Seti@Home optimized science apps and information

Optimized Seti@Home apps => Windows => Topic started by: Simon on 03 Oct 2006, 04:34:39 pm

Title: New version of the CPUID program
Post by: Simon on 03 Oct 2006, 04:34:39 pm

Hi folks,

Ben Herndon has compiled a new version with the updates from the first test run.

Below is a list of the models already tested and correctly identified - if your CPU is in that list, you don't need to try this new version (though you can, of course). Please only post your results if your CPU was incorrectly identified or is missing in the list below.

Thanks for your help!


Please download the attached file (http://lunatics.at/index.php?module=Downloads;sa=download;id=27;mirror=1) and extract and run "CPUID 10-04-A.exe" by double-clicking it.

  It will display a 'command' window with information about your CPU, and then pause, asking you to press enter for memory speed.
  Press enter once to perform some benchmarks.
  We hope the information is completely accurate.
  If it is not accurate or...
  If the model of CPU you have is not currently listed on the table of tested CPUs
  - and please, only then -
  please copy this window of information and attach it to a reply to this thread (http://lunatics.at/index.php?action=post;topic=89.0).
  How to copy the command window content depends on your version of windows.
 

• Win XP- Look for the "C:\" tiny icon in the upper left of the window.
• click on this tiny icon to bring up a menu
• choose the edit -> mark   option (continue)
• Win 2000 - Win SE
• With the left mouse button - drag select the entire text area of the window
• click the right mouse button or press a key
• Now this text is in your clipboard
• next launch [start menu] -> programs -> Accessories -> Notepad (or use Windows Key + R and just type "notepad" and hit enter)
• paste the text into the notepad window.
• Save the file, giving it a descriptive name, such as "cpuid <my processor model type - including core type>.txt"
• attach it to a post in this thread (http://lunatics.at/index.php?action=post;topic=89.0)

Current Computed List

Intel
ID Brand Computed Model

5-4-3 02 P1m Pentium-MMX (P55C)
6-1-9 00 P1m Pentium-Pro ()
6-5-2 02 P2 Pentium-II (Deschutes)*
6-6-5 00 P2 Celeron (Mendocino)
6-8-6 02 P3 Pentium-III (Coppermine)
6-8-A 01 P3 Celeron (Coppermine)
6-9-5 16 PM Pentium M (Banias)
6-B-1 2004 P3 Pentium-III (Tualatin)
6-D-8 16 PM Pentium M (Dothan) {error:  PM Core Duo (Yonah) when L2=4Meg}
6-E-8 220 PM Core Duo (Yonah)*
6-F-6 220 C2D Core 2 Duo (Conroe)
6-F-6 2220 C2Dx Xeon 51xx (Woodcrest)
F-2-5 09 P4 Xeon (Gallatin) {P4 Xeon (Prestonia B) L1=512K, L2=2Meg)
F-2-5 200B P4 Xeon (Prestonia B)
F-2-7 09 P4 Pentium-IV (Northwood)
F-2-7 0E P4 Pentium-IV (Northwood)* {error: PD Pentium D (Presler)}
F-2-7 200B Xeon DP (Prestonia) {error: PDx Xeon D 70 (Paxville) }
F-2-9 09 P4 Pentium-IV (Northwood)*
F-2-9 0B P4 Xeon DP (Prestonia)
F-3-4 00 P4 Celeron (Prescott)
F-3-4 00 P4 Pentium-IV (Prescott)
F-3-4 00 P4 Pentium-IV (Prescott)* {error: Xeon P4 (Nocona) when L2=2M}
F-4-1 00 P4 Pentium-IV (Prescott)*
F-4-1 00 P4 Celeron (Prescott) {error: Pentium-IV (Prescott) when L2 = 256K}
F-4-1 2000 P4 Xeon MP (Cranford)*
F-4-3 00 P4 Pentium-IV (Prescott)* {error: PD Pentium D (Smithfield) when L2=2M}
F-4-3 2000 P4 Xeon DP (Irwindale)
F-4-4 00 PD Pentium D (Smithfield)
F-4-7 210 PD Pentium D (Smithfield)*
F-6-4 210 PD Pentium D (Presler)*

AMD
5-7-0 00 K6 (Little Foot)
5-8-C 00 K6! K6-II 3D (Chompers)
5-9-1 00 K6! K6-III K6-3 (Sharptooth)
6-3-1 00 K7 Duron (Spitfire)
6-4-4 00 K7 Athlon (Thunderbird)*
6-6-2 00 K7x Athlon XP (Palomino)
6-8-0 00 K7x Athlon XP (Thoroughbred)
6-A-0 00  K7x Athlon XP (Barton)*
F-3-2 B04 K8 Opteron 1 DC (xxx) [Denmark]*
F-7-1 304 K8 Opteron 1 (Sledgehammer) [Venus]*
F-1-2 C05 K8 Opteron DC 2 (xxx)[Italy] *
F-4-A 106 K8 Athlon 64 (Clawhammer)
F-C-0 108 K8 Athlon 64 (Newcastle)*
F-7-A 10D K8 Athlon 64 (Newcastle)*
F-F-0 10D K8 Athlon 64 (Winchester)
F-7-1 10F K8 Athlon 64 (San Diego)*
F-4-2 20F K8 Athlon 64 (Newark)*
F-B-1 150 K8 Athlon 64 X2 (Manchester)
F-3-2 15A K8 Athlon 64 X2 (Toledo)*
F-C-2 8C6 K8 Sempron (Albany)
F-2C-2 982 K8 Sempron (Manila)*
F-24-2 2CA K8 Turion 64 (Lancaster)*

Title: Re: New version of the CPUID program
Post by: hiamps on 03 Oct 2006, 04:39:13 pm

Where is the list of already tested?

