: TimeInterval set to 3600.
01/12/2008 19:52:09: DEBUG CWinAFCDlg::parse_input_file: Found profile line: >>C:\fah1\FahCore_*.exe := PAIR0::CPU0+PAIR1::CPU0+PAIR2::CPU0+PAIR3::CPU0 [assign=1,resource=cpuuse,policy=pseudobalanced,priority=idle]<<
01/12/2008 19:52:09: DEBUG process_nested_attributes::operator (): Priority attribute set to idle.
01/12/2008 19:52:09: DEBUG CWinAFCDlg::parse_input_file: Found profile line: >>C:\fah2\FahCore_*.exe := PAIR3::CPU1+PAIR2::CPU1+PAIR1::CPU1+PAIR0::CPU1 [assign=1,resource=cpuuse,policy=pseudobalanced,priority=idle]<<
01/12/2008 19:52:09: DEBUG process_nested_attributes::operator (): Priority attribute set to idle.
01/12/2008 19:52:09: DEBUG CWinAFCDlg::parse_input_file: Found profile line: >>C:\Users\me\AppData\Roaming\Folding@home-gpu\FahCore_*.exe := SKIP [priority=belownormal]<<
01/12/2008 19:52:09: DEBUG process_nested_attributes::operator (): Priority attribute set to belownormal.
01/12/2008 19:52:09: DEBUG CWinAFCDlg::TraverseProcessList: Process >>C:\fah1\FahCore_a1.exe<< matches a profile!
01/12/2008 19:52:09: DEBUG CWinAFCDlg::TraverseProcessList: Process >>C:\fah1\FahCore_a1.exe<< matches a profile!
01/12/2008 19:52:09: DEBUG CWinAFCDlg::TraverseProcessList: Process >>C:\fah1\FahCore_a1.exe<< matches a profile!
01/12/2008 19:52:09: DEBUG CWinAFCDlg::TraverseProcessList: Process >>C:\fah1\FahCore_a1.exe<< matches a profile!
01/12/2008 19:52:09: DEBUG CWinAFCDlg::TraverseProcessList: Process >>C:\fah2\FahCore_a1.exe<< matches a profile!
01/12/2008 19:52:09: DEBUG CWinAFCDlg::TraverseProcessList: Process >>C:\fah2\FahCore_a1.exe<< matches a profile!
01/12/2008 19:52:09: DEBUG CWinAFCDlg::TraverseProcessList: Process >>C:\fah2\FahCore_a1.exe<< matches a profile!
01/12/2008 19:52:09: DEBUG CWinAFCDlg::TraverseProcessList: Process >>C:\fah2\FahCore_a1.exe<< matches a profile!
01/12/2008 19:52:09: C:\Users\me\AppData\Roaming\Folding@home-gpu\FahCore_11.exe with PID 3364, set priority to Below Normal (TestMode is ENABLED; priority was not modified)
01/12/2008 19:52:09: DEBUG CWinAFCDlg::TraverseProcessList: Process >>C:\Users\me\AppData\Roaming\Folding@home-gpu\FahCore_11.exe<< matches a profile!
01/12/2008 19:52:09: DEBUG ApplicationProfile::setAffinity:
Found 4 new C:\fah1\FahCore_*.exe processes.
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: processing process 2588 of process 2588 (C:\fah1\FahCore_a1.exe)
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: thread 2588 of process 2588 (C:\fah1\FahCore_a1.exe) uses  67.9% of CPU
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: processing process 4116 of process 4116 (C:\fah1\FahCore_a1.exe)
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: thread 4116 of process 4116 (C:\fah1\FahCore_a1.exe) uses  65.8% of CPU
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: processing process 5664 of process 5664 (C:\fah1\FahCore_a1.exe)
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: thread 5664 of process 5664 (C:\fah1\FahCore_a1.exe) uses  66.2% of CPU
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: processing process 6032 of process 6032 (C:\fah1\FahCore_a1.exe)
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: thread 6032 of process 6032 (C:\fah1\FahCore_a1.exe) uses  63.1% of CPU
01/12/2008 19:52:09: C:\fah1\FahCore_a1.exe with PID 2588 using 67.9% of CPU time, set affinity to CPU0 (TestMode is ENABLED; affinity was not modified).
01/12/2008 19:52:09: C:\fah1\FahCore_a1.exe with PID 2588, set priority to Idle (TestMode is ENABLED; priority was not modified)
01/12/2008 19:52:09: C:\fah1\FahCore_a1.exe with PID 5664 using 66.2% of CPU time, set affinity to CPU1 (TestMode is ENABLED; affinity was not modified).
01/12/2008 19:52:09: C:\fah1\FahCore_a1.exe with PID 5664, set priority to Idle (TestMode is ENABLED; priority was not modified)
01/12/2008 19:52:09: C:\fah1\FahCore_a1.exe with PID 4116 using 65.8% of CPU time, set affinity to CPU2 (TestMode is ENABLED; affinity was not modified).
01/12/2008 19:52:09: C:\fah1\FahCore_a1.exe with PID 4116, set priority to Idle (TestMode is ENABLED; priority was not modified)
01/12/2008 19:52:09: C:\fah1\FahCore_a1.exe with PID 6032 using 63.1% of CPU time, set affinity to CPU3 (TestMode is ENABLED; affinity was not modified).
01/12/2008 19:52:09: C:\fah1\FahCore_a1.exe with PID 6032, set priority to Idle (TestMode is ENABLED; priority was not modified)
01/12/2008 19:52:09: DEBUG ApplicationProfile::setAffinity:
Found 4 new C:\fah2\FahCore_*.exe processes.
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: processing process 1304 of process 1304 (C:\fah2\FahCore_a1.exe)
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: thread 1304 of process 1304 (C:\fah2\FahCore_a1.exe) uses  69.4% of CPU
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: processing process 4092 of process 4092 (C:\fah2\FahCore_a1.exe)
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: thread 4092 of process 4092 (C:\fah2\FahCore_a1.exe) uses  67.1% of CPU
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: processing process 4184 of process 4184 (C:\fah2\FahCore_a1.exe)
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: thread 4184 of process 4184 (C:\fah2\FahCore_a1.exe) uses  68.6% of CPU
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: processing process 5904 of process 5904 (C:\fah2\FahCore_a1.exe)
01/12/2008 19:52:09: DEBUG computeProcessCpuTime: thread 5904 of process 5904 (C:\fah2\FahCore_a1.exe) uses  53.4% of CPU
01/12/2008 19:52:09: C:\fah2\FahCore_a1.exe with PID 1304 using 69.4% of CPU time, set affinity to CPU7 (TestMode is ENABLED; affinity was not modified).
01/12/2008 19:52:09: C:\fah2\FahCore_a1.exe with PID 1304, set priority to Idle (TestMode is ENABLED; priority was not modified)
01/12/2008 19:52:09: C:\fah2\FahCore_a1.exe with PID 4184 using 68.6% of CPU time, set affinity to CPU6 (TestMode is ENABLED; affinity was not modified).
01/12/2008 19:52:09: C:\fah2\FahCore_a1.exe with PID 4184, set priority to Idle (TestMode is ENABLED; priority was not modified)
01/12/2008 19:52:09: C:\fah2\FahCore_a1.exe with PID 4092 using 67.1% of CPU time, set affinity to CPU5 (TestMode is ENABLED; affinity was not modified).
01/12/2008 19:52:09: C:\fah2\FahCore_a1.exe with PID 4092, set priority to Idle (TestMode is ENABLED; priority was not modified)
01/12/2008 19:52:09: C:\fah2\FahCore_a1.exe with PID 5904 using 53.4% of CPU time, set affinity to CPU4 (TestMode is ENABLED; affinity was not modified).
01/12/2008 19:52:09: C:\fah2\FahCore_a1.exe with PID 5904, set priority to Idle (TestMode is ENABLED; priority was not modified)

