Re: System Panic Makes My Life Easier
Ivan Sergio Borgonovo <[email protected]>
| Newsgroups | gmane.org.user-groups.linux.svlug |
|---|---|
| Organization | http://www.WebThatWorks.it |
| Message-ID | <[email protected]> |
On 07/28/2016 11:14 PM, Steve Litt wrote: > On Thu, 28 Jul 2016 20:58:25 +0000 (UTC) > Daniel Gimpelevich <daniel-R/[email protected]> wrote: > >> On Thu, 28 Jul 2016 13:34:59 -0400, Steve Litt wrote: >>> My prime number generator hits the processor and RAM hard, and also >>> exercises the hard disk. It's pretty easy to compile and run it on a >>> live CD: >>> >>> http://www.troubleshooters.com/codecorn/primenumbers/ >> primenumbers.htm#_Implementation:_Scalability_Through_Paging >>> >>> If you run my prime number generator, watch your processor >>> temperature like a hawk. >> >> It may have hit the processor hard in a desirable way when you wrote >> it, but today, SMP is the norm, and the code above is >> single-threaded, and due to its other resource usage, you can't just >> run multiple instances. > > What if I run four of them, and three of them are piped to /dev/null? Hitting *hard* hardware in the proper way to test it is *hard*. You could get an idea about it looking at what memtest does, and I'm not even sure memtest covers all memory technology. I haven't used memtest in years. eg. logically adjacent bytes/bit may not be in physically adjacent positions on different memory technologies. Let simplify the problem, suppose you'd like to know if the CPU has some thermal issue. You may be tempted to run an "intensive load". Even considering a single core CPU you've functional units in different places, you may do a lot of int computation without touching the FPU or hitting the RAM without hitting the various cache layers. When the CPU get near to its thermal limit somewhere it may throttle down the frequency without you get the chance to test the faulty area. CPU, memory, network cards, video board... are "stateless", testing storage media may be trickier if you care about your data. Testing network stuff may be trickier for other reasons... etc... If I had to test hardware I would, as Rick suggested boot from a live distro, possibly one specialized in testing hardware... and well that's exactly what you find if you google it ;) http://www.inquisitor.ru/about/ If there wasn't any specialized distribution I'd look for testing tools written by people that did much more research than I did and later find the simplest live distro that satisfy most of the dependencies of the tools. HW vendor may have simple tools to test their hardware that may run from usb (possibly Linux or DOS based). I'm not aware of any Windows live madia ;) But I guess a live Windows may crash for different reasons other than hardware problems and it could run much more software than a live Linux distro making diagnosis harder. -- Ivan Sergio Borgonovo http://www.webthatworks.it http://www.borgonovo.net