Igor Nunes igorsnunes

## 0-report.txt
Compiling cmpldi.s and running with "perf stat" the output was "195.508661159 seconds time elapsed".
Doing the same with nocmpldi.s the output was "195.614111789 seconds".

The files were compiled using gcc version 4.9.2 and the following line was used to compile:
gcc cmpldi.s -o cmpldi.o
gcc nocmpldi.s -o nocmpldi.o

Based on these results, no performance gain was detected.

## disclaimer.txt
When running the attached example in x86_64  and powerpc64le, different results are presented.

The attached file calls the function getDisplayName imported from libicu.

It is expected that in both architectures the program crashes in a core dump as pointed in [1] and
[2]. But in powerpc64le the last printed character is 247 and in x86_64 it is 255, i.e. in x86_64
the overflow happens 8 bytes ahead of ppc64le.

In order to compile the below code, the following command line is used:
	Compiling cmpldi.s and running with "perf stat" the output was "195.508661159 seconds time elapsed".
	Doing the same with nocmpldi.s the output was "195.614111789 seconds".

	The files were compiled using gcc version 4.9.2 and the following line was used to compile:
	gcc cmpldi.s -o cmpldi.o
	gcc nocmpldi.s -o nocmpldi.o

	Based on these results, no performance gain was detected.
	When running the attached example in x86_64 and powerpc64le, different results are presented.

	The attached file calls the function getDisplayName imported from libicu.

	It is expected that in both architectures the program crashes in a core dump as pointed in [1] and
	[2]. But in powerpc64le the last printed character is 247 and in x86_64 it is 255, i.e. in x86_64
	the overflow happens 8 bytes ahead of ppc64le.

	In order to compile the below code, the following command line is used: