zydirtyfish / blockcachesim

BlockCacheSim

Download & Compile

• Download via the following command.

git clone https://github.com/zydirtyfish/BlockCacheSim.git

• Enter the BlockCacheSim directory, and use command make to compile.

• Use command make run to execute the sample program.

To download the trace file

• The sample program uses the trace file example.csv under the directory trace. You can get more trace file from UMassTraceRepository or SNIA Website. You can match the trace file format by adjusting the trace_type in the configuration file.

Configuration file

• All the configuration information of BlockCacheSim is in the file config, it uses equals sign to connnect the config attributes and their values.
• Comments begin with #

#algorithm type
algorithm_type=0
#cache size
block_num_conf=65536
#block size
block_size_conf=4096
#write strategy
write_algorithm_conf=1

#the name of the log file
log_prefix=./trace/example.csv

The program entry is in the main function of cache-sim.cpp. It can be found that the configuration file is passed in through the command line parameters. That is we should using the following command to execute the program.

./cache-sim.o config

To statistic the basic information of traces

./cache-sim.o config 100

The basic structure of BlockCacheSim

• The initialization funtion init_cache().
• The destruction function destroy_cache().
• The function main() is the entrance to the program which calls the method exec() of the instance of class Run.
• The method exec() calls the method kernel() of the instance of class Algorithm. kernel() is the core method of BlockCacheSim.
• Different cache replacement algorithm classes inherit the parent class Algorithm and implemente the virtual function map_operation() of the parent class. The function kernel() implements different replacement strategies by calling different map_operation() functions.
• This is similar to the implementation of flashcache (an open source block level cache for flash cache). The duty of function map_operation() is similar to the duty of flashcache, and the function kernel() is similar to the device-mapper layer.

Sample output

<---------------------Statistical Reuslts--------------------->
File Name:./trace/example.csv
throughput	14.00GB
unique data	2.18GB
re-access data	11.82GB
read ratio	76.17
----------reuse_dis_cdf----------
4KB	8KB	16KB	32KB	64KB	128KB	256KB	512KB	1MB	2MB	4MB	8MB	16MB	32MB	64MB	128MB	256MB	512MB	1GB	2GB	4GB	8GB	16GB
0.00	0.01	0.02	0.05	0.08	0.09	0.10	0.12	0.15	0.18	0.19	0.20	0.21	0.51	0.81	0.83	0.85	0.87	0.89	0.91	0.93	0.97	1.00	
----------frequency_cdf----------
1	2	4	8	16	32	64	128	256	512	1024	2048	4096	8192	16384	32768	
0.00	0.83	0.91	0.96	0.98	0.98	0.98	0.99	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	

--------------------------------------------------------------------------------
cache_size: 10000blocks	write_policy: write through
--------------------------------------------------------------------------------
name		hit_ratio		ssd_write
--------------------------------------------------------------------------------
LRU		18.69		3601641
--------------------------------------------------------------------------------

--------------------------------------------------------------------------------
cache_size: 10000blocks	write_policy: write through
--------------------------------------------------------------------------------
name		hit_ratio		ssd_write
--------------------------------------------------------------------------------
OPT		63.73		1990351
--------------------------------------------------------------------------------

--------------------------------------------------------------------------------
cache_size: 10000blocks	write_policy: write through
--------------------------------------------------------------------------------
name		hit_ratio		ssd_write
--------------------------------------------------------------------------------
ARC		32.85		3080631
--------------------------------------------------------------------------------

To add algorithms

• First create a replacement algorithm header file x.h according to the lru.h template, and then modify the logic of the function map_operation().
• After modifying the header file x.h, add the line #define X n to the location of the cache algorithm configuration area in the header file cache-sim.h. X is the name of the algorithm and n is a positive integer, but it cannot be repeated with the previously defined integer.
• Finally modify the header file run.h. Firstly it needs to import the header file #include "x.h". Then add an X object (__X * x) as a private variable for the class Run, and then modify the initialization part of function RUN(). Lastly modified the cache operation part of class RUN (exec()).

Update

BlockCacheSim will also implement more algorithms proposed recently, so stay tuned!

About us

Yu Zhang is currently working toward the PhD degree in Wuhan National Laboratory for Optoelectronics and School of Computer Science and Technology, Huazhong University of Science and Technology, Wuhan, China.

Statement

Non-commercial reprint please indicate the author and source. Commercial reprint please contact the author himself.