26 grading this material, particularly as it relates to the inexperienced user, In general we are attempting to up-grade the various DEC- originated subsystems to the newest versions to increase the chance of compatibility. We have recently done this with FORTRAN and MACRO and will bring the other programs along as soon as possible. The whole issue of compatibility is one which will receive attention. We will not be able to commit the necessary resources within available manpower to redo the TOPS-10 emulator correctly, but will keep chipping away. Some requests to look into additional software subsystems have been received and we will consider mounting them if the community develops a definite need, Suggested augmentations include the Simulation language SIMULA, the medical information system language MUMPS, BPL or APL, and subroutine libraries such as the biomedical Statistics package. We will continue our efforts in the building of user subroutine libraries, concentrating initially on SAIL. Informal Information Access: One characteristic of the SUMEX community is the diversity of information, formal and informal, which flows around the system or is available from users. We want to work on ways to capture that information and direct it to other interested individuals. We have two major repositories for information, the directory which contains the formal documentation of system facilities and procedures and the directory which is an accumulation of informal and dynamic information. We will be working on capabilities both to ease the entry and cataloging of information into these areas and to assist in guiding the user to that subset which is of interest to him at a given time, These user-oriented lookup protocols are, of course, strongly related to the problems of adaptive user interfaces to the system and each will benefit from the experience of the other. Community Management: We will continue to work with the management committees to recruit the additional high quality projects which can be accommodated and to evolve resource allocation policies which appropriately reflect assigned priorities and project needs. We hope to make more generally available information about the various projects both inside and outside of the community and thereby to promote the kinds of exchanges exemplified earlier and made possible by network facilities, The first workshop will provide much useful information about the Strengths and weaknesses of the performance programs both in terms of criticisms from other AI projects and in terms of the needs off practicing medical people. We plan to use this experience to guide the community building aspects of SUMEX-AIM, 27 II.B SUMMARY OF RESOURCE USAGE The following data give an overview of the resource usage from August 1974 (when we began keeping detailed statistics) through March 1975. There are four sub-sections containing data respectively for 1) resource usage by community (AIM, Stanford, and system), 2) resource usage by project, 3) Network usage data, and 4) measures of diurnal variations in system loading. II.B.1 RELATIVE SYSTEM LOADING BY COMMUNITY The SUMEX resource is divided, for administrative purposes, into 3 major communities: user projects based at the Stanford Medical Sehool, user projects based outside of Stanford (national AIM projects), and systems development efforts. As defined in the resource management plan approved by BRB at the start of the project, the available resource will be divided between these communities as follows: CPU Usage - Stanford HOS AIM 4OZ System 20% File Space - Stanford 25,000 pages AIM 25,000 pages system 30,000 