Benchmarks
Normally
after selecting a menu item, the corresponding analysis will be automatically
performed at once. This is different for all drive benchmarks, that require
the user to select the devices to be examined first. A checklistbox allows to
mark the desired drives. To run a benchmark please click the "Run benchmark".
Benchmark
results will be inserted both in the table grid and in the diagrams - colored
differently compared to the values that existed before.
Use the grids checkbox column (with chart icon title) to insert/delete results
into/from charts.
To save your own results click the button "Save own results".
Processor,
Memory
Measures the performance of processor, memory and chipset. Hard- and Softstones
are overall performance index values, indicating the integer/memory resp. the
floating point performance.
Multithread
Benchmark
Measures the performance of multi core processor systems. (However it works
also for single coreprocessor systems). It consists of 4 different threads that
are executed "simultaneously". That means that up to four cores are
currently supported by this benchmark.
The resulting increase of performance is indicated in the field "SMP Factor"
that shows the relationship between execution on all and execution on only one
processor. In order to get this relationship the system will be forced to run
the benchmark first on all available and afterwards only one thread.
Note that the results are
partially and intentionally influenced by graphical output performance. For
authentic results it is necessary to consider GDI performance because even on
multiprocessor systems there is only one GDI available!
The field "CPUs used by OS" indicates the number of cores the operating
system relies on.
This benchmark is an application by it´s own and may (and should) be executed outside of Dr. Hardware too. To get most reliable results any overhead should be avoided.
Hint: You may study the Processor counters of the system monitor to find out more about the usage of the installed processors of your system.
Video
Measures the 2-D video performance via GDI API calls.
Harddisk
Drives
Measures the performance of harddisk drives via API32 calls.
There is a checklistbox where you may select the drives you want to be benchmarked
before running the tests.
The tests include both read and write operations. If possible any buffering
is disabled during these operations so that the results should report the native
drive performance rather than virtual cache efficiency of the operating system.
However the disabling of buffering seems to fail sometimes and irrational results
come out of the tests.
CD
ROM-/DVD Drives
Measures the performance of optical drives like CD ROM, DVD or Blue Ray Disc
drives.
There is a checklistbox where you may select the drives you want to be benchmarked
before running the tests.
The benchmark includes both low level and API level performance testing.
Access times and 1/3 or full stroke are measured through direct sector reading
operations. API calls provide measurement of the filebased data transfer rate,
provided a file of sufficient size (about 2 Meg) can be found on the CD ROM.
Important: A data CD/DVD
ROM is required. If the benchmark fails use another disk instead. Note that
the disks minimum capacity is 100 MB while 600 MB is recommended.
If a native CD writer without reading capabilities is installed in your system
please uncheck it before executing the test because it cannot be distinguished
from a normal CD ROM drive and test may fail.
Also uncheck any virtual CD ROM drive installed by software. They too cannot
be distinguished from a normal CD ROM drive and "accessing" them may
cause the system to hang.
SCSI/ASPI
Drives
Measures the performance of ASPI/SCSI/Atapi drives.
There is a checklistbox where you may select the drives you want to be benchmarked
before running the tests.
The tests are based on read accesses of different block size. On a lot of systems
the maximum block size is 64 KB.
The ASPI benchmark is based on true SCSI commands so the results may be more
reliable for native SCSI devices than the results from the harddrives benchmark
tests.
If a CD writer is installed in your system please uncheck it before executing
the test because it cannot be distinguished from a normal CD ROM drive and test
may fail.
Net
Drives
Measures the performance of remote drives via API calls.
There is a checklistbox where you may select the drives you want to be benchmarked
before running the tests.
The tests include both read and write operations. If possible any buffering
is disabled during these operations so that the results should report the native
drive performance rather than virtual cache efficiency of the operating system.
Important: you must have
writing operation rights for the tested remote drives otherwise the test will
fail.
Hints:
1. The Internet Benchmark is available only in the registered version.
2. This test is optimized for modern DSL technolgy. For Modems, please enable Option <Internet-Benchmark: Use small file> in the Program Settings, which reduces the test duration.
Measures
the performance of your internet connection via HTTP transfers (to and from
Dr. Hardware server www.drhardware.de) and a Ping Echo Request.
You may also define different URLs as a test basis though (Program Settings).