Title: Re: New version of the CPUID program
Post by: Simon on 03 Oct 2006, 04:40:04 pm

It was just an edit away :)

Fixed, thanks for noticing.

Regards,
Simon.

Title: Re: New version of the CPUID program
Post by: Pepo on 03 Oct 2006, 06:18:05 pm

I thought OK, my Banias is already listed, but I noticed that the 10-03-A output differs :-\ from the output of the 10-02-A substantially (L1/code and Extd:family-model  Brand description) and I've nowhere read about having such inconsistence solved...

Code: [Select]

  cpus cores hyperT =threads L1    L2    L3   code
7c7
<   1 *   1  *   1   =  1     32K 1024K    0K   16K
>   1 *   1  *   1   =  1     32K 1024K    0K   32K
9c9
<  family-model  Brand  Extd:family-model  Brand [9-30b]
>  family-model  Brand  Extd:family-model  Brand [10-03a]
      6-9-5       16            00-0       000

So here is my output.

Peter

[attachment deleted by admin]

Title: Re: New version of the CPUID program
Post by: Roy on 05 Oct 2006, 11:50:04 am

Ran the new version. Both my family-model and computed name are listed in your table.
But they don't match with each other - so here's my result.

 CPUID
 -tag: 'GenuineIntel'
 -fullname: 'Mobile Intel(R) Pentium(R) 4 - M CPU 1.80GHz'
Computed name: 'Intel P4 Pentium-IV (Northwood)'
 -notes:
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   1  *   1   =  1      8K  512K    0K   12K

 family-model  Brand  Extd:family-model  Brand [10-03a]
    F-2-7       0E            00-0       000

 features: mmx sse sse2

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   00000002 756E6547 6C65746E 49656E69
 0..1   00000F27 0001080E 00004400 BFEBF9FF
 0..2   665B5101 00000000 00000000 007B7040
 8..0   80000004 00000000 00000000 00000000
 8..1   00000000 00000000 00000000 00000000
 8..2   69626F4D 4920656C 6C65746E 20295228
 8..3   746E6550 286D7569 34202952 4D202D20
 8..4   55504320 382E3120 7A484730 007A4800

 < Press [enter] for memory speed >
 Speed=1794 Mhz
    Test  :    L1   L2   RAM
    8 ints:   6880  4652 1227 - Reads
    8 ints:   3400  3366  629 - Writes
     4 sse:  13617  8613 1368 - Reads
     4 sse:   7752  7777  568 - Writes
 pre-cache:     33    33 1365 - Reads
not-cached:     33    33 1354 - Writes

 < Press [enter] key to close >

Title: Re: New version of the CPUID program
Post by: BenHer on 05 Oct 2006, 12:27:11 pm

Thanks Roy,

They match as deep as the ID program is written to go.  Not significant enough to the optimized crucher whether a system is Mobile or not.  Only CPU speed, core type, and RAM speed.  Different RAM/CPU speed meaning is obvious, core type tells certain manufacturer (Intel/AMD) improvements in timings.

Title: Re: New version of the CPUID program
Post by: Roy on 05 Oct 2006, 09:56:06 pm

Wasn't sure which factors might be important to you, so....

Are you planning to have a version of this for Linux?

Roy

Title: Re: New version of the CPUID program
Post by: htrae on 05 Oct 2006, 11:24:15 pm

Hey ChickenMan,

I saw you were looking for an AMD Opteron 1 Venus tested with your newer CPUID test so here it is. I also replied in the new thread you started on the Seti forums.

Regards.

[attachment deleted by admin]

Title: Intel P4 (Northwood) 2.26Ghz
Post by: Evil_Rich on 06 Oct 2006, 08:11:36 am

 CPUID
 -tag: 'GenuineIntel'
 -fullname: 'Intel(R) Pentium(R) 4 CPU 2.26GHz'
Computed name: 'Intel P4 Pentium-IV (Northwood)'
 -notes:
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   1  *   1   =  1      8K  512K    0K   12K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    F-2-4       09            00-0       000

 features: mmx sse sse2

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   00000002 756E6547 6C65746E 49656E69
 0..1   00000F24 00010809 00000000 3FEBFBFF
 0..2   665B5001 00000000 00000000 007B7040
 8..0   80000004 00000000 00000000 00000000
 8..1   00000000 00000000 00000000 00000000
 8..2   65746E49 2952286C 6E655020 6D756974
 8..3   20295228 50432034 2E322055 48473632
 8..4   2034007A 20555043 36322E32 007A4847

 < Press [enter] for memory speed >
 Speed=2267 Mhz
    Test  :    L1   L2   RAM
    8 ints:   8368  5999 1739 - Reads
    8 ints:   4408  4458 1014 - Writes
     8 sse:  17143 16958 1928 - Reads
     8 sse:   9759  9866 1013 - Writes
 pre-cache:     33    33 1919 - Reads
not-cached:     33    33  440 - Writes

 < Press [enter] key to close >


[attachment deleted by admin]

Title: Re: New version of the CPUID program
Post by: Pepo on 06 Oct 2006, 08:17:24 am

Quote from: Pepo on 03 Oct 2006, 06:18:05 pm

...I noticed that the 10-03-A output differs :-\ from the output of the 10-02-A substantially (L1/code and Extd:family-model  Brand description)...

Code: [Select]

<  family-model  Brand  Extd:family-model  Brand [9-30b]
>  family-model  Brand  Extd:family-model  Brand [10-03a]


Now after seeing the result of 10-04-A I've finally understood what the text "Brand [9-30b]" mean - the CPUID test program version :-) (which was already named 10-02-A). I've previously thought that it would be fine to have the version number somewhere in the output  ;D

For my Banias the only difference is the usage of 8sse instead of 4sse and slightly higher memory speed scores - probably not worth attaching the output.