pages We have recently brought an additional 40,000 pages of file space on-line (this happened since March and hence is not shown in these data) - this additional space will be equally divided, for the most part, between the Stanford and AIM user groups. The system file requirements include all of the subsystem files, documentation files, and other system-related files shared by all users in addition to staff directories. We expect the system allocation to grow somewhat more slowly than user space requirements, reflecting primarily the addition of new subsystems, documentation, and user information from time to time. The following plots show monthly usage of CPU and file space resources for each of these three communities relative to their respective aliquots. Note that file space data are not available for November and December 1974 because we were making a transition between numeric and alphameric account designations. As is evident in the data, many of the national AIM projects joined the SUMEX community during this year (as communication facilities became available) and their use of the facility began Slowly. The rate of remote use is expected to increase substantially during the next few months. 402% 40% 20% CENTRAL PROCESSOR USE BY COMMUNITY 28 National AIM (Entitled to 402) Aug Stanford (Entitled to 40%) Sep Oct 1974 Nov Dec Jan Feb 1975 Mar Aug System (Entitled to 20%) Sep Oct 1974 Nov Dec Jan Feb 1975 Mar Aug Sep Oct 1974 Nov Dec Jan 29 FILE SPACE USE BY COMMUNITY 25,000 National AIM (25,000 pages) _ a SS ee 0 T q t r t 1 q ' Aug Sep Oct Nov Dec Jan Feb Mar 1974 1975 25,000 Stanford (25,000 pages) 0 T T T T T Tt T T Aug Sep Oct Nov Dec Jan Feb Mar 1974 1975 __ _ — — — System (30,000 pages) 0 Tc T T T T T Aug Sep Oct Nov. Dec Jan Feb Mar 1974 1975 30 II,B.2 INDIVIDUAL PROJECT AND COMMUNITY USAGE The table following shows average resource usage by project in the past grant year. The data displayed include a description of the operational funding sources (outside of SUMEX-supplied computing resources) for currently active projects, average monthly CPU consumption by project (Hours/month), average monthly terminal connect time by project (Hours/month), and average file space in use by project (Pages/month, 1 page = 512 computer words). Averages were computed for each project for the months since August 1974 or since the project was admitted to the resource if after August 1974. Many of the national AIM projects joined the SUMEX community during this year (as communication facilities became available) and their use of the facility began slowly. For this reason, many of the average usage figures for these projects are lower than is representative of their activity currently. We expect the remote usage to increase significantly in the next months. 31 RESOURCE USE BY INDIVIDUAL PROJECT STANFORD COMMUNITY 1) 2) 3) 4) DENDRAL PROJECT "Resource Related Research Computers and Chemistry" NIH RR-00612 (3 yr award) $323,933 this year MYCIN PROJECT "Computer-based Consult. in Clin. Therapeutics" HEW HSO-1544 (3 yr award) $124,000 this year PROTEIN STRUCT MODELING "Heuristic Comp. Applied to Prot. Crystallog." NSF DCR74-23461 (2 yrs.) $88,436 total PILOT PROJECTS (see reports in See IV.B.1) COMMUNITY TOTALS NATIONAL AIM COMMUNITY 1) 2) 3) DIALOG PROJECT "Computer Model of Diagnostic Logic" HEW MB-00144 (3 yrs.) $125,027 this year Higher Mental Functions "Computer Models in Psychiatry and Psychother." NIH MH-06645 (3 yrs.) $170,000 this year NIH MH-27132 (2 yrs.) $130,000 this year MISL PROJECT “Medical Information systems Laboratory” HEW MB-00114 (2 yrs.) $380,619 this year CPU (Hrs/mo) 46.38 10.17 3.62 8.54 0.68 CONNECT (Hrs/mo) 728.5 186.1 139.0 1262.0 159.9 53.1 FILE SPACE (Pages/mo) 12325 3264 933 2063 1021 242 4) RUTGERS PROJECT “Computers in Biomedicine" NIH RR-00643 (3 yrs.) $285,240 this year 5) AIM Administration — COMMUNITY TOTALS SUMEX SYSTEM 1) Development 2) Operations COMMUNITY TOTALS RESOURCE TOTALS 32 .26 351.5 2659 33 II.B.3 NETWORK USAGE STATISTICS NETWORK USAGE PLOTS The following are plots of the major billing components for SUMEX-AIM TYMNET usage. These include the total connect time for terminals coming into SUMEX and the total number of characters transmitted over the net. The ratio of characters received at SUMEX to characters sent to the terminal is about 1:14 over the past couple of months, The plot on page 3 is for TOTAL character traffic. Note that the high usage in October is because we were doing a great deal of testing of the net at that time. As of March, we announced that the IN-WATS lines would be terminated soon and that users should transfer to the TYMNET. Extrapolation of these data should be done very carefully because the apparent steep rise in February and March has apparently been tempered by the ARPANET connection becoming operational in late March. The Rutgers community, at least in part, is using the ARPANET in place of the TYMNET because of more convenient access through their TIP. 34 TYMNET CONNECT TIME (IN HOURS) 1000- 800- 600- 400- 200- Om tea eta nn pan wpe a pen np ee pe ep ee pn ap nn ep ee OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC JAN 1974 1975 1976 TYMNET CHARACTER TRANSMISSION (MILLIONS OF CHARS) 20-} 15- 10- On tanta ante n npn np 4 4 pn ng ng ng ep eg gp ne OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC JAN 1974 1975 1976 35 II.B.4 SYSTEM DIURNAL LOADING PROFILE Since late February, we have been collecting more detailed information about time variations in system performance and loading, These data include user-oriented measurements such as load average, number of jobs, and percent of LISP use. In addition, we collect data characterizing internal system operating parameters such as paging rate, trap overhead, scheduler overhead, and drum use. Systematic measurements are taken every 20 minutes to give an overall picture of diurnal variations. Periodically we observe system dynamics on a much shorter time scale as well to observe more of the internal workings of the system - such data are so voluminous, however, that they cannot be collected continuously. The following graphs give a feeling both for the type of data available and the loading characteristics they reveal. The plots are organized 2 per page with the quantity displayed labeled at the top of each graph along with average and extremum values of the data. The first 3 pages display composite data for the month of April 1975 - derived by averaging together all of the weekday data for that month as a function of time of day. The next 3 pages show somewhat different data for a single day recently in May (note the hash marks on the right side of these plots signify that the machine was down [for repair]). The most striking feature from the monthly data is the expected peak loading during prime time and in fact a bi-modal distribution is apparent in some of the plots approximating the 3 hour shift between prime times on the east coast and west coast. Another striking feature is the dramatic difference between the monthly average data for April and the individual daily data. We feel there are at least two effects taking place to account for this difference in addition to simple day-to-day loading level variations. First, daily data is