BIT config and test time

[WaRRioR]

Hi,
Do you recommends special settings in BIT ?

specially:
"Test configuration and duty cycles"

and how long do i have to run the test?

We are Systembuilders and need this test to go for shure that the computerhardware
and config runs stabil when our customers pick up thier computer.

Normally we test our computers with 2 Test (benchmarks) at the same time. We test them with 3dmark 2000-2003 and run the BIT in the background to increase the system requirements. Is that good or ist it better to run the BIT standallone?

[David (PassMark)]

How long should I run BurnInTest for?

Not an easy question. In our opinion, the chances or finding a problem in the first hour are relatively high, (the system gets hot, it's the first run across the disk / CD and the first use of some of the drivers). Then every hour after that, the chance of finding a hardware problem drops significantly. The extra benefit of doing 12hours compared to 6hours is thus probably not great. Other nice technique is temperature cycling. All major manufacturers of electronic equipment do this, they have large ovens and fridges in which they test equipment. The expansion and contraction of components and solder joins brings to light many problems. You could do 6 hours On, 6 hours Off, then 6 hours On, to get some limited temperature variation like this.

As for the duty cycle settings. The disk and CD test take the longest so put them on 100%. Then move up the percentages on the other tests that you want to run until the CPU is fully loaded pretty much all the time. This might take a little bit of experimentation and the levels with be different on different machines.

---
David
PassMark

[WaRRioR]

Hm and what is with the 2 benchmarks? Is this good or not so good?

[David (PassMark)]

There is no doubt that running benchmarks puts a lot of load on the machine, put I see several issues with this.

1/ Benchmarks are written to run as fast as possible and achieve maximum data throughput. This comes at the expense of checking the correctness of data. The disk test is an easy to understand example. In BurnInTest the data read from the disk is checked byte by byte to ensure each byte has the correct value. This checking slows down the disk test. A disk benchmark will not do this check. BurnInTest aims to test your computer is working correctly. Benchmarks measure how fast things can be done.

2/ If you keep test records, BurnInTest will not record the activities of other applications that are running. If you don’t keep records this doesn’t matter.

3/ You don’t have much control over the type of load with benchmarks. For example running 3D Mark will obviously load the CPU and Video card. But this will almost definitely slow down the disk test because it is getting less CPU time.

(Note that we don't classify BurnInTest as a benchmark application, we have another product called PerformanceTest)

-----
David
PassMark