Peter

Title: Re: New version of the CPUID program
Post by: Pepo on 06 Oct 2006, 08:41:12 am

I've found one Prescott, as requested on main Seti page...
[edit]But forgot to attach the output  :-[

[attachment deleted by admin]

Title: Re: New version of the CPUID program
Post by: BadWolf63 on 09 Oct 2006, 03:37:50 am

Here is My CPU... it wasn't in the list...

Best Regards
BadWolf63

CPUID
 -tag: 'GenuineIntel'
 -fullname: 'Intel(R) Pentium(R) 4 CPU 1.80GHz'
Computed name: 'Intel P4 Pentium-IV (Willamette)'
 -notes:
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   1  *   1   =  1      8K  256K    0K   12K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    F-1-2       08            00-0       000

 features: mmx sse sse2

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   00000002 756E6547 6C65746E 49656E69
 0..1   00000F12 00010808 00000000 3FEBFBFF
 0..2   665B5001 00000000 00000000 007A7040
 8..0   80000004 00000000 00000000 00000000
 8..1   00000000 00000000 00000000 00000000
 8..2   65746E49 2952286C 6E655020 6D756974
 8..3   20295228 50432034 2E312055 48473038
 8..4   2034007A 20555043 30382E31 007A4847

 < Press [enter] for memory speed >
 Speed=1816 Mhz
    Test  :    L1   L2   RAM
    8 ints:   6747  4805 1415 - Reads
    8 ints:   3535  3570 1002 - Writes
     8 sse:  13672 13732 1825 - Reads
     8 sse:   7798  7899 1007 - Writes
 pre-cache:     33    33 1811 - Reads
not-cached:     33    33  321 - Writes

Title: Re: New version of the CPUID program
Post by: Arnulf on 10 Oct 2006, 05:27:28 pm

A few differences from the previous version, so I have attached the new 10-04-A result from my Opteron machine.

[attachment deleted by admin]

Title: Re: New version of the CPUID program
Post by: svenie25 on 13 Oct 2006, 04:38:58 pm

Hi. Here is my CPU:

  CPUID
 -tag: 'AuthenticAMD'
 -fullname: 'AMD Athlon(tm) 64 Processor 3500+'
Computed name: 'AMD K8 Athlon 64 (Venice)'
 -notes: XX00+
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   1  *   1   =  1     64K  512K    0K   64K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    F-F-0       42F            00-2       10D

 features: mmx 3Dnow 3Dnow+ sse sse2 sse3 em64t

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   00000001 68747541 444D4163 69746E65
 0..1   00020FF0 00000800 00000001 078BFBFF
 8..0   80000018 68747541 444D4163 69746E65
 8..1   00020FF0 0000010D 00000001 E3D3FBFF
 8..2   20444D41 6C687441 74286E6F 3620296D
 8..3   72502034 7365636F 20726F73 30303533
 8..4   0000002B 00000000 00000000 00000000
 8..5   FF08FF08 FF20FF20 40020140 40020140
 8..6   00000000 42004200 02008140 00000000

 < Press [enter] for memory speed >
 Speed=2420 Mhz
    Test  :    L1   L2   RAM
    8 ints:  13312  6808 3239 - Reads
    8 ints:  16138  6260 2195 - Writes
     8 sse:  18304  9800 3318 - Reads
     8 sse:  18364  9102 2301 - Writes
 pre-cache:     33    33 3711 - Reads
not-cached:     33    33  552 - Writes

 < Press [enter] key to close >

Could you please tell me, for what this informations are needed?  ;)

Title: Re: New version of the CPUID program
Post by: Hal2710 on 15 Oct 2006, 03:45:15 am

Hi All,

Here is my cpu code.

I am looking for a seti@home client to use on this NEW baby.

She's running sweat @ 2.804 Ghz 1.4v

See you soon, Hal2710


 CPUID
 -tag: 'AuthenticAMD'
 -fullname: 'Dual Core AMD Opteron(tm) Processor 170'
Computed name: 'AMD K8 Opteron DC 1 (xxx)'
 -notes:
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   2  *   1   =  2     64K 1024K    0K   64K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    F-3-2       2C23            00-2       B05

 features: mmx 3Dnow 3Dnow+ sse sse2 sse3 em64t

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   00000001 68747541 444D4163 69746E65
 0..1   00020F32 01020800 00000001 178BFBFF
 8..0   80000018 68747541 444D4163 69746E65
 8..1   00020F32 00000B05 00000003 E3D3FBFF
 8..2   6C617544 726F4320 4D412065 704F2044
 8..3   6F726574 6D74286E 72502029 7365636F
 8..4   20726F73 00303731 00000000 00000000
 8..5   FF08FF08 FF20FF20 40020140 40020140
 8..6   00000000 42004200 04008140 00000000

 < Press [enter] for memory speed >
 Speed=2799 Mhz
    Test  :    L1   L2   RAM
    8 ints:  15397  5224 3546 - Reads
    8 ints:  18666  3706 2390 - Writes
     8 sse:  21211  5860 3566 - Reads
     8 sse:  21242  4206 2656 - Writes
 pre-cache:     33    33 3928 - Reads
not-cached:     33    33 2964 - Writes

 < Press [enter] key to close >

Title: Re: New version of the CPUID program
Post by: David Ball on 15 Oct 2006, 02:04:43 pm

Hi,

Murphy is alive and well. 
I tried the cpuid program and it got an error.