considerably more "noisy" than the average data with peak loading occurring in cycles spaced about 2-3 hours apart. We observe the phases of the various sub-peaks relative to time of day to be unpredictable other than gross effects leading to the bi-modal average data corresponding to heavy mid-afternoon use of the respective coasts. For example, whereas the peak in the monthly load average curve occurs around 1500 PDT, the peak for the daily plot is around 1600-1700 and the daily plot displays much higher "momentary" (20 minute average) excursions. In fact, over shorter intervals (1-2 minutes) load average peaks are even higher (load averages approaching 10 or more are not uncommcn under present loading in the afternoon), Thus in taking the average for a month, the peak loading extremes which strongly influence the subjective "feel" of the system on a daily basis are blurred out and lost. Secondly, we are apparently on a rather steeply rising curve in terms of system utilization. The composite load average data for March, for example, had a peak value of 2.7 as opposed to 3.7 for April. Of course, two points do not necessarily constitute a trend but this data corresponds with the subjective comments contributed by users in their project reports and other commentaries on system response. There is general agreement that loading has been increasing noticeably. 36 Several points mentioned in the report can be observed in the daily loading data. During peak loading, drum use somewhat exceeds the 3300 page fixed head capacity we now have on-line, We expect to mitigate this overflow through more intelligent management of Swapping Storage. The total number of jobs follows closely the average data plots but other measures of system use like load average and LISP job consumption tends to skew toward the evening hours. This is evidently a natural leveling process wherein the very large jobs have more difficulty contending during the day and the program developers have shifted their schedules to work later at night. Observe the very heavy (and efficient - 96%) use during the evening hours when the load average has dropped to about 3 or 4 and is arising to a large extent from LISP usage. It should also be noted that load average and overall percent usage are somewhat independent. Usage measures how many cycles are being consumed and load average measures between how many users these cycles are being divided (one CPU-bound job would Show a usage of 100% but a load average of only 1.0). These data, while preliminary in the sense that we have only the trends of a couple of months to go on, are consistent with user comment and are the basis of our preliminary proposal to up-grade the SUMEX~AIM CPU (see page 4 AND Appendix C). While load-leveling (such as moving work on LISP programs to evening hours) may be acceptable (and desirable) for some program development work, it is unrealistic to ask physicians to adapt in a similar manner in trying out the AI programs. Our development and managerial incentives are directed toward making better use of the off-hours, freeing up prime time where possible. But, we must anticipate the need for more computing capacity during prime time, especially based on these measurements and the anticipated growth of the user community. We are continuing active work on the CPU up-grade plan discussed elsewhere in this report (see page 51). ave Tne 2 Pac 37 RAGE DATA PLOTS © WEEKDAYS, APRIL 1975 ludes dates? TUESDAY, 1LeAPRe75 WEDNESDAY, 2eAPRe7§ THURSDAY, 3eAPRe7S§ FRIDAY, GeAPRw75 MONDAY, TeAPRe7§ TUESDAY, BeAPR=75§ WEDNESDAY, 9*APRe75 THURSDAY, LO*APRaTS5 FRIDAYs {1eAPRe7§ MONDAY, {&nAPRe75 TUESDAY, 15eAPRe75§ WEDNESDAY, 16*APRa7§ THURSDAY, L7eAPReT§ FRIDAY, 18eAPR=75 MONDAY, 21"APR75 TUESDAY, 22eAPRe75S WEDNESDAY, 23"APRaTS§ THURSDAY, 24"APRa75 FRIDAY, 25eAPRe7§ MONDAY, 288APRO7S TUESDAY, 29sAPROTS§ WEDNESDAY, 30*APRe75§ CUMULATIVE UPTIME: Number of Days! Data in Average Plots Qe} { ( eee | Ty) CI I 00 6 8 PerDem eeeneereeeoenore «| OOOO C08 FEPEERenooRoneseSODOROED fe 6 { OOHPOH OR DERE EERO DRO RRO O ODER OOOO ROR REORO everens { ORPOPOOROHOO ERODE OR ORE OPO ERO DO ROR EREDAR DO HORROR ROR OOO ORO ODE fe POOCOODEERACOEOE AD ROOOO OOOOH RR OSEO ODEO DOOR ADORED PERFOR AO RR RHODES [POOPROLOOOO EOE E DCEO RE ROOORODOREROEOEREORE ROR EDO OO ORO ORR ORR RODODED ee @ | PCHOOCOOOHE ODO ECOOOOREREDOO ODOR ODEO HORE O ROR DOOR REDDER ROR SOR ROEE eee | POOOCOOOODO OCOD EOE RSE DODO ODEO ERED OR ODEHOOD ORO DELO DE DERE RR REC RO REE OEDE [COCOOPOPODE RODE DOO DODO DOOR ORORO DODO OORODODEOOR ODED RO OR DOR OR RO ORO RRO ERO [AOCPROD ROPER DEOL OR DOR EOP ODE ROO HDRE DRO OO DORR DORR RCRD RRR OR DOOR OR ORR ORE [OPPOROOORECEHEHORORODEREAEOOOHODEDEDE OOOO OO ROD ERO R ORR ROHR OO ORDER ODER REE =| POPHCOORRAP ODODE ODED ER DE OPO POO DEODORR OOOH EERO ROR ORE ORO RROD OCDE ORO OO OR ORD [OOOOH ERROOO EEE OOREOROROOEOO ODO OORO DEO OREDOORDOR ERED RESO RROD ORO ROR ROR RROD 1 OOSOOOLEDEOEOHE DOE OEAODO LODE OOOO OOOO ROE PPOROREOD ERODED ODODE O DOS O OREO NORD | OPROCOH ERO ROR ODO OPE DE OOOO ERO DOOOO DODO DEO OR ORR ORO ROR ORE DOR O ROOD ERR OEEE | OOCOHORO LORE POO DORE DORE OE OOOO OOOO REORDER OR OOORODED ROO REDO ROR OR ORR RRORE Seer emp nmemntomenatmesendaacenenenantaennetonneedaccentnanectheeucudnececs TIME 0 2 4 6 6 10 ie i4 16 18 20 22 24 38 AVERAGE DATAS Load Average (Lows 0, Ave® 1,5, High# 3,7) 12°| WEEKDAYS, APRIL 1975 . e oe 8 | epenecenoone CORO BEHEO HEH HH CHOOOHED ® e OKEKeRasooReeRenereEHeeaeees eee 60 {ee PHPHOOCHHOEHOCHOCHENEHSOHHOP OHH AEHHE Ceeoneeoree {[eseeese 6 POOOOPELOOORORPAE PEPER OHOCOR OPE SOOO RCEOOOO EOP OH FOOD PAC sn we pa tnenwesensend naenn dew aaatananagevesatasenntewsengaaamapamaaabenanad TIME 4 2 4 6 8 40 12 i4 16 18 20 22 ea AVERAGE DATAS Total Number of Jobs (Lowa 0, Ave® Sige, Highs 19,8) 326] WEEKDAYS, APRIL 1975 | { { { of | { © ote erenenns { CReO ERODE RESO ORES ef Poon ERROR ROE RECO R OO REE { ORK Co eR EH OER EHR COREE OEE FE { ITI ee ( POKER LE RACE EAHEHEHOHOROR OOO EEE eesse [ rYYITITIIILII I »fe OOPPORCARPEEPOOEOROSOO OOOO OREO ADOO ERED OOOH OREO DO FEE (eevee PREOCOHE ROO REFODEOPOH OE HE HORORAEO CHER ODOR ORE HOO REOOR fecenecece @ CORORREPOPOEROODEERE DODO HOPPER HEHE RODE REECE HEHEHE DOD OOE | PPPHOPO EPA DOEEE POORER LODO OHO OO DERE PREROL OPED ODODE PEOO OOOO OHO PO DOOEOOOEO { OOO OP DO OOOOROR OREO ORE OPEOHODORAOOLODOPEOOPDO OOO HO EORODOROP OOOO DORR OO ORO PAC Seem asenenetaseecteuwanterussataenentuvowateseuetreuaagepoantuaaempouuuad TIME 0 2 a 6 § 10 12 i4 16 18 20 22 24 39 AVERAGE DATA’ Number of LISP Jobs (Lows ,0¢ Ave® ied, Highs 3,2) 16! WEEKDAYS, APRIL 1975 . = . eenee Coeereneneneeoees oe POPP COHEHHO HSH OHH OOH OAOHEEEED eeeeneeese eereeneseee OPOPHOOHHEO OOH OHHOHEKROEHOOREOHOSHOOHHOCEHOEOS OREO HEEROE PAC $86 Oh HORE POTEET HEHE SREEE HSER ERTEREHNKEDSTTERHUE EES EPUNEHR HERA NEM HE OHS TIME 0 2 4 6 8 10 t2. °° «14 16 16 20 22 24 AVERAGE DATAL X% LISP Usage (Lows ,0, Avem 10,5, High 24,0) 1001 WEEKDAYS, APRIL 1975 | e f s » e ee ee e eeoe Che ee eeeee seeeeeoese # 66 6 eeceoeroes eee COOHOCEEHEHE SOL OHEHEOOHHOEHES © 6 eveseeceee noe eee Jeesee @ ee CoPeHeeCEeRECEeHeROCO HEHE HR RER LEB R OH RHE REE OB HO ROO ERE HF Pac ‘ Peww eed arash we awat ar ew SSE SaRe CEH ERE RET