Racket_parallel:improve the performance

[Antigen-1]

I've experimented many times according to the Racket Reference and it runs slightly faster than the implementations of Racket and Racket_compiled now.

---------Results--------	(Diff):	KMinAvg	Min	Avg	Max	Stdev	Cnt
Racket_parallel	(100%):	519.94	512.06	542.92	634.90	32.85	20
Racket	(100%):	544.95	526.47	561.27	599.69	20.97	20
Racket_compiled	(100%):	547.92	537.09	563.61	587.08	17.57	20

(Tested on Racket 8.4[CS], Ubuntu 20.04)

[Antigen-1]

Sorry to bother you again,sir. This is my last commit. I'm just a lover of computer programming and a freshman of Racket,for which the performance of my implementation never exceeds largely that of the Racket implementation even if it runs in parallel. Also, I have no idea whether the program will present worse performance or not when the computation it faces increases in that the usage of swap is always 0 on my machine although it is on.

[lilydjwg]

I get an error (it presents before and after this change):

map: all lists must have same size
  first list length: 473
  other list length: 472
  procedure: #<procedure:swapview.rkt:57:27>
  context...:
   /usr/share/racket/collects/racket/private/map.rkt:257:2: gen-map
   body of "/home/lilydjwg/workspace/swapview-rosetta/Racket_parallel/compiled/swapview_rkt.zo"

You can test against the Rss: field to test if you are not using swap.

[Antigen-1]

I need to use place-channel-put/get to fix the bug, but the cost for communicating with the places can be fairly high.

[Antigen-1]

---------Results--------	(Diff):	KMinAvg	Min	Avg	Max	Stdev	Cnt
Racket_parallel	(100%):	689.73	680.27	700.79	725.76	13.70	20
Racket_compiled	(100%):	696.66	685.11	704.51	727.93	10.45	20
Racket	(100%):	697.58	690.65	704.46	735.48	10.36	20

I find myself having a poor understanding of the place-channel.