SMP Client, Folding dgn Multicore Processor

[roedzzz]

Tweet

wekekekekek

kowe arep njaluk opo le?
*sambil menengadah dan ngusap2 janggut* :laugh:

Kulo nyuwun toge kemawon nDoro Gandrik

[raido]

Tweet

Kulo nyuwun toge kemawon nDoro Gandrik

Wah, ini ni... Mulai deh...

[zungkono]

Tweet

karena susah dapat yg lain (selalu dapat p2665) tadi coba yg dari mac os tak pindah ke win xp ternyata bisa diproses ... jadinya sekarang proses p2605 di win xp.

[g4ndr1k]

Tweet

PPD sama atau berubah mas?

[zungkono]

Tweet

baru proses mas, install win xp dulu

[testament85]

Tweet

mau nanya ni...
kalo giling smp pake vista 64bit sebenernya bisa gak???
saya udah aplikasinya yg utk vista 64bit, setelah instal n download wu malah dapet project 2606???
apa itu termasuk smp???

[04:09:55] Folding@Home Gromacs Core
[04:09:55] Version 1.90 (March 8, 2006)
[04:09:55]
[04:09:55] Preparing to commence simulation
[04:09:55] - Looking at optimizations...
[04:09:55] - Files status OK
[04:09:59] - Expanded 3678978 -> 18878715 (decompressed 513.1 percent)
[04:09:59]
[04:09:59] Project: 2606 (Run 34, Clone 11, Gen 94)
[04:09:59]
[04:10:00] Assembly optimizations on if available.
[04:10:00] Entering M.D.
[04:10:21] (Starting from checkpoint)
[04:10:21] Protein: p2606_tet_1499
[04:10:21]
[04:10:21] Writing local files
[04:10:24] Extra SSE boost OK.
[04:10:24] Writing local files
[04:10:24] Completed 0 out of 125000 steps (0%)

[agungy]

Tweet

mau nanya ni...
kalo giling smp pake vista 64bit sebenernya bisa gak???
saya udah aplikasinya yg utk vista 64bit, setelah instal n download wu malah dapet project 2606???
apa itu termasuk smp???

Bukan, SMP itu bro....sepertinya anda kurang ngasih flag "-smp".

[zungkono]

Tweet

Bro disini saya ingin memberi sedikit info perbandingan folding smp dari windows (xp 32 bit), mac os x iDeneb dan linux (ubuntu 64bit)

misal :
day = 24 x 3600 = 86400 detik
m = menit
d = detik
p = point
w = waktu yg diperlukan untuk proses 100%
1% = m menit d detik

Perhitungan PPD :
w = (( m x 100 x 60 ) + ( d x 100 ))
Ppd = 86400 / w x p
= 86400 / ((m x 6000) + (dx100)) x p
Ppd = 86400p / (6000m+100d) <==== RUMUS




1. Windows XP 32bit (1 client @ 4core)
Q6600 @2.40 Ghz (266x9)
Super talent 4x1Gb @400 mhz Cl 5-5-5-15
p2605 = 1760 point

[04:01:00] Completed 150000 out of 500000 steps (30 percent)
[04:12:31] Writing local files
[04:12:32] Completed 155000 out of 500000 steps (31 percent)
[04:24:03] Writing local files
[04:24:04] Completed 160000 out of 500000 steps (32 percent)
[04:35:34] Writing local files
[04:35:35] Completed 165000 out of 500000 steps (33 percent)
[04:47:06] Writing local files
[04:47:06] Completed 170000 out of 500000 steps (34 percent)
[04:58:36] Writing local files
[04:58:36] Completed 175000 out of 500000 steps (35 percent)
[05:10:06] Writing local files
[05:10:06] Completed 180000 out of 500000 steps (36 percent)
[05:21:36] Writing local files
[05:21:37] Completed 185000 out of 500000 steps (37 percent)
[05:33:06] Writing local files
[05:33:06] Completed 190000 out of 500000 steps (38 percent)
[05:44:35] Writing local files
[05:44:36] Completed 195000 out of 500000 steps (39 percent)
[05:56:04] Writing local files
[05:56:05] Completed 200000 out of 500000 steps (40 percent)
[06:07:34] Writing local files
[06:07:34] Completed 205000 out of 500000 steps (41 percent)
[06:19:04] Writing local files
[06:19:04] Completed 210000 out of 500000 steps (42 percent)
[06:30:33] Writing local files
[06:30:33] Completed 215000 out of 500000 steps (43 percent)
[06:42:02] Writing local files
[06:42:02] Completed 220000 out of 500000 steps (44 percent)
[06:53:31] Writing local files
[06:53:32] Completed 225000 out of 500000 steps (45 percent)