Code: [Select]

C:>"cpuid 10-04-a"
 CPUID
 -tag: 'GenuineIntel'
 -fullname: 'Intel(R) Celeron(R) CPU 2.30GHz'
Computed name: 'Intel P4c Celeron (Northwood)'
 -notes: (mobile)
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   1  *   1   =  1      8K  128K    0K   12K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    F-2-7       0A            00-0       000

 features: mmx sse sse2

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   00000002 756E6547 6C65746E 49656E69
 0..1   00000F27 0001080A 00004400 BFEBFBFF
 0..2   665B5101 00000000 00000000 003B7040
 8..0   80000004 00000000 00000000 00000000
 8..1   00000000 00000000 00000000 00000000
 8..2   65746E49 2952286C 6C654320 6E6F7265
 8..3   20295228 20555043 30332E32 007A4847
 8..4   20295228 20555043 30332E32 007A4847

After that, it offered to send an error report to Microsoft. The machines is an eMachines D2386 desktop machine. I don't know why it says the celeron 2.3 in it is a mobile cpu. It's running WinXP Pro SP2 with all patches including this months Patches and has 1.5 GB of ram. I'm also trying the "Windows Live Onecare" antivirus on that machine. It's behind a firewall/NAT and the mail server for that domain also runs a virus check on email and uses a couple of the more conservative blacklists that don't block huge IP blocks.

I saved the output from CPU-Z
Here (http://www.booksnbytes.com/book_covers/cpuz_eMachines_D2386.gif).

-- David Ball

Title: Re: New version of the CPUID program
Post by: S Cubed Sound on 18 Oct 2006, 02:48:03 am

the brand was different.  Not sure if that matters for what you are doing.

[attachment deleted by admin]

Title: Re: New version of the CPUID program
Post by: KWSN - Sir Brian on 25 Oct 2006, 04:13:12 pm

not sure about this, everythinlooks fine except the intel code name is different, I was expecting conroe

Quote

CPUID
 -tag: 'GenuineIntel'
 -fullname: 'Intel(R) Core(TM)2 CPU          6300  @ 1.86GHz'
Computed name: 'Intel C2D Core 2 Duo (Allendale)'
 -notes: E6xxx
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   2  *   1   =  2     32K 2048K    0K   32K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    6-F-6       220            00-0       000

 features: mmx sse sse2 sse3 sse4 em64t

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   0000000A 756E6547 6C65746E 49656E69
 0..1   000006F6 01020800 0000E3BD BFEBFBFF
 0..2   05B0B101 005657F0 00000000 2CB4307D
 0..3   00000000 00000000 00000000 00000000
 0..4   04000121 01C0003F 0000003F 00000001
 8..0   80000008 00000000 00000000 00000000
 8..1   00000000 00000000 00000001 20100000
 8..2   65746E49 2952286C 726F4320 4D542865
 8..3   43203229 20205550 20202020 20202020
 8..4   30303336 20402020 36382E31 007A4847
 8..5   00000000 00000000 00000000 00000000
 8..6   00000000 00000000 08006040 00000000

 < Press [enter] for memory speed >
 Speed=1863 Mhz
    Test  :    L1   L2   RAM
    8 ints:   7006  5923 4430 - Reads
    8 ints:   6692  5536 2539 - Writes
     8 sse:  18739 12907 5222 - Reads
     8 sse:  24490 11255 2625 - Writes
 pre-cache:     33    33 5223 - Reads
not-cached:     33    33  541 - Writes

 < Press [enter] key to close >

Title: Re: New version of the CPUID program
Post by: BenHer on 25 Oct 2006, 04:37:43 pm

Sir Brian,

Allendale's have a 2 Meg L2 Cache per core (2 x 1024)
Conroes have a 4 Meg L2 Cache  (2 x 2048)

So you do have a Conroe, but my identification table (corrected just now) listed the doubled cache values, but compared them to  the single cache sizes.

Thanks for the assist.

Title: Re: New version of the CPUID program
Post by: S Cubed Sound on 26 Oct 2006, 07:34:18 am

Assistance from S Cubed Sound

CPUID
 -tag: 'AuthenticAMD'
 -fullname: 'AMD Athlon(tm) 64 X2 Dual Core Processor 4200+'
Computed name: 'AMD K8 Athlon 64 X2 (Manchester)'
 -notes: XX00+
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   2  *   1   =  2     64K  512K    0K   64K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    F-B-1       52B            00-2       154

 features: mmx 3Dnow 3Dnow+ sse sse2 sse3 em64t

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   00000001 68747541 444D4163 69746E65
 0..1   00020FB1 01020800 00000001 178BFBFF
 8..0   80000018 68747541 444D4163 69746E65
 8..1   00020FB1 00000154 00000003 E3D3FBFF
 8..2   20444D41 6C687441 74286E6F 3620296D
 8..3   32582034 61754420 6F43206C 50206572
 8..4   65636F72 726F7373 30323420 00002B30
 8..5   FF08FF08 FF20FF20 40020140 40020140
 8..6   00000000 42004200 02008140 00000000

 < Press [enter] for memory speed >
 Speed=2216 Mhz
    Test  :    L1   L2   RAM
    8 ints:  12190  6234 2751 - Reads
    8 ints:  14778  5732 2051 - Writes
     8 sse:  16761  8974 2783 - Reads
     8 sse:  16817  8336 2080 - Writes
 pre-cache:     33    33 3087 - Reads
not-cached:     33    33 1407 - Writes

 < Press [enter] key to close >

Title: Re: New version of the CPUID program
Post by: troythewdblue on 02 Nov 2006, 06:17:44 pm

hi i tried the cpuid and it says i have a 64kb l1cache when i have 2 so it is actuly 128kb amd 3800+ venice single core

Title: Re: New version of the CPUID program
Post by: BenHer on 02 Nov 2006, 09:12:19 pm

Quote from: troythewdblue on 02 Nov 2006, 06:17:44 pm

hi i tried the cpuid and it says i have a 64kb l1cache when i have 2 so it is actuly 128kb amd 3800+ venice single core

Hi Troy,

1.  It only reports what exists on a single core...so that is how it is intended to run.
2.  If you have a 3800+ "venice single core", then I'm confused what you mean when you say "...when i have 2..."

