COMPUTER SUPPORT FOR PROJECTS PERINSUR AND CAPER
Document Type:
Collection:
Document Number (FOIA) /ESDN (CREST):
CIA-RDP85B01152R000700910029-2
Release Decision:
RIPPUB
Original Classification:
K
Document Page Count:
4
Document Creation Date:
December 21, 2016
Document Release Date:
June 26, 2008
Sequence Number:
29
Case Number:
Publication Date:
June 29, 1983
Content Type:
MEMO
File:
Attachment | Size |
---|---|
CIA-RDP85B01152R000700910029-2.pdf | 144.52 KB |
Body:
Approved For Release 2008/06/27: CIA-RDP85BO1152R000700910029-2
?
ROUTING AND RECORD SHEET
SUBJECT: (Optional) DD/A Registry
Computer Support for Projects
FRO
EXTENSION
NO.
ODP-83-904
D
ODP
/
2D00 Hd
s
DATE
q
.
TO: (Officer designation, room number, and
building)
DATE
OFFICER'S
COMMENTS (Number each comment to show from whom
RECEIVED
FORWARDED
INITIALS
to whom. Draw a line across column after each comment.)
EO/DDA
7D18 Hdqs.
983
;
`~~~~
2.
ADDA
7D18 Hdqs.
3.
DDA
7D18 Hdqs.
5 JU
1983
e -4.
7.
8.
9.
10.
11.
12.
13.
14.
15.
FORM 610 USE PREVIOUS
I-79 EDITIONS -
Approved For Release 2008/06/27: CIA-RDP85BO1152R000700910029-2
Approved For Release 2008/06/27: CIA-RDP85BO1152R000700910029-2
DD/A Registry
ODP-b3-904
2 9 JUN 1983
NF;t,i0RANDUiyi FO1i: Director of eersoniiel
F'hOi'A :
SUEJ1sCT:
iiEH'LELI'NCE:
Director of Data rrocessine,
STAT
Coruputer Support for Projects
Your kemo, Same Subject, Ltd. 21 Plar.' b3
(ODP-63-456)
1. We have investigated all of the reasonable, orations for
ssuu_~uoting your two Hicrodata Reality applications,
at two different sites. Both applications currently run on a
Microdata 8000 system with an older 6000 system being used for
development work. cost-effective o tion would be to move
the 8000 system to support and e the
memory of the 6000 system to 128K and use it for The
disadvantages of this alternative, however, are the questionable
reliability of the 6000 system and its inability to provide good
response at high concurrent user levels.
2. Attachment A presents the tabulated results of a series of
benchmark tests run on-both the 6000 and 8000 systems by ODP in late,
1980. Another series of benchmarks were run recently which re-
verified these earlier results. A significant point to be noted
from the Attachment A tabulation is that the performance of the 6000
series machine begins to degrade significantly beyond the six
concurrent user level even at its maximum memory configuration (128K
bytes).
3.- Since it is our understanding that th -application
will have to support up to nine concurrent users, using the 6'000
system for this application may'present performance problems during
peak loading periods. Because of the performance problem just cited
and because of the potential for significant hardware outage on the
60Q0 (it is a very old machine which uses ferrite core memory),
Processing recommends that, if funds can be ied, a second
Micro Data 8000 system be acquired to support when it..moves.
STAT
STAT
STAT
SIAI
STAT
SIAI
STAT
STAT
Approved For Release 2008/06/27: CIA-RDP85BO1152R000700910029-2
Approved For Release 2008/06/27: CIA-RDP85BO1152R000700910029-2
0
?
l'tnis would leave tine current 6000 anu b000 machines in their current
location to support Iand development activities. SIAI
4. ata bUUU system can be acquirea at an approximate [-TAT
bytes of memory for the 6UdO would cost about) Since ODP STAT
did riot anticipate or budget for increased ilicrodata Reality
capacity requirements, no funds ar.e available for this purpose in
USP. 11.1 S P n "I~ fund this increase as a cost of moving its
personnel STAT
Please let us know how you wish to proceed.
STAT
cc: deputy Director for Administration, w/att
Approved For Release 2008/06/27: CIA-RDP85BO1152R000700910029-2
Approved For Release 2008/06/27: CIA-RDP85BOl 152R000700910029-2
Attachment A 13 February 1981
Benchmark Results
All programs used in the benchmark display menus at the terminal. The
menus are "filled in" automatically with data that has been stacked in the
terminal's input buffer. In a user environment, the ENTER key would be hit
after each field in the menu was "filled in" and again after the entire menu is
complete. Average response time was computed, therefore, as the total time for-
a complete menu cycle divided by the number of fields in the menu plus one.
Since the time for a_-complete menu cycle includes the time required to display
the menu, the values in the table below may be higher than the response time
that the user would actually experience.
-------------------?----------------=--?------???--?....???-?
? Average Response Time (seconds)
Users 6000-64K 6000-128K 8000-256K
-------------------------------------------------- m -------
?
1 ..1.0 10.0*
3.0
2 3.4 6.0
3.0
3 4.7 6.0
3.7
4 8.0 6.7
4.9
5 10.8 7.4
5.2
6 18.0 9.0
6.1
7 24.2 11.4
7.2
8 - 11.7 .
8.1 ?
9 - 14.2
9.4
it _ ~ -
?
11.2
1.. - 19.4
-
`~ - -
15.9
---------------------?---------------------------
--------
* An anomaly probably associated with system start-up procedures.
Approved For Release 2008/06/27: CIA-RDP85BOl 152R000700910029-2