The Intel Processor Diagnostic Tool for Windows* - User Guide

The Intel® Processor Diagnostic Tool - Windows® User Guide R7
Page 23 of 26
Test Library
Cache Test
Description
The purpose of CPU Cache test library is to verify the size of the L1, L2, and L3
Cache, whichever is present. The CPUID instruction is used to read ECX
register. The L1, L2, & L3 Cache size is in 1024-byte units.
The test retrieves the L1, L2, and L3 Cache size information and compare with
entry in configuration file for CPU unit under test.
The cache is tested using one either presence test mode or verification mode.
Which mode is used is determined by the SkipConfig parameter in the
IPDTconfig.xml configuration file. If the value of SkipConfig parameter is set to
“Yes”, then the test library will simply report the cache size.
If the value of SkipConfig parameter is set to “No” then the test library will verify
the L2 or L3 cache size detected against the expected cache size using the
L2Cache or L3Cache parameter found in the CPU node list section of the
configuration file.
Options
CacheTest="Enable", CacheTest="Disable"; SkipConfig="Yes", SkipConfig="No"
Parameters
N/A
Default Settings
CacheTest="Enable"; SkipConfig="Yes"
Test Library
Prime Number Generation Test
Description
The purpose of this test library is to test how fast the CPU can search for Prime
numbers, reported as operations per second. Prime number generation will be
based on the Sieve of Eratosthenes algorithm. The test program will test for
unsigned integer type and within the numbers of 0-2³² for 32 bit platform.
Random comparison will be done to make sure the prime number generated is a
valid prime number.
Options
PrimeNumber="Enable", PrimeNumber="Disable"
Parameters
Time in seconds
Default Settings
PrimeNumber="Enable"; PrimeNumberTimer="2"
Test Library
Floating Point Test
Description
The purpose of this test is to perform addition, subtraction, multiplication and
division, to test how fast the CPU can perform mathematical floating point
numbers operations, and do a sum comparison to make sure the floating point is
giving the correct sum value at the end of each cycle.
The test program will display Million Floating Point Operations per Second
(MFLOPS), time started, cycle completed and quantity of errors detected.
Options
FloatingPoint="Enable", FloatingPoint="Disable"
Parameters
Time in seconds
Default Settings
FloatingPoint="Enable"; FloatingPointTimer="2"