ETE RE RTOT ROR NAN PEE ERatePEE nen EaEah TIME 0 2 4 6 8 10 12 14 i6 18 20 22 ed 40 MONDAY, 19=MAY753 Total Number of Jobs (Lows 6,7, Avet 18,5, Highs 30,7) 320] “NINA { 6 ® NANAN [ a een /N\INA { eee eoeee NANAN [ © 8 #00 senennene 6 ‘NINA of © 8 P88PePoeCOOeFDe CF \/NAN { PLCC Re HD OBHA Ren O ROR BB RAE HORE Be8e /\/N\/ | © COnDE POE DEFOR COE ARE Rp REC ERA OROEeROe N/N/\ { COSC FOC ER EOP RP ORO HCC ORE ROC ROROR EE HBAe /\/\/ { ROGHROROR REPO PREAH OBR ECC ROR OD REO ROO RO \/N\/\ efe Ir eaves [ee ee 00 PPPHRHCE DORE P OOO RE HOOPER DOLE DORE AOOOROROAON/\/\ {eseeece POPHHCEKEO HELE EREHEP HHL O DEL HCAD RECO RE RH RE RO ADR O/\/\/ {ecevesce PHOHOSO POOH OPPO OOOO ESP O OOO OHP ORES O OOO DOCH O RE ARON/\/\ TIT I III rr r PAV AV, SLOCHCPOOLROCR LAO HOEOH OHAOEOKOHEH OH OEHAHAPE SE RERE DEO OR nH BRA EE AHAEHEHOE\/\/\ [CORO HOHE HE RAEHEHHAHHREHACHERAHHO CHEN ROHERHLREA PA RRPRDAC HE ReHO HORACE /\/\/ [OOOOH HAGHOHADHDORERORADEOPHOAARODOENAGHEHOHAODE OHO HHRDELORADHEECEDOOAN/N/\ RIT AVA, [@COPOPHREEECHEOOOEEEHARSEHOCOOEDE ERR O CREF OCHRE LEP EO RHE RO EOE HOO RHE RRO AN/N\/\ Pate ‘PS awenansewevstoreawse wane aeshbasaenpavesansusventeauseteasec¢usocuspurmanmasann sah TIME 0 e 4 6 8 10 12 44 16 18 20 22 24 MONDAY, S9=MAY=753 Load Average (Lowe 40s Aves 2,4, HighB® 7,7) t2e| 4\fN/ { \/NAN “N/M NANAN NINA o NANAN { 4NIN/ e N/NZN @ 4N\/N/ oe e AYA . e eeee oe 4N\/N/ e seen ene N/NAN ee Coepeneoes e0/N/N/ ee ee ePeRCoeOD @ COON/N/N ee COCR eb eeRoneceoS @ #ee/\/N/ | eee CopeoseresceEeREeHOH 686 BOON/N/N | e CCFO ef cEPenoeRsoo HORE ere R HORROR Ee/\/\/ e CPORHOCHR CHOG RAH O FORE HO OEHHA BOO OHO OREO /\/\ e 6 HEPC HRE HERO HERR O DERE C RFR B RE CORE HOR RO HOE/\/\/ 60 eefeee ff 8e Be COOH COHHERECEOHROREDP EEO OOO ROROHOHO HO OHO R OOOO /N/\ Pac | PORES RSENS TAREE RESET HECOT ANE RSH E Ad HERS EhesaNesaenantvanEahrsemuas¢aennansd TIME 0 2 4 6 8 10 12 14 46 18 20 22 24 4l MONDAY? 19°MAYe7S1 Number of LISP Jobs (Lowm 0» Avem 2el, Hight 5,9) jee /NAN/ . NANAN “NINA AYA 4NfNA e \/NAN /NINA N/NAN “NINA { \/NAN « “NINA N\/NAN @ ‘NINA ee V/NAN ease INANA a © © ePenecenroene NANAN eee e806 envoceeosecocooepeeees /\/\/ ee e CFO HOKE SEE OE RAH OR ESHOP HERERO OE OREN/\/\ [eeee eee COHOROO PEREHHOCREDE CH PERO E HoH AHO ROR ORE /\/N\/ [eeveeorcenvcononoee eaece TIT I III ILL LATAY Pac ‘Sew wwese sewed eesretansevevsvnastaeranadaqnvantavesadousraevosvatenawapaverny TIME: 0 2 4 6 8 10 $2.0 «4G 16 18 20 22 24 MONDAY, 19@MAY@7S2 % LISP Usage (Lows .0, Avemw 25,07 Highs 72,0) 100"| /N\fN4 , \/NAN { ‘NIN/ NANAN 4N\/N/ . NANAN oe 4NINC e even AYA ' e eeceee 0 4N\/N/ ® ® 6 eaereeee N/NAN e e e ee eoesenee 4‘NfN/ e e e 900 @ eeneseceee N\/N/N 6 e e een & eopeceeeens /\/N\/ e e ee Oe peeneceeneenea\/\/\ e e OCP ppeepceRnoR Ae EeRRO/\/N\/ ° ee e e e 0008 HODevesonooEPeneeREN/\/\ fo @ @e oe ee 06 6@ ee OCH HORE OR EHH SH COREEOE/\/N/ fe © e060 ene0cee e ee ef COHOROPOROO EERO OROOOOONOREOOEN/N/\ {@ © Sppsereronone eee COOCHEH EPEC EsEEOOoCCERB ERC ERO ROE ORBEO/\/\/ {oe @ eveeeroneorore eee COLEKOH HEHE HEHE HERERO CR EO RRL OBOE HE RORON/N/N Pac "owe eae we need ee ewes ROE POS PSEA R PO UEHESE FAME H SPEER ENT TTH HER ETASTANEasUEUNes TIME: 0 2 q 6 8 30 32 %@ 16 18 @ 22 24 42 MONDAYs 19=MAY=751 0 -% T(me Used (Lowm 2,07 Avem 56,0, Higham 96,0) 1008 {NIN¢ poner oo N/NAN | 6 e foebaR ane OO/N/N\/ { e e e OOOOH ORREROCEPEB\/\/\ { ee a0 OOF HEHO REDE R EH EOROORR/\/\/ ee F00G Ree PReRD ORR oHL HERE ORAN/N/\ POOH COOP SOOPER HED RA ROPE DEA OE EO ROA/\/\/ COCCH POR EALOHEROO HE OAH EAHA HAHA ORBAN /N\/\ © PHPCR POOH ER ERDF ER EPOEHAL EHO ROHL RB RE/\/\/ POPHOH EOE CHER EHRL OLE DR OA HE HOO HERO DOO