[molpro-user] Solid State Drives
Joseph Lane
jlane at waikato.ac.nz
Thu May 5 05:46:41 BST 2011
Dear Jacek
Thank you for sharing those results. Do you (or does anyone else on
the list) know what the major I/O bottleneck is for medium-large
CCSD(T) calculations with Molpro? Is it sequential read, sequential
write, seek times? We are in the process of ordering a new server and
are trying to evaluate whether to purchase a conventional HDD SAS RAID
0 array or some form of SSD setup.
Kind regards
Jo
On Thu, May 5, 2011 at 2:05 AM, Jacek Antoni Klos <jklos at umd.edu> wrote:
> Dear Jo
>
> Yes, I did some IO_test benchmarks and small_normal_ccsd bench using 512GB SSD that comes with Apple Mac Pro machines. Apple used Toshiba brand SSDs.
>
> Here are some excerpts from benchmark results on SSD as a scratch:
>
> IO_TEST: 200MB file
>
> Test I/O on file length 204.5Mb with transfer size 0.50Mb ( 409 segments) and 10000 seeks
>
> Test name Description Time/s Rate/(Mb)/s
> SW Sequential write, 1 process file 1 1.37 149.27
> SR Sequential read, 1 process file 1 0.12 1704.16
> SL Seek, 1 process 0.08 125000.12
>
>
> **********************************************************************************************************************************
> PROGRAMS * TOTAL IOTEST IOTEST
> CPU TIMES * 0.10 0.00 0.00
> REAL TIME * 3.77 SEC
> DISK USED * 214.76 MB
> **********************************************************************************************************************************
>
>
>
> IO_TEST: 10GB file
>
> Test I/O on file length 10239.5Mb with transfer size 0.50Mb ( 20479 segments) and 10000 seeks
>
> Test name Description Time/s Rate/(Mb)/s
> SW Sequential write, 1 process file 1 58.61 174.71
> SR Sequential read, 1 process file 1 2.59 3953.48
> SL Seek, 1 process 0.08 124999.75
>
>
> **********************************************************************************************************************************
> PROGRAMS * TOTAL IOTEST IOTEST
> CPU TIMES * 0.36 0.08 0.19
> REAL TIME * 124.01 SEC
> DISK USED * 10.74 GB
> **********************************************************************************************************************************
>
>
>
>
> IO_TEST: two simultaneous jobs using same SSD as scratch each 10GB file
>
> Test I/O on file length 10239.5Mb with transfer size 0.50Mb ( 20479 segments) and 10000 seeks
>
> Test name Description Time/s Rate/(Mb)/s
> SW Sequential write, 1 process file 1 118.73 86.24
> SR Sequential read, 1 process file 1 2.68 3820.71
> SL Seek, 1 process 0.08 125000.12
>
>
> **********************************************************************************************************************************
> PROGRAMS * TOTAL IOTEST IOTEST
> CPU TIMES * 0.48 0.07 0.32
> REAL TIME * 247.99 SEC
> DISK USED * 10.74 GB
> **********************************************************************************************************************************
>
>
>
> SMALL_NORMAL_CCSD benchmark on SSD :
>
> CPU and I/O time analysis:
> Routine CPU(%) MFLOP SYS WALL Call Routine CPU(%) MFLOP SYS WALL Call
> TOTAL: 514.5(99.9) 0.0 6.9 525.6 1 TRIPLES: 355.9(69.2) 6872.5 3.4 359.3 1
> KEXTA: 96.1(18.7) 0.0 0.7 99.4 8 CCVIJ3: 20.8( 4.0)10322.2 0.2 21.0 8
> YZMAT: 18.4( 3.6)10759.8 0.0 18.3 1152 CCKINT: 10.3( 2.0) 2145.2 0.5 11.8 1
