[ Date Index ] [ Thread Index ] [ <= Previous by date / thread ] [ Next