OO\/\/\ CHOOEHE HE CHAOS HREL OEP HOBO BONER ARO REO HOO/\/\/ PPOEPEFOOEORO ROO EHOROORORHOOED CORE HERO B\/\/\ OOPFPEEHOROD OH OP HDOCAAPEONADADEDO OOO BBOA/\/\/ { e e SCOPOOP DOMED OOO RE OO OOOOH ORHEO NOE HED RHOHON/\/\ f @ S060 e COOOHORHOHOOAHO HEHE ROCHE DOOR ERROR BAHAR AHE/\/\/ {ee © @e00 6 6 6e FOF = AODOOROR OO OS OOO HO RERO ROE LO ORE EREP OOOO OMEN /N/\ [ee seeeee eeseses OHO AAPOOEP OOOOH OPED OROHOOOA RODE R OOOOH O LEO ROR/\/\/ [ecarecneescenosace SOORHPCC HLF OO AHO RHE OPER OREO HREOC OROnHOR RO RAAHAOOAN/\/\ {| eeerecancsoeaonsen SOHPOFH OKO ROOHORE HO LOH CORO RRO RO RAH OBHAOB RRNA ORBEA /\/\/ [@OPOROCECOO HEED OOOO RAD AOAEREOOORAOO PERO OO EEO DOO OO ROCH EOE REDOROOOOCERO\/\/\ PAC Ferment aweR atom eae THEE ROestEePEEtEw TENE HUET ESTEE ENS PORTER TURES HOE HRS TIME 0 2 4 6 8 19 je 14 16 18 20 22 24 esoeoeseevss = e ® e e e 6 MONDAY, 190MAYo7S3 Deum Pages {nm Use (Lows 405,07 Aves 2099.1, Highs 3690,0) 4500e) “NINA { NANAN { ‘NIN [ e N/NAN { eee 4NIN/ of eee N/NAN { © ercncee ee oe ANIN/ { e PPR E ROC ROReHe #08 eee ON/NAN [ oe POPOOHHHPCHEOOREREOR HOH ERO/\/\/ {e S08 PROROREDOR CERO OOOH FEHR RR EO OREN /\/\ «|e POPPER OOOO OROPEROROAH ORE OOOO EDOREO/\/\/ {ee 0 PORODOESH ODOR OR FOR ESH HOARD AERO RORP EN /\/\ leoee oe AAV, [esee see OPOPOCOROH ORO PE DECOO OHA H OOOH OREREO ESPON /\/\ {eceveeees ee 6 POPOCEFCOOSORORODOR ERE O ODORS EEREOHORR/\/\/ ol esesereoseeoe ee OF = POPEERPHOHOOOAHE DOO OH RO ROEHL ODOEHHADEROOON/N/\ | eeseneeonseoeneono POOOE POOLED E HOR EPAHOOOREROOOROOHEDEO ROR R AB AARE/\/\/ |eeeearenscenosnefae CHOKE D RABE E HOR ORO EE AEO OH RO NRO HE OL ORE HDR ERO RO CHB A\/\/\ [HOCHROELO OO ALDOR ORD OO OOOOCOOHODEEOOOPADO ORO OOO DOR SORORAERO ROH OOROR/\/\/ [OPOPPO LORE OOHOR OOOH OOOH OOO RAO OOPPE REED EOOREROP HOSE READ OHERODHOREOE\/\/\ PAC PERE SMH OKO TEPAM SOK eS TEBE ETHTARH EHO MASE OAH N +H EETSP HOSE TE HEARED TEER HN TEBAS TIME. © 2 4 6 8 10 12 44 16 18 20 22 24 43 II.C RESOURCE EQUIPMENT SUMMARY The following table gives a list of the items of equipment purchased to date for the SUMEX resource along with details on vendor, description, price, and date. MODEL DATE DATE PURCHASE SOURCE ITEM QUANTITY DESCRIPTION MANUFACTURER NUMBER INSTALLED ACCEPTED PRICE FUNDS KI-10 CPU 1 Central processor, Digital Equipment KI-10 3/1/74 4/24/74 $178,500 NIH including console Corporation Memory 3 Core memory (64K Digital Equipment MF-10G 3/1/74 4/24/74 $224,910 NIH words including 4 Corporation MC-10 memory parts 1 Core memory (64K Digital Equipment MF-10G | 11/74 12/74 $ 63,754 NIH words including 4 Corporation MC-10 memory ports 1 Memory port Digital Equipment MX~10 8/74 9/74 $ 4,770 NIH multiplexer C orporation Clock 1 Programmable clock Digital Equipment DK~10 3/1/74 4/24/74 $ 2,678 NIH Corporation Disk System 1 Single double density Digital Equipment RP-10C 3/1/74 5/1/74 disk controller Corporation 1 Memory data channel Digital Equipment DF-10 3/1/74 4/24/74 Corporation 4 Double density disk Digital Equipment RP-03 3/1/74 4/24/74 $108,153 NIH drives and disk packs Corporation 3 Double density disk Digital Equipment RP-03R 2/75 3/75 $ 44,636 NIH drives and disk packs Corporation Y MODEL DATE DATE PURCHASE SOURCE ITEM QUANTITY DESCRIPTION MANUFACTURER NUMBER INSTALLED ACCEPTED PRICE FUNDS Swapping 2 Fixed head disk with Digital Development A-7312-D-8 1/75 3/75 $ 37,206 NIH Storage 1.7M word capacity Corporation and 4 track parallel access 1 Special systems Digital Equipment RES-10 10/74 11/74 $ 81,090 NIH controller for Corporation DDC disks DEC Tapes 1 DEC tape control Digital Equipment TD-10 3/1/74 4/24/74 (TU-56) Corporation 1 Dual DEC