> CCVIJ1: 6.0( 1.2) 1588.4 0.1 6.1 8 CC3EXT: 3.8( 0.7) 0.0 0.4 4.3 7
> CCKEXT: 2.1( 0.4) 4956.1 0.1 2.8 8 CCMP2: 0.4( 0.1) 1116.8 0.1 0.5 9
> CCDIIS: 0.4( 0.1) 0.0 0.5 1.4 17 CIOPTR: 0.3( 0.1) 9649.6 0.0 0.3 1
> CCVIJ2: 0.2( 0.0) 3186.9 0.0 0.2 8 CIORTH: 0.0( 0.0) 0.0 0.1 0.0 1
> TRANSFORM: 0.0( 0.0) 0.0 0.0 0.0 1
>
>
> PROGRAM ENERGY USER SYS TOTCPU ELAPSED USER(%) TOTCPU(%)
> AOINT 0.00000000 6.11 0.30 6.41 6.42 95.17 99.84
> AOSORT 0.00000000 3.45 1.28 4.73 10.75 32.09 44.00
> INT(TOT) 0.00000000 9.56 1.58 11.14 17.17 55.68 64.88
> HF-SCF -156.15930716 9.41 0.85 10.26 10.27 91.63 99.90
> TRANSFORM 0.00000000 10.54 0.55 11.09 12.14 86.82 91.35
> CCSD -156.83775164 148.06 2.18 150.24 154.14 96.06 97.47
> TRIPLES -0.02781518 355.94 3.37 359.31 359.32 99.06 100.00
> CCSD(T) -156.86556682 514.57 6.11 520.68 525.65 97.89 99.05
> TOTAL 0.00000000 533.54 8.54 542.08 553.09 96.47 98.01
>
>
>
>
>
> Now we switch temporary file directory to be 3TB RAID0 composed of 3 7200RPM 1TB disks.
>
> Here is IO bench with 10GB scratch file using our RAID0 composed of 3 1TB disks of 7200RPM speed as a scratch directory:
>
> IO_TEST 200 MB
>
> Test I/O on file length 204.5Mb with transfer size 0.50Mb ( 409 segments) and 10000 seeks
>
> Test name Description Time/s Rate/(Mb)/s
> SW Sequential write, 1 process file 1 0.76 269.08
> SR Sequential read, 1 process file 1 0.19 1076.32
> SL Seek, 1 process 0.15 66666.73
>
> IO_TEST 10GB
>
> Test I/O on file length 10239.5Mb with transfer size 0.50Mb ( 20479 segments) and 10000 seeks
>
> Test name Description Time/s Rate/(Mb)/s
> SW Sequential write, 1 process file 1 33.09 309.44
> SR Sequential read, 1 process file 1 2.84 3605.46
> SL Seek, 1 process 0.15 66666.73
>
>
> As a single drive SSD was faster than the single regular drive. I think single disk of our RAID has around 90-100 MB/s transfer, so 3 of them in RAID0 give around 300 MB/sek. Also Apple does not use the best SSD drives on the market, Intel's one or OCZ models would perform better.
>
> Hope that helps
> Jacek Klos
> Chemistry at University of Maryland
>
>
>
>
>
>
>
> On May 3, 2011, at 9:46 PM, Joseph Lane wrote:
>
>> Has anyone had experience running Molpro using solid state drives for
>> the temporary files instead of conventional hard drives? I am
>> particularly interested in any benchmark jobs obtained at the coupled
>> cluster theory level of theory.
>>
>>
>> Kind regards
>>
>>
>> Jo
>>
>> --
>> --------------------------------------------------------
>>
>> Dr Joseph Lane
>>
>> Lecturer in Physical and Theoretical Chemistry
>>
>> Department of Chemistry
>>
>> University of Waikato
>>
>> Private Bag 3105
>>
>> Hamilton 3240
>>
>> New Zealand
>>
>>
>>
>> Ph: +64-7-838-4466 ext 8549
>>
>> Fax: +64-7-838-4219
>> _______________________________________________
>> Molpro-user mailing list
>> Molpro-user at molpro.net
>> http://www.molpro.net/mailman/listinfo/molpro-user
>
>
>
>
>
--
--------------------------------------------------------
Dr Joseph Lane
Lecturer in Physical and Theoretical Chemistry
Department of Chemistry
University of Waikato
Private Bag 3105
Hamilton 3240
New Zealand
Ph: +64-7-838-4466 ext 8549
Fax: +64-7-838-4219
More information about the Molpro-user
mailing list