waktu pengerjaan untuk 1% = 11 menit 32 detik
Ppd = 86400 x 1760 / ( ( 6000x11) + (100x32) )
= 152064000 / ( 66000 + 3200 )
= 152064000 / 69200
= 2197,45 point


2. Mac OS X iDeneb 10.5.4 (1 client @ 4core)
Q6600 @2.40 Ghz (266x9)
Super talent 4x1Gb @400 mhz Cl 5-5-5-15
p2605 = 1760 point

[07:19:33] Completed 135000 out of 500000 steps (27 percent)
[07:30:38] Writing local files
[07:30:38] Completed 140000 out of 500000 steps (28 percent)
[07:41:41] Writing local files
[07:41:41] Completed 145000 out of 500000 steps (29 percent)
[07:52:43] Writing local files
[07:52:44] Completed 150000 out of 500000 steps (30 percent)
[08:03:45] Writing local files
[08:03:46] Completed 155000 out of 500000 steps (31 percent)
[08:14:48] Writing local files
[08:14:48] Completed 160000 out of 500000 steps (32 percent)
[08:25:50] Writing local files
[08:25:50] Completed 165000 out of 500000 steps (33 percent)
[08:36:52] Writing local files
[08:36:52] Completed 170000 out of 500000 steps (34 percent)
[08:47:54] Writing local files
[08:47:54] Completed 175000 out of 500000 steps (35 percent)
[08:58:56] Writing local files
[08:58:56] Completed 180000 out of 500000 steps (36 percent)
[09:09:59] Writing local files
[09:09:59] Completed 185000 out of 500000 steps (37 percent)

waktu pengerjaan untuk 1% = 11 menit 2 detik
Ppd = 86400 x 1760 / ( ( 6000x11) + (100x2) )
= 152064000 / ( 66000 + 200 )
= 152064000 / 66200
= 2297,04 point

p2669 = 1920 point

[04:40:26] Preparing to commence simulation
[04:40:26] - Ensuring status. Please wait.
[04:40:26] Files status OK
[04:40:27] - Expanded 4838813 -> 23982265 (decompressed 495.6 percent)
[04:40:27] Called DecompressByteArray: compressed_data_size=4838813 data_size=23982265, decompressed_data_size=23982265 diff=0
[04:40:28] - Digital signature verified
[04:40:28]
[04:40:28] Project: 2669 (Run 9, Clone 139, Gen 7)
[04:40:28]
[04:40:28] Assembly optimizations on if available.
[04:40:28] Entering M.D.
[04:40:37] (Run 9, Clone 139, Gen 7)
[04:40:37]
[04:40:38] Entering M.D.
[04:40:44] Node 0 initialized
[04:48:17] Completed 2500 out of 250000 steps (1%)
[04:55:47] Completed 5000 out of 250000 steps (2%)
[05:03:14] Completed 7500 out of 250000 steps (3%)
[05:10:43] Completed 10000 out of 250000 steps (4%)
[05:18:11] Completed 12500 out of 250000 steps (5%)
[05:25:40] Completed 15000 out of 250000 steps (6%)
[05:33:08] Completed 17500 out of 250000 steps (7%)
[05:40:37] Completed 20000 out of 250000 steps (8%)
[05:48:05] Completed 22500 out of 250000 steps (9%)

kalau ini gilingnya pakai fahcore_a2.exe point 1920
waktu pengerjaan untuk 1% = 7 menit 30 detik
Ppd = 86400 x 1920 / ( ( 6000x7) + (100x30) )
= 165888000 / ( 42000 + 3000 )
= 165888000 / 45000
= 3686.40 point



