LAB3

3.1.1
a) Worst-case delay : 1780ps
b) Estimated area : 4564

3.1.2
a) WCD : 989ps -> +11ps slack
	estimated area : 6074
	
b) The RCA is compiled with a medium level of optimization. This results in more complex structures such as multiplexers being "converted" into simpler compound gates such as AOI:s of various, relatively high drive strength. 

The Sklansky meets the timing requirements at a low level of optimization. At medium effort a slack of 0ps was obtained. The critical path followed : The opcode DFF -> add_sub_module -> mux21 (conditional invert B) -> sklansky -> progresses through dotops levels 1-5 -> gdotops -> E -> sum -> slt_mux -> output mux -> output DFF

Simpler gates than with lower level of opt. compound gates, half adders (HA1), buffers, various flipflops, various muxes

c) From worst to best
Sklansky adder
shift right 
shift left

d) 
Estimated area : 4842

e)


3.3.1

b)
def: togg = how many times a signal toggles during 1ns. togg fuses information on f and Ai as defined by P_sw=f*vdd^2*sum(A_i)*C_i

togg = (1780e−12^−1Hz/1GHz)×(0.01×2)     @ nominal f=1780ps^-1	

Total Power: @1780ps
RCA ALU :  839321.767 nW
	- comp_rca : 85.235  54699.372  54784.607 
SKL ALU: 816692.263 nW
	- comp_sklansky : 48.226  42797.935  42846.161 

c) Changing the activity factor Ai from 0.01 to 0.25
Total Power: @1780ps | Ai = 0.25 
RCA ALU :  418.138 5824708.937 5825127.075nW
	- comp_rca : 108.172 1230840.640 1230948.812  
SKL ALU: 351.938 5296036.869 5296388.806 
	- comp_sklansky : 48.099  784948.366  784996.465 


d) GO BACK AND DO!

e)
====== RCA =======
Exactly the same as sklansky!


===== SKLANSKY =======
     Net         Net            Toggle             
(asserted *) Power (nW) Prob. Rate (/ns) Cap. (nF) 
---------------------------------------------------
clk           81303.371 0.500      1.124   100.500 

And using our normal expression : 

P = 561.8Mhz*1.2V^2*100.5nF = 81uW

3.3.2
a)
Total Power: @1780ps @ Ai=0.01
RCA ALU :   765187.300 4758624.095 5523811.395 
	- comp_rca : 91346.940  576872.429  668219.369 

Increased power dissipation by factor of 10(x10)

c) 
Leakage increased x1000
Dynamic power x10

3.4
Using the generated VCD file a new power report was generated:
ALU                              1260   361.941 4267632.778 4267994.719 
    comp_rca_generic              160    83.364  629229.226  629312.590 


Static power dissipation is more or less constant,  whereas dynamic power has increased by a factor x10.



Using the SAIF file
ALU                              1260   361.949 4264120.338 4264482.288 
    comp_rca_generic              160    83.364  629134.932  629218.295 
	
The VCD and SAIF file formats give almost identical results.

The file formats generate file sizes wit approximately a x10 difference. 5.6M for VCD vs. 788k for SAIF.

      "clk" : "0.50000 1290323000";