Please copy the text inside the window when the program is waiting for  "< Press [enter] key to close >".  Instructions on how to do so are on the first post in this thread.

Title: Re: New version of the CPUID program
Post by: Simon on 03 Nov 2006, 06:41:15 am

64 K L1-I, 64K L1-D, 512 K L2 in that single core Athlon64, to my knowledge, all in all 640K.

troythewdblue, it says 64K twice instead of 128K once because 64K of that 128K are meant to cache Instructions and 64K are meant to cache Data.

HTH,
Simon.

Title: Re: New version of the CPUID program
Post by: FloridaBear on 03 Nov 2006, 09:34:58 am

Quote from: BenHer on 25 Oct 2006, 04:37:43 pm

Sir Brian,

Allendale's have a 2 Meg L2 Cache per core (2 x 1024)
Conroes have a 4 Meg L2 Cache  (2 x 2048)

So you do have a Conroe, but my identification table (corrected just now) listed the doubled cache values, but compared them to  the single cache sizes.

Thanks for the assist.

To my knowledge, the 6300 and 6400 ARE Allendales and have the 2MB cache versus the Conroe's 4 MB. My E6400 shows up as an Allendale with 2 MB L2 cache in CPU-Z, which I believe is correct. Also, for both of these chips, I believe the cache is shared, not per core.

So basically, I think your CPUID was entirely correct for that E6300. Just my 2¢

Title: Re: New version of the CPUID program
Post by: talaktalan on 05 Nov 2006, 03:19:13 pm

Hi,

do you also need the output file if only the brand number is different, but not the family-model?

Please find attached a F-4-A 404 AMD K8 Athlon 64 (Clawhammer).

Btw, I always wondered if it would be possible to create a client optimized for 3dnow? Is there any compiler for this?

Regards

Alex

[attachment deleted by admin]

Title: Re: New version of the CPUID program
Post by: gleckert on 09 Nov 2006, 03:53:13 am

 CPUID
 -tag: 'GenuineIntel'
 -fullname: 'Intel(R) Core(TM)2 CPU         E6600  @ 2.40GHz'
Computed name: 'Intel C2D Core 2 Duo (Conroe)'
 -notes: E6xxx
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   2  *   1   =  2     32K 4096K    0K   32K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    6-F-5       220            00-0       000

 features: mmx sse sse2 sse3 sse4 em64t

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   0000000A 756E6547 6C65746E 49656E69
 0..1   000006F5 01020800 0000E3BD BFEBFBFF
 0..2   05B0B101 005657F0 00000000 2CB43049
 0..3   00000000 00000000 00000000 00000000
 0..4   04000121 01C0003F 0000003F 00000001
 8..0   80000008 00000000 00000000 00000000
 8..1   00000000 00000000 00000001 20100000
 8..2   65746E49 2952286C 726F4320 4D542865
 8..3   43203229 20205550 20202020 45202020
 8..4   30303636 20402020 30342E32 007A4847
 8..5   00000000 00000000 00000000 00000000
 8..6   00000000 00000000 10008040 00000000

 < Press [enter] for memory speed >
 Speed=2400 Mhz
    Test  :    L1   L2   RAM
    8 ints:   9030  7770 5023 - Reads
    8 ints:   8625  7307 3242 - Writes
     8 sse:  24100 16714 5763 - Reads
     8 sse:  31292 15059 3319 - Writes
 pre-cache:     33    33 5752 - Reads
not-cached:     33    33  534 - Writes

 < Press [enter] key to close >

Title: Re: New version of the CPUID program
Post by: FluffyChicken on 10 Nov 2006, 07:34:45 am

This is a Pentium-M Dothan but not the one on the list ;-)




[attachment deleted by admin]

Title: Re: New version of the CPUID program
Post by: swmpthng on 11 Nov 2006, 08:55:01 am

where it says F-6-4 210 PD Pentium D (Presler)* does that mean all presler cpu's?

ps- you should animate the kitty to move side to side , i was just burning time during regisrtation doing that and it was funny, you can make it an opera singer or it could be in a horror movie screaming, or you could have just stepped on it's tail! we americans are easily amused!  :D

Title: Re: New version of the CPUID program
Post by: BenHer on 11 Nov 2006, 12:14:42 pm

Quote from: swmpthng on 11 Nov 2006, 08:55:01 am

where it says F-6-4 210 PD Pentium D (Presler)* does that mean all presler cpu's?

All presler's so far have been F-6-4.  Intel is fixated on Core 2, so not likely to make more sandwich core CPUs (like Pentium D).

Title: Re: New version of the CPUID program
Post by: Tank on 13 Nov 2006, 01:23:01 pm

Here's one more missing in the list

 CPUID
 -tag: 'AuthenticAMD'
 -fullname: 'AMD Athlon(tm) 64 X2 Dual Core Processor 5000+'
Computed name: 'AMD K8 Athlon 64 FX (Orleans)'
 -notes: -ZZ
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   2  *   1   =  2     64K  512K    0K   64K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    F-B-2       244B            00-4       919

 features: mmx 3Dnow 3Dnow+ sse sse2 sse3 em64t

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   00000001 68747541 444D4163 69746E65
 0..1   00040FB2 01020800 00002001 178BFBFF
 8..0   80000018 68747541 444D4163 69746E65
 8..1   00040FB2 00000919 0000001F EBD3FBFF
 8..2   20444D41 6C687441 74286E6F 3620296D
 8..3   32582034 61754420 6F43206C 50206572
 8..4   65636F72 726F7373 30303520 00002B30
 8..5   FF08FF08 FF20FF20 40020140 40020140
 8..6   00000000 42004200 02008140 00000000

 < Press [enter] for memory speed >

Title: Re: New version of the CPUID program
Post by: KWSN - Sir Brian on 14 Nov 2006, 02:24:17 pm

Quote from: FloridaBear on 03 Nov 2006, 09:34:58 am

Quote from: BenHer on 25 Oct 2006, 04:37:43 pm

Sir Brian,

Allendale's have a 2 Meg L2 Cache per core (2 x 1024)
Conroes have a 4 Meg L2 Cache  (2 x 2048)

So you do have a Conroe, but my identification table (corrected just now) listed the doubled cache values, but compared them to  the single cache sizes.

