PF06 DishaGidwani CourseOutcome

You might also like

Download as pdf
Download as pdf
You are on page 1of 2
HPC Cource outcomes wri viladedpal fa —————— HPC OMe page write up oN CoOurce OUICOMES | Paratet procecci cing O hac been cheveloped | as eee Lan effective Fechnology in modern computers — to Meet demand for higher perfomance » (ower coct and accurate recuite om reo AuSS appli catio Arotiel Compurorion ‘con be achieved Via __|Mutriproceccors. Moet common chared memory. Multi procece or models are Uni Form _snemory accece, Non- uniform memory Qccece ANA Cache Only Memory architecture re (COMA) it fe Specrat AU NUMA Mode] where all dictributed main Memoriec are | converted to coche memories. A _ distributed memory multicomputer | Concicte “oF Multripre coms puters, KNOWN Ac Nodec, inter-connected by Meccage paceing HetwWork. Each Thode here acts as an autonomous computer ‘having Oo Proceccor, O Vocol Memory and alco i" t ilo devices. In this cose, alt local memo — ee are privote and _Acceccible, In fimo Super compu VI : Trers .' N' number of ‘Proceccors are connected ro Ee a_ control unit And all proceceors “hove | ‘their iduol Memory units. AL rocece Te ea by eq ntee Sane a ea Sceieo a | connect i C! Else cee fom Network. In the | 4 multiple Proceceor track, it te aecummed ince | |) [different threads execure concurrentiy on ~ To _| dite fferent procetcors and |_COMen uty Neore Mhrough shored memoryor Meccage Pacerna tyetems or Mutticomputer track cyetem, Scanned with CamScanner —+— ‘Interconnection network in tHe parollel machine refers trance fer information From any Cource Mode to ‘any decired dectination Sode. Tic Eack should be completed with Ac cmatl latency ac porcible. Network te compriced of linke and cunt ches swhich Helps cend Information They are _compoced of er view pr egram ac _Sollection meccage pa lis compoced of by ¥opotogy, Fouring Agonrhen, | cwitching stroteay and flow control Mechanicm. elemente. Topology tc porrern to connecr (oO “cuitehes. There ore “four typec of paratie\ Progmm _ ming models t{\eluding chared Memory Model, Meccage Paccing Model, hreade Model and dota Parole! Model. Tn shared Memory Model, Progromm F proceccec which use common or chared Variable: In Me : ing Models Everyprocece hac ike Oton _iset (BF data and _ data _trancfer [s done ucing | _[eo- operation between ‘every Procece. . Thread mode! fe defined as, a short. sequence of Tineeruers one, Within oO “process. The re a at frerent, fhreads con be executed 00 fame Proceccor oF ifferent. In data parotiel “model, “dotacet is orgomzed ioto Common structure. It Could be an al le Concurrent Events Are Poca dle in tedoys orray alto. computers due to practice OF Multiprogramming, Multiprocecemg and Parole tictn owing to hah performance computing. Scanned with CamScanner [From source Dode to destination node. A networks |

You might also like