Download Load balancing in parallel computers: theory and practice by Cheng-Zhong Xu, Francis C.M. Lau PDF

Posted by

By Cheng-Zhong Xu, Francis C.M. Lau

Load Balancing in Parallel pcs: thought and perform is ready the fundamental software program means of load balancing in allotted reminiscence message-passing parallel desktops, also known as multicomputers. each one processor has its personal tackle house and has to converse with different processors via message passing. normally, a direct, point-to-point interconnection community is used for the communications. Many advertisement parallel pcs are of this type, together with the Intel Paragon, the pondering laptop CM-5, and the IBM SP2. Load Balancing in Parallel pcs: concept and perform offers a complete therapy of the topic utilizing rigorous mathematical analyses and functional implementations. the point of interest is on nearest-neighbor load balancing equipment during which each processor at each step is particular to balancing its workload with its direct neighbours simply. Nearest-neighbor equipment are iterative in nature simply because a world balanced country should be reached via processors' successive neighborhood operations. considering nearest-neighbor tools have a particularly cozy requirement for the unfold of neighborhood load info around the approach, they're versatile when it comes to permitting one to manage the balancing caliber, potent for conserving conversation locality, and will be simply scaled in parallel desktops with a right away verbal exchange community. Load Balancing in Parallel pcs: concept and perform serves as an exceptional reference resource and will be used as a textual content for complicated classes at the topic.

Show description

Read or Download Load balancing in parallel computers: theory and practice PDF

Best design & architecture books

A+ Complete Lab Manual

This new version of the A+ entire Lab handbook has been completely up-to-date to hide the newest CompTIA goals. it is also been revised for less complicated navigation and a tighter healthy with David Groth's best-selling A+ entire examine consultant. Use those assets jointly to achieve the information, abilities, and self belief you must cross the assessments and start a lucrative occupation.

Web 2.0 Architectures : What Entrepreneurs and Information Architects Need to Know

Net 2. zero is extra pervasive than ever, with company analysts and technologists suffering to realize the chance it represents. yet what precisely is internet 2. 0--a advertising time period or technical truth? This attention-grabbing ebook eventually places substance at the back of the phenomenon by means of picking out the center styles of net 2.

High Performance Data Mining: Scaling Algorithms, Applications and Systems

Excessive functionality info Mining: Scaling Algorithms, purposes andSystems brings jointly in a single position very important contributions and updated learn ends up in this fast-paced zone. excessive functionality facts Mining: Scaling Algorithms, functions andSystems serves as an outstanding reference, delivering perception into the most hard learn concerns within the box.

Integrated Circuits for Wireless Communications

"High-frequency built-in circuit layout is a booming region of progress that's pushed not just via the increasing features of underlying circuit applied sciences like CMOS, but additionally via the dramatic elevate in instant communications items that rely on them. built-in CIRCUITS FOR instant COMMUNICATIONS comprises seminal and vintage papers within the box and is the 1st all-in-one source to deal with this more and more very important subject.

Additional resources for Load balancing in parallel computers: theory and practice

Example text

For the structures of chain, ring,mesh and torus, they derived the optimal exchange parameters that would lead to the fastestconvergence [215]. Boillatpresented a version of the dimension exchange method (ca[ledCayley diffusion) in [19],in which he derived the optimal exchange parameters for Cayley graphs. The D E algorithm discussed above is based on a priori edge-coloring of system graphs. The edge-coloring is deterministic in the sense that the colors of incident edges of a node are determined by the system graph.

Suppose A = 1. Then E is a permutation matrix because for all 1 < c < n, Ee is a permutation matrix. Therefore, E has the eigenvalue 1 in modulus with multiplicity N w h e n A = 0 or 1, and hence 7(E) = 1. Thus, limt-,~o E t does not exist or if it exists, does not converge to ~. Consequently, l i m t _ ~ E t = ~ if and only if A E (0, 1). (2) Since limt-~o E t = E--,for any initial workload distribution W °, we have lira W t~ = ~ W ° = b W t--~ 42 3. T H E G D E M E T H O D where b = ~I<~

S ~ a r convergence resulCs ~ ~ e sCatic w o r ~ o a d model were ob[a~ed ~dependenfly b y Boilla¢ ef al. [17, 18]. Moreove5 Bo~a¢ ef al. derived ~ upper b o ~ d O(N ~) for ~ e convergence fac[or of ~ e dff~sion m e ~ o d , where N is ~ e n ~ b e r of processors. ~ a ¢ is, load b a l ~ c ~ g w i ~ ~ e dff~sion m e ~ o d w ~ convergence to ~ eq~bri~ sCa~e~ p o l ~ o ~ a l ~ e . S u b r ~ ~ d S ~ e r s o n derived lower ~ d upper b o n d s on ~ e ~ g ~ e of ~ e dff~sion m e ~ o d ~ arb9 ~ a ~ n e ~ o r ~ [184].

Download PDF sample

Rated 4.70 of 5 – based on 36 votes