Quantcast

32 v 64 bit SPSS

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

32 v 64 bit SPSS

Osterlind, Steven J.
I am either buying or building a new computer and want to learn from your experiences. I rely on SPSS for intense processing with large data sets. Two questions:
1.      Is 64-bit versus 32 bit SPSS noteworthy better?
2.      Does SPSS respond best to disk-writing speed or in processing speed?

Thanks,
Steve Osterlind


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Steven J. Osterlind, Ph.D., Professor
Measurement & Statistics
University of Missouri
(573) 268-7514 - cell
[hidden email]

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: 32 v 64 bit SPSS

Barnett, Adrian (DECD)
Hi Steven
There are a few 'it depends' with this.

64-bit is better with big data sets IF you've got lots (in this context I'd say > 4 GB) RAM. Below that it won't make any difference. The more you've got above 4 GB, the more work can be done in RAM, thus reducing the amount of paging to your hard drive.

Speed of your hard drive will affect your speed of processing because it does thrash the HDD even when you've got plenty of RAM. If you can afford it, and if your IT Department Stalinists will allow it, I'd recommend a solid-state disk to use as the place that SPSS and your operating system will use for their paging files. But under any circumstances, when you are running with big data sets, I'd suggest that getting fast hard drives is always better than slow ones. I'd stay right away from 5400 rpm drives if you can, and go for a decent 7200 rpm drive.

One other factor to consider in the 64-bit versus 32-bit decision is whether you need to use ODBC. If you do, then the other applications you do ODBC transfers to and from needs to be the same bit-width as SPSS. As has been stated here earlier, ODBC transfers can only be done between applications of the same bit-width: 32-bit to 32-bit, or 64-bit to 64-bit. So if you are likely to need to talk to MS Office applications, for example, via ODBC, you will need them to be the same width as SPSS. This situation is painful at the moment. Not many organisations will have deployed 64 bit Office (which means Office 2010, if memory serves. It comes in both 32-bit and 64-bit versions). Office 2010 at present will not let you have both 32 and 64 bit versions of its applications installed at once. So if you need to have 32 bit Access to talk to a corporate database, that means you won't be able to write to Access from 64 bit SPSS. The only workarounds for that situation that I've been abl!
 e to think up are (1) writing stuff out from SPSS as a bit-width-neutral format like CSV or (2)using a format translator like Stat/Transfer to go from SPSS to say Access. Not insurmountable problems but a nuisance.

When I compare how well 64 bit SPSS performs compared to 32 bit, my home system with 64 bit SPSS 18 flies by comparison to the 32 bit version under XP at work (although the CPU is much faster at home and has more RAM - 8GB vs 4GB).

Hope this helps

Regards


Adrian Barnett
Project Officer
Educational Measurement and Analysis
Data and Educational Measurement
DECS
ph 82261080

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Osterlind, Steven J.
Sent: Tuesday, 17 August 2010 11:55 AM
To: [hidden email]
Subject: 32 v 64 bit SPSS

I am either buying or building a new computer and want to learn from your experiences. I rely on SPSS for intense processing with large data sets. Two questions:
1.      Is 64-bit versus 32 bit SPSS noteworthy better?
2.      Does SPSS respond best to disk-writing speed or in processing speed?

Thanks,
Steve Osterlind


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Steven J. Osterlind, Ph.D., Professor
Measurement & Statistics
University of Missouri
(573) 268-7514 - cell
[hidden email]

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: 32 v 64 bit SPSS

John McConnell-2
In reply to this post by Osterlind, Steven J.
Steve

On 2. I've always recommended faster disks ... especially for larger
data volumes. When spss runs any procedure it is sequentially reading
and writing from/to disk. Of course a better CPU will help with the
processing but faster disks probably help more (IMO).

I have just switched to 64 bit but haven't perceived a difference. I'll
run a test and let you know.

Regards
john


John McConnell
analytical-people
m.      +44 (0)7711 786600
o.      +44 (0)845 680 1871
500 Chiswick High Road
London, W4 5RG
UK

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Osterlind, Steven J.
Sent: 17 August 2010 03:25
To: [hidden email]
Subject: 32 v 64 bit SPSS

I am either buying or building a new computer and want to learn from
your experiences. I rely on SPSS for intense processing with large data
sets. Two questions:
1.      Is 64-bit versus 32 bit SPSS noteworthy better?
2.      Does SPSS respond best to disk-writing speed or in processing
speed?

Thanks,
Steve Osterlind


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
Steven J. Osterlind, Ph.D., Professor
Measurement & Statistics
University of Missouri
(573) 268-7514 - cell
[hidden email]

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Regression Interaction

Salbod
In reply to this post by Osterlind, Steven J.
Good morning, Folks,

I'm planning on doing a series of tests (6) for moderation. For example, I want to test if gender (female=1, male=0) moderates the relationship between belief in UFO visitation (UFO,  scale ) and Loneliness (LONE, mean-centered). I plan to use a hierarchical multiple regression in this case to test for the moderation effect. At Step 1 I will enter LONE and gender and at Step 2 I will enter LONE, gender, and LONE*gender. At Step I will test for R square change.

In this situation couldn't I have used a simultaneous regression analysis, and entered, LONE, gender, and LONE*gender in one step?

Any suggestions or references that discuss this issue will be greatly appreciated.

TIA

Stephen Salbod, Pace University, NYC

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: 32 v 64 bit SPSS

Francien Berndsen
In reply to this post by John McConnell-2
In my opinion 64 bit is much faster. So fast that sometimes I'm not
able to do anything else than SPSS, because SPSS consumes my RAM
totally.

I changed from winXP & SPSS 17 (32 bit) to win7 & SPSS 18 (64 bit) on
the same computer, and I noticed that it runs almost twice as fast (with
very large files). But it will depend on your computer hardware.

The only con, but for me it's a big one, that it's not possible to
export a file to excel or access if you don't have 64 bit version of
Microsoft office.

Regards,
Francien


On Tue, 17 Aug 2010 07:37:16 -0500, John McConnell
<[hidden email]> wrote:

> Steve
>
> On 2. I've always recommended faster disks ... especially for larger
> data volumes. When spss runs any procedure it is sequentially reading
> and writing from/to disk. Of course a better CPU will help with the
> processing but faster disks probably help more (IMO).
>
> I have just switched to 64 bit but haven't perceived a difference. I'll
> run a test and let you know.
>
> Regards
> john
>
>
> John McConnell
> analytical-people
> m.      +44 (0)7711 786600
> o.      +44 (0)845 680 1871
> 500 Chiswick High Road
> London, W4 5RG
> UK
>
> -----Original Message-----
> From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
> Osterlind, Steven J.
> Sent: 17 August 2010 03:25
> To: [hidden email]
> Subject: 32 v 64 bit SPSS
>
> I am either buying or building a new computer and want to learn from
> your experiences. I rely on SPSS for intense processing with large data
> sets. Two questions:
> 1.      Is 64-bit versus 32 bit SPSS noteworthy better?
> 2.      Does SPSS respond best to disk-writing speed or in processing
> speed?
>
> Thanks,
> Steve Osterlind
>
>
> . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
> .
> Steven J. Osterlind, Ph.D., Professor
> Measurement & Statistics
> University of Missouri
> (573) 268-7514 - cell
> [hidden email]
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Loading...