3. Linux Ubuntu 64bit (2 client, masing2 dihandle 2 core) under vmware Win XP 32 bit
Q6600 @2.40 Ghz (266x9)
Super talent 4x1Gb @400 mhz Cl 5-5-5-15
karena satu virtual machine cuman bisa pakai 2 core, maka dibuat 2 virtual machine untuk memanfaatkan 4 core

A. mengerjakan p2665 dan p2669
p2665 = 1920 point
p2669 = 1920 point

waktu pengerjaan p2665 untuk 1% = 21 menit 40 detik
Ppd = 86400 x 1920 / ( ( 6000x21) + (100x40) )
= 165888000 / ( 126000 + 4000 )
= 165888000 / 130000
= 1276,06 point

waktu pengerjaan p2669 untuk 1% = 13 menit 20 detik
Ppd = 86400 x 1920 / ( ( 6000x13) + (100x20) )
= 165888000 / ( 78000 + 2000 )
= 165888000 / 80000
= 2073,06 point
Total Ppd = 1276,06 + 2073,06 = 3349,12 point

B. mengerjakan 2 x p2669
p2669 = 1920 point

waktu pengerjaan p2669 (vm1) untuk 1% = 13 menit 10 detik
Ppd = 86400 x 1920 / ( ( 6000x13) + (100x10) )
= 165888000 / ( 78000 + 1000 )
= 165888000 / 79000
= 2099,84 point

waktu pengerjaan p2669 (vm2) untuk 1% = 13 menit 15 detik
Ppd = 86400 x 1920 / ( ( 6000x13) + (100x15) )
= 165888000 / ( 78000 + 1500 )
= 165888000 / 79500
= 2086,64 point
Total Ppd = 2099,84 + 2086,64 = 4186,48 point

C. mengerjakan 2 x p2605
p2669 = 1760 point

waktu pengerjaan p2605 (vm1) untuk 1% = 17 menit 37 detik
Ppd = 86400 x 1760 / ( ( 6000x17) + (100x37) )
= 152064000 / ( 102000 + 3700 )
= 152064000 / 105700
= 1438,63 point

waktu pengerjaan p2605 (vm2) untuk 1% = 17 menit 23 detik
Ppd = 86400 x 1760 / ( ( 6000x17) + (100x23) )
= 152064000 / ( 102000 + 2300 )
= 152064000 / 104300
= 1457,94 point
Total Ppd = 1438,63 + 1457,94 = 2896,57 point


4. Linux Ubuntu 64bit (2 client, masing2 dihandle 2 core) under vmware Win XP 32 bit + GPU Nvidia Pov 8600GT Exo
Q6600 @2.40 Ghz (266x9)
Super talent 4x1Gb @400 mhz Cl 5-5-5-15
karena satu virtual machine cuman bisa pakai 2 core, maka dibuat 2 virtual machine untuk memanfaatkan 4 coremengerjakan 2 x p2605
p2669 = 1760 point

waktu pengerjaan p2605 (vm1) untuk 1% = 17 menit 10 detik
Ppd = 86400 x 1760 / ( ( 6000x17) + (100x10) )
= 152064000 / ( 102000 + 1000 )
= 152064000 / 103000
= 1476,34 point

waktu pengerjaan p2605 (vm2) untuk 1% = 17 menit 25 detik
Ppd = 86400 x 1760 / ( ( 6000x17) + (100x25) )
= 152064000 / ( 102000 + 2500 )
= 152064000 / 104500
= 1455,15 point
Sub Total Ppd = 1476,34 + 1455,15 = 2931,49 point

mengerjakan p5229 (GPU under windows xp)
p5229 = 480 point

waktu pengerjaan p5229 untuk 1% = 9 menit
Ppd = 86400 x 480 / ( ( 6000x9) + (100x0) )
= 41472000 / ( 54000 + 0 )
= 41472000 / 54000
= 768 point