So it seems to be working as expected now, I shall turn the testmode off and see what effect it has on my ppd.

I'd had a look on xs before I posted here and was suprised to see no threads about it there.

[shatteredsilicon]

I thought there were some 8 core WUs around for Skulltrails. "-smp 8" ought to get them fetched, IIRC.

[toTOW]

-smp X flag is only supported with A2 core at the moment ... and this core is only available for Linux or OSX.

We're stuck with A1 core under windows which is hardcoded for 4 cores unless some progress is made porting A2 core

[baobab]

The CPU detection is architecture oblivious and it is just an heuristic. It looks like it is not very robust for Nehalem. It may be that some cores go into a low power mode or something else. I will have to look at those numbers later and think a bit more about it. Perhaps I will have to adjust it somewhat.
But I am thinking that some heuristic should work to find the pairs of virtual cores that share a physical core because the two virtual cores use the same L1/L2 cache hierarchy, while different physical cores communicate through the shared L3 cache. Unfortunately I do not have a Nehalem machine to test different strategies.

Also, it might be the case that for Nehalem the cores are numbered consistently by Windows. That is, CPUs 0 to 3 are the physical cores (or the first virtual cores as you may want to think about it) and CPUs 4-7 are the secondary virtual cores that match those primary cores so that virtual cores (0,4), (1,5), (2,6), (3,7) are the pairs corresponding to the four physical cores.

If they are always numbered like this, the CPU detection does not even have to be executed, just use the correct CPU indices.

Pootank, I do not think the system restart had anything to do with the heuristic succeeding. It is just that the measured bandwidths are not very consistent and the heuristic may fail in some cases. if that happens, you could try running the heuristic again from the context menu (right click the systray icon). You'll have the option to rerun it until it completes successfully. Also, you do not need to run it with the debug mode on. I asked you that just to see those numbers the first time.

Also, the detection heuristic is independent of the application profiles used. You can change the application profiles at run-time, without restarting the program.

Looking through your log file, I see that the TestMode is enabled. You need to comment out the line that sets TestMode to 1 in affinityinput.txt, or just set the value to 0. In TestMode affinities are not modified.
Also, you have the TimeInterval set to 1 hour. I think that is a bit long, but it is your call. The program is not using many CPU cycles.

Thanks for your reports and you can ask me if you have other questions about settings, etc.

[Pootank]

After letting it run as per the 2nd set up ive got the following ppds
gpu 2386
fah1 965
fah2 1405
total 4457

Although these figures are changing up and down after every frame.

I cpu detect works when i dont have the gpu client working but doesnt when I do so thats easy to get around.

I've changed the timeinterval back to 30 sec now so ill see if this helps, I changed it not quite realising how the program was suppose to work :S.