tape drive Digital Equipment TU-56 3/1/74 4/24/74 $ 17,850 NIH Corporation Magnetic Tapes 1 Magnetic tape Digital Equipment T™-10A 3/1/74 4/24/74 (2 x TU-30) controller Corporation 2 Tape transports Digital Equipment TU-30 3/1/74 4/24/74 § 31,238 NIH Corporation Line Printer 1 Special systems line Digital Equipment Special 6/74 7/74 $ 7,208 NIH printer control for Corporation Data Products 2410 1 Line printer with Data Products 2410 6/74 7/74 $ 18,963 NIH 96 character drum, vertical format control, parity check SY MODEL DATE DATE PURCHASE SOURCE GT-40 1 Graphics terminal Digital Equipment GT-40 3/1/74 4/24/74 $11,156 NIH Corporation Line Scanner 1 Data line scanner Digital Equipment DC~10A 3/1/74 4/24/74 Corporation 1 8-line unit Digital Equipment DC~LOB 3/1/74 4/24/74 $ 16,275 NIH Corporation TYMNET 1 PDP-10 TYMNET TYMSHARE 8/74 10/74 $50,774 NIH Interface communications controller ARPANET 1 BB&N ARPANET/KI-10 Bolt, Beranek 1/75 2/75 $ 21,200 NIH Interface interface & Newman PDP-11/10 1 Communications Digital Equipment PDP-11/10 2/75 3/75 $ 13,445 NIH processor Corporation 9¥ MODEL DATE DATE PURCHASE SOURCE ITEM QUANTITY DESCRIPTION MANUFACTURER NUMBER INSTALLED ACCEPTED PRICE FUNDS Terminals 1 Terminal Data Terminals DTC-300 3/18/74 6/74 $ 4,597 NIH Communications 2 Terminals - Computer 311-3 3/18/74 6/74 $ 6,402 NIH Execuport portable Transceiver with carry case Systems, Inc. 6 Terminals - Datamedia 2500 9-10/74 11/74 $ 13,250 NIH elite CRT with edit capabilities Keyboards 3 Keyboards, special, Datamedia 70DVK7019 11/74 12/74 $ 1,156 NIH for leased Datamedia elite 2500 CRT terminals at - NIH Johns Hopkins Univ. Washington Univ. LY MODEL DATE DATE PURCHASE SOURCE ITEM QUANTITY DESCRIPTION MANUFACTURER NUMBER INSTALLED ACCEPTED PRICE FUNDS Modems 16 Auto answer modems Prentice P~113B 5/6/74 5/6/74 Electronics 5 Auto answer modems Prentice P~1200/150 5/6/74 5/6/74 Electronics 5 Originate modems Prentice P-1200/150 5/6/74 5/6/74 Electronics 4 Modem enclosure Prentice P-~100 5/6/74 5/6/74 with loopback Electronics switch and cables 4 Modem enclosures for Prentice P-850 5/6/74 5/6/74 8 modems with cables, Electronics power supply, digital loopback, line loopback, indicator lights 2 Acoustic coupler Prentice DC-22 3/74 3/74 modems Electronics Y Vv 3 Modem enclosure with Prentice P-100 3/74 3/74 $ 9,415 NIH live loopback switch Electronics to house P-103F modems Oscilloscope 1 Oscilloscope Tektronix, Inc. 475DM43 1/75 $ 3,476 NIH BY 49 II.D PUBLICATIONS Publications for the SUMEX staff have included papers describing the SUMEX-AIM resource coincident with its dedication last November (SIGART Newsletter, Sept. 1974; and ACM SIGBIO Meeting, San Diego, November 1974 (oral presentation)). In addition, a substantial effort has gone into developing, upgrading, and extending documentation about the SUMEX-AIM resource, the SUMEX-TENEX system, and the many subsystems available to users. These efforts include a number of major documents (such as SOS, PUB, and TENEX-SAIL manuals) as well as a much larger number of document upgrades, user information and introductory notes, and policy guidelines (see Appendix E, Appendix H, and Appendix I). Publications for individual user projects are summarized in the respective reports and Appendix F, page 167. 50 III RESOURCE FINANCES III.A REFERENCE TO BUDGETARY DETAILS The budgetary materials for the SUMEX project covering past actual costs, current performance, and estimates for the next grant year are submitted in separate document to the NIH. Only that section referenced earlier and describing preliminary plans for SUMEX-AIM CPU up-grading are included here