Total Ppd = 2931,49 + 768 = 3699,49 point




Kesimpulan :
1. Windows XP 32bit (1 client @ 4core)
p2605 => Ppd = 2197,45 point

2. Mac OS X iDeneb 10.5.4 (1 client @ 4core)
bila mengerjakan p2605 => Ppd = 2297,04 point
dan
bila mengerjakan p2669 => Ppd = 3686,40 point

3. Linux Ubuntu 64bit (2 client, masing2 dihandle 2 core) under vmware Win XP 32 bit
A. p2665 + p2669
p2665 => Ppd = 1276,06 point
p2669 => Ppd = 2073,06 point
Total Ppd = 3349,12 point

B. 2 x p2669
p2669-1 => Ppd = 2099,84 point
p2669-2 => Ppd = 2086,64 point
Total Ppd = 4186,48 point

C. 2 x p2605
p2605-1 => Ppd = 1438,63 point
p2605-2 => Ppd = 1457,94 point
Total Ppd = 2896,57 point

4. Linux Ubuntu 64bit (2 client, masing2 dihandle 2 core) under vmware Win XP 32 bit + GPU Nvidia POV8600GT Exo
2 x p2605 + p5229
p2605-1 => Ppd = 1476,34 point
p2605-2 => Ppd = 1455,15 point
p5229 ===> Ppd = 768 point
Total Ppd = 3699,49 point

hasil point #4 agak aneh karena dengan menambah GPU yg juga membutuhkan resource CPU, ternyata tidak terjadi penurunan point pada SMP client tetapi yg terjadi SMP Client jadi lebih tinggi sedikit pointnya. Hal ini masih memerlukan pengujian lebih lanjut.





----------------------------------------------------------------------------------------------------------------------
data tambahan sebagai perbandingan AMD vs Intel untuk 2 core (menyusul):
Windows XP 32bit
X2 4050e @2.782 Ghz
Super talent 4x1Gb @463 mhz Cl 4-4-4-12
p2605 = 1760 point

[02:52:06] Completed 145000 out of 500000 steps (29 percent)
[03:07:07] Timered checkpoint triggered.
[03:17:50] Writing local files
[03:17:51] Completed 150000 out of 500000 steps (30 percent)
[03:32:52] Timered checkpoint triggered.
[03:43:38] Writing local files
[03:43:38] Completed 155000 out of 500000 steps (31 percent)
[03:58:39] Timered checkpoint triggered.
[04:09:26] Writing local files
[04:09:26] Completed 160000 out of 500000 steps (32 percent)
[04:24:27] Timered checkpoint triggered.
[04:35:15] Writing local files
[04:35:16] Completed 165000 out of 500000 steps (33 percent)

waktu pengerjaan untuk 1% = 25 menit 45 detik

[redwood]

Tweet

nice pak....
btw, kalau di mac procie nya make yang dual core aja deh, imho procie quadcore under hackintosh nggak gitu ngaruh, dibanding sama E2160 misalnya, malah bisa lebih cepet dari kuat kor...

mungkin perlu dibikin benchmark lagi antara vista dan linux native ya?...ini aku lagi make amd 4800 dapet wu 2665 minta ampun rasanya ...

[agungy]

Tweet

Thanks, bro Zungkono atas tabelnya......

berarti kemungkinan ada yang salah sama gilinganku ini....SMP on Vista, with dual core....