Thanks for the assist.

To my knowledge, the 6300 and 6400 ARE Allendales and have the 2MB cache versus the Conroe's 4 MB. My E6400 shows up as an Allendale with 2 MB L2 cache in CPU-Z, which I believe is correct. Also, for both of these chips, I believe the cache is shared, not per core.

So basically, I think your CPUID was entirely correct for that E6300. Just my 2¢

yep you're right.....
my bad....

 :-[ :-[ :-[

Ni!

Title: Re: New version of the CPUID program
Post by: FluffyChicken on 15 Nov 2006, 05:24:53 am

Quick question,  I posted mine above since it had a different family model.
I have 2 more compters with different family models agian.

Is it just hits you are after are more information ?

Title: Re: New version of the CPUID program
Post by: Kiwit on 01 Dec 2006, 07:02:46 pm

This is for my XP-M 3000 Athlon Laptop in a Asus A2500D

CPUID
 -tag: 'AuthenticAMD'
 -fullname: 'AMD Athlon(tm) Processor'
Computed name: 'AMD K7x Athlon XP (Barton)'
 -notes:
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   1  *   1   =  1     64K  512K    0K   64K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    6-A-0       00            00-0       000

 features: mmx 3Dnow 3Dnow+ sse

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   00000001 68747541 444D4163 69746E65
 0..1   000006A0 00000000 00000000 0383F9FF
 8..0   80000008 68747541 444D4163 69746E65
 8..1   000007A0 00000000 00000000 C1CBF9FF
 8..2   20444D41 6C687441 74286E6F 5020296D
 8..3   65636F72 726F7373 00000000 00000000
 8..4   00000000 00000000 00000000 00000000
 8..5   0408FF08 FF20FF10 40020140 40020140
 8..6   00000000 41004100 02008140 00000000

 < Press [enter] for memory speed >
 Speed=2199 Mhz
    Test  :    L1   L2   RAM
    8 ints:  11773  5530  522 - Reads
    8 ints:  13192  5248  554 - Writes
     8 sse:  16562  6460  884 - Reads
     8 sse:  16594  6446  724 - Writes
 pre-cache:     33    33  964 - Reads
not-cached:     33    33  675 - Writes

 < Press [enter] key to close >

Title: Re: New version of the CPUID program
Post by: citroja on 26 Dec 2006, 03:08:02 pm

I didn't see this computer on the list but someone else may have posted for it...
The only thing that I noticed a difference on was the extended family values.

I have attached the .txt files for comparison.

**************************************

here is the 10-04-A results:

**************************************

 CPUID
 -tag: 'GenuineIntel'
 -fullname: 'Intel(R) Pentium(R) 4 CPU 1.80GHz'
Computed name: 'Intel P4 Pentium-IV (Willamette)'
 -notes:
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   1  *   1   =  1      8K  256K    0K   12K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    F-1-2       08            00-0       000

 features: mmx sse sse2

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   00000002 756E6547 6C65746E 49656E69
 0..1   00000F12 00010808 00000000 3FEBFBFF
 0..2   665B5001 00000000 00000000 007A7040
 8..0   80000004 00000000 00000000 00000000
 8..1   00000000 00000000 00000000 00000000
 8..2   65746E49 2952286C 6E655020 6D756974
 8..3   20295228 50432034 2E312055 48473038
 8..4   2034007A 20555043 30382E31 007A4847

 < Press [enter] for memory speed >
 Speed=1793 Mhz
    Test  :    L1   L2   RAM
    8 ints:   6662  4745  933 - Reads
    8 ints:   3482  3525  512 - Writes
     8 sse:  13499 12113  988 - Reads
     8 sse:   7661  7794  513 - Writes
 pre-cache:     33    33  986 - Reads
not-cached:     33    33  336 - Writes

-citroja

[attachment deleted by admin]

[attachment deleted by admin]

Title: Re: New version of the CPUID program
Post by: Hal2710 on 04 Jan 2007, 06:38:13 am

Hi,

Heres my baby :-



CPUID
 -tag: 'AuthenticAMD'
 -fullname: 'Dual Core AMD Opteron(tm) Processor 170'
Computed name: 'AMD K8 Opteron DC 1 (xxx)'
 -notes:
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   2  *   1   =  2     64K 1024K    0K   64K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    F-3-2       2C23            00-2       B05

 features: mmx 3Dnow 3Dnow+ sse sse2 sse3 em64t

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   00000001 68747541 444D4163 69746E65
 0..1   00020F32 01020800 00000001 178BFBFF
 8..0   80000018 68747541 444D4163 69746E65
 8..1   00020F32 00000B05 00000003 E3D3FBFF
 8..2   6C617544 726F4320 4D412065 704F2044
 8..3   6F726574 6D74286E 72502029 7365636F
 8..4   20726F73 00303731 00000000 00000000
 8..5   FF08FF08 FF20FF20 40020140 40020140
 8..6   00000000 42004200 04008140 00000000

 < Press [enter] for memory speed >
 Speed=2899 Mhz
    Test  :    L1   L2   RAM
    8 ints:  15952  8138 3782 - Reads
    8 ints:  19333  7486 2711 - Writes
     8 sse:  21927 11654 3765 - Reads
     8 sse:  22004 10897 2921 - Writes
 pre-cache:     33    33 4070 - Reads
not-cached:     33    33 3063 - Writes

