


Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Community
Ask the community for help and clear up your study doubts
Discover the best universities in your country according to Docsity users
Free resources
Download our free guides on studying techniques, anxiety management strategies, and thesis advice from Docsity tutors
Questions from the fall 1993 cs 152 midterm ii exam, covering topics such as instruction set architecture, ieee floating point numbers, microcode, vhdl, datapath design, register files, and cache organization. Students are expected to answer questions related to these topics, some of which may require problem-solving and explanation.
Typology: Exams
1 / 4
This page cannot be seen from the preview
Don't miss anything!
You have 120 minutes to complete 125 of the 140 points. You should do 3 of te last 3 problems and indicate which. You may use a calculator or the book if you like, but you will probably do better withou it. The exam is long and hard so pace yourself and think before you write.
Show your work. Write neatly and be well-organized. Good Luck!
page max score 2 25 3 15 4 15 5 25 6 15 7 15 8 15 9 25 TOTAL 125
1.[15]Indicate which of the following are specified in the instruction set architecture (ISA) and which are determined by the machine organization (MO).
a.instruction format b.cycle time c.set of operations d.data types and representations e.power consumption f.programmable storage g.pipeline hazard resolution h.CPI i.MFLOPS j.handling of conditions k.number of branch delay slots l.cache line size m.how memory is addressed n.number of explicit operands o.virtual to physical address translation
2.[5]The normalized single precision IEEE floating point number of smallest magnitude is
Fall 1993 CS 152 Midterm II 1
a.1.02^- b.1.02^- c.-1.02^- d.1.12^-
3.[5]During the post-normalization step of an IEEE floating point addition it may be necesssary to shift the result
a.right b.right or left c.right by one or left by some d.right by some or left by one e.right by one or left by one
4.[5]Briefly explain the difference between horizontal and vertical microcode and indicate the advantages/disadvantages of each.
5.[5]Draw the waveform for y produced by the VHDL statement
y<=a NOR b after 0.5ns
6.[5]Consider a datapath stage containing combinational logic with a critical path of 10ns between two edge-triggered flip-flops with a setup time of 1ns, propagation delay of 2ns, and a hold time of 0ns. The maximum clock skew is 1ns. What is the minimum clock period at which the datapath can be run?
7.[7]The following diagram shows the internal structure of a register file with one read port and one write port. Label each of the components and fill in the missing pieces of the cell detail.
diagram missing
8.[8]Compute the timing and load characteristics of the new cell C shown below from the cells A and B in the table.
Input Load Internal Delay Delay per Load
A 1 5 2
B 2 3 1
Cix
Ciy
diagram missing
Store 10% Branch 20%
The measured bandwidth (instruction and data) at the memory is 36MB/s. What is the CPI? (Show very clearly how you calculate this! Draw yourself a picture. Make sure the units check in each step of the calculation.)
15.[15]We have used a linear model to estimate the performance of many aspects of computer system. A model frequently applied to network is T(n)=T0+n/BW , where T0 is the start-up cost to send a zero length message and n is the message size. You are comparing ethernet and FDDI for a system you are designing. A little table is shown below to help you recall the key performance parameters. Suppose that the start-up cost
EtherNet FDDI Peak BW 10 Mbits/S 100 Mbits/S Max Packet Size 1500 Bytes 4500 Bytes
to send a 1500 byte packet is 1 millisecond on either network. What is the bandwidth achieved on each network sending packets of this size (assuming no contention)?
How much additional bandwidth is achieved on FDDI if a full size packet is sent with this same start-up cost?
Instruction RTL add rd, rs, rt R[rd]<R[rs]+R[rt] sub rd, rs, rt R[rd]<R[rs]-R[rt] lw rt, rs, im16 R[rt]<M[R[rs]+SX(im16)] sw rt, rs, im16 M[R[rs]+SX(im16)<-R[rt] bz rt, rs ifR[rt]==0 then PC<-R[rs]