 < Press [enter] key to close >



Hal2710

Title: Re: New version of the CPUID program
Post by: talaktalan on 10 Jan 2007, 02:23:49 pm

Here is the output for an oldie, it's neither Intel nor AMD  ;D :

 CPUID
 -tag: 'CyrixInstead'
 -fullname: ''
Computed name: 'Cyrix 6x86MX 6x86MX (MII)'
 -notes:
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   1  *   1   =  1     16K    0K    0K   16K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    6-0-0       43            00-0       000

 features: mmx

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   00000001 69727943 64616574 736E4978
 0..1   00000600 69727943 00000000 0080A135
 8..0   [CPU doesn't have xtd ID]

 < Press [enter] for memory speed >
 Speed= 249 Mhz
    Test  :    L1   L2   RAM
    8 ints:    964    33  741 - Reads
    8 ints:    128    33  322 - Writes
     8 mmx:   1513    33 1517 - Reads
     8 mmx:    289    33  346 - Writes
 pre-cache:     33    33   33 - Reads
not-cached:     33    33   33 - Writes

 < Press [enter] key to close >

There is, however, an error in the data: This cpu has a 64 KB unified L1 cache, not 16KB L1 data + 16KB L1 code as listed in the output. The other information should be correct. Maybe the doubled 6x86MX in the 'Computed name' output is a bit odd. This cpu was called Cyrix MII PR333 (by Cyrix) and IBM 6x86MX PR333 (by IBM), I have the IBM version here, but afaik only the package is different.

Regards

Alex

[attachment deleted by admin]

[attachment deleted by admin]

Title: Re: New version of the CPUID program
Post by: troythewdblue on 10 Jan 2007, 04:18:56 pm

 CPUID
 -tag: 'AuthenticAMD'
 -fullname: 'AMD Athlon(tm) 64 Processor 3800+'
Computed name: 'AMD K8 Athlon 64 (Venice)'
 -notes: XX00+
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   1  *   1   =  1     64K  512K    0K   64K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    F-F-2       42F            00-2       110

 features: mmx 3Dnow 3Dnow+ sse sse2 sse3 em64t

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   00000001 68747541 444D4163 69746E65
 0..1   00020FF2 00000800 00000001 078BFBFF
 8..0   80000018 68747541 444D4163 69746E65
 8..1   00020FF2 00000110 00000001 E3D3FBFF
 8..2   20444D41 6C687441 74286E6F 3620296D
 8..3   72502034 7365636F 20726F73 30303833
 8..4   0000002B 00000000 00000000 00000000
 8..5   FF08FF08 FF20FF20 40020140 40020140
 8..6   00000000 42004200 02008140 00000000
it is actuly 2 64 kb l1 cache for total of 128 only shows 64

Title: Re: New version of the CPUID program
Post by: dkenny on 04 Feb 2007, 12:55:53 am

My copy of the cpuid run for my system is attached.

[attachment deleted by admin]

Title: Re: New version of the CPUID program
Post by: v4wheelman on 09 Feb 2007, 02:38:39 pm

Good Day !

A BIG thank you to those who take the time to write, publish and support optomized applications. Great work !!!

I am running an optomized Seti App ... details below. My computer is not in the list and I am submitting it here. Is there any chance there will be an SSE3 app for AMD processors?


 CPUID
 -tag: 'AuthenticAMD'
 -fullname: 'AMD Turion(tm) 64 X2 Mobile Technology TL-52'
Computed name: 'AMD K8 Opteron 8 (Sledgehammer)'
 -notes: YY HE
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   2  *   1   =  2     64K  512K    0K   64K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    F-8-2       1648            00-4       597

 features: mmx 3Dnow 3Dnow+ sse sse2 sse3 em64t

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   00000001 68747541 444D4163 69746E65
 0..1   00040F82 01020800 00002001 178BFBFF
 8..0   80000018 68747541 444D4163 69746E65
 8..1   00040F82 00000597 0000001F EBD3FBFF
 8..2   20444D41 69727554 74286E6F 3620296D
 8..3   32582034 626F4D20 20656C69 68636554
 8..4   6F6C6F6E 54207967 32352D4C 00000000
 8..5   FF08FF08 FF20FF20 40020140 40020140
 8..6   00000000 42004200 02008140 00000000

 < Press [enter] for memory speed >
 Speed=1607 Mhz
    Test  :    L1   L2   RAM
    8 ints:   8839  4496 2266 - Reads
    8 ints:  10753  4153 1746 - Writes
     8 sse:  12134  6372 2521 - Reads
     8 sse:  12184  5999 1969 - Writes
 pre-cache:     33    33 2651 - Reads
not-cached:     33    33 1933 - Writes

 < Press [enter] key to close >





<core_client_version>5.8.8</core_client_version>
<![CDATA[
<stderr_txt>
Optimized SETI@Home Enhanced application

Optimizers: Ben Herndon, Josef Segur, Alex Kan, Simon Zadra
  Version: Windows SSE2 32-bit 'Ni!' based on seti V5.15  'Chicken Good!'
      Rev: (R-2.0|QxW|FFT:IPP_SSE2|Ben-Joe)
    CPUID: 'AMD K8 Opteron 8 (Sledgehammer)'
     cpus: 1 cores: 2 threads: 1   cache: L1=64K  L2=512K L3=0K
 features: mmx 3Dnow 3Dnow+ sse sse2 sse3 
    speed: 1607 MHz  -- read megs/sec: L1=8805, L2=4442, RAM=2399

Title: Re: New version of the CPUID program
Post by: BenHer on 23 Feb 2007, 04:47:02 pm

Ok...

Now I've gotten a version of CPUID to work with an external identification table.
(note it still has an internal table, and will use that if external table file is not in same directory)

This version has no checksum or md5 checking to see if the external file is valid...which is good for this test in case some end user can update the table for a new CPU and send me a working update.

The version which will be put into seti cruncher (and maybe boinc eventually) will have some protections against being modified.

Now I have to look through all the postings here and elsewhere about mis-identification and update the table wtih those newer CPUs.  Once I have that I will probably post something.

Title: Re: New version of the CPUID program
Post by: Ronon Dex on 25 Feb 2007, 09:37:32 pm

I let run the CPUID 10-04-A program when the CPU (QX6700) was in idle (energy savings mode -> @1.6GHz)


 CPUID
 -tag: 'GenuineIntel'
 -fullname: 'Intel(R) Core(TM)2 Quad CPU           @ 2.66GHz'
Computed name: 'Intel C2Dx Xeon 51xx (Woodcrest)'
 -notes: 51xx
 cpus cores hyperT =threads L1    L2    L3   code
  1 *   4  *   1   =  4     32K 4096K    0K   32K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    6-F-7       420            00-0       000

 features: mmx sse sse2 sse3 sse4 em64t

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   0000000A 756E6547 6C65746E 49656E69
 0..1   000006F7 03040800 0000E3BD BFEBFBFF
 0..2   05B0B101 005657F0 00000000 2CB43049
 0..3   00000000 00000000 00000000 00000000
 0..4   0C000121 01C0003F 0000003F 00000001
 8..0   80000008 00000000 00000000 00000000
 8..1   00000000 00000000 00000001 20100000
 8..2   65746E49 2952286C 726F4320 4D542865
 8..3   51203229 20646175 20555043 20202020
 8..4   20202020 20402020 36362E32 007A4847
 8..5   00000000 00000000 00000000 00000000
 8..6   00000000 00000000 10008040 00000000

 < Press [enter] for memory speed >
 Speed=2666 Mhz
    Test  :    L1   L2   RAM
    8 ints:   6028  5176 5734 - Reads
    8 ints:   5753  4880 3126 - Writes
     8 sse:  16046 11060 6373 - Reads
     8 sse:  21409 10055 3189 - Writes
 pre-cache:     33    33 6316 - Reads
not-cached:     33    33  446 - Writes

 < Press [enter] key to close >

Title: Re: New version of the CPUID program
Post by: KWSN - Sir Brian on 28 Mar 2007, 05:13:07 pm

here's an update for A clvertown E5320...    currently being wrongly detected as a woodcrest

Whilst in idle
 CPUID
 -tag: 'GenuineIntel'
 -fullname: 'Intel(R) Xeon(R) CPU           E5320  @ 1.86GHz'
Computed name: 'Intel C2Dx Xeon 51xx (Woodcrest)'
 -notes: 51xx

 cpus cores hyperT =threads L1    L2    L3   code
  2 *   4  *   1   =  8     32K 4096K    0K   32K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    6-F-7       2420            00-0       000

 features: mmx sse sse2 sse3 sse4 em64t

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   0000000A 756E6547 6C65746E 49656E69
 0..1   000006F7 07040800 0004E3BD BFEBFBFF
 0..2   05B0B101 005657F0 00000000 2CB43049
 0..3   00000000 00000000 00000000 00000000
 0..4   0C000121 01C0003F 0000003F 00000001
 8..0   80000008 00000000 00000000 00000000
 8..1   00000000 00000000 00000001 20100000
 8..2   65746E49 2952286C 6F655820 2952286E
 8..3   55504320 20202020 20202020 45202020
 8..4   30323335 20402020 36382E31 007A4847
 8..5   00000000 00000000 00000000 00000000
 8..6   00000000 00000000 10008040 00000000

 < Press [enter] for memory speed >
 Speed=1866 Mhz
    Test  :    L1   L2   RAM
    8 ints:   7018  5853 3211 - Reads
    8 ints:   6703  5693 2122 - Writes
     8 sse:  18724 12958 3317 - Reads
     8 sse:  24987 11730 2152 - Writes
 pre-cache:     33    33 3289 - Reads
not-cached:     33    33  127 - Writes

 < Press [enter] key to close >


whilst running at 100% with 8 optimised clients....

 CPUID
 -tag: 'GenuineIntel'
 -fullname: 'Intel(R) Xeon(R) CPU           E5320  @ 1.86GHz'
Computed name: 'Intel C2Dx Xeon 51xx (Woodcrest)'
 -notes: 51xx

 cpus cores hyperT =threads L1    L2    L3   code
  2 *   4  *   1   =  8     32K 4096K    0K   32K

 family-model  Brand  Extd:family-model  Brand [10-04a]
    6-F-7       2420            00-0       000

 features: mmx sse sse2 sse3 sse4 em64t

 cpuid   -eax-    -ebx-    -ecx-    -edx-
 0..0   0000000A 756E6547 6C65746E 49656E69
 0..1   000006F7 07040800 0004E3BD BFEBFBFF
 0..2   05B0B101 005657F0 00000000 2CB43049
 0..3   00000000 00000000 00000000 00000000
 0..4   0C000121 01C0003F 0000003F 00000001
 8..0   80000008 00000000 00000000 00000000
 8..1   00000000 00000000 00000001 20100000
 8..2   65746E49 2952286C 6F655820 2952286E
 8..3   55504320 20202020 20202020 45202020
 8..4   30323335 20402020 36382E31 007A4847
 8..5   00000000 00000000 00000000 00000000
 8..6   00000000 00000000 10008040 00000000

 < Press [enter] for memory speed >
 Speed=1866 Mhz
    Test  :    L1   L2   RAM
    8 ints:   7018  5676 1472 - Reads
    8 ints:   6498  5414  786 - Writes
     8 sse:  18811 12715 1488 - Reads
     8 sse:  24529 11166 1053 - Writes
 pre-cache:     33    33 1679 - Reads
not-cached:     33    33   38 - Writes

 < Press [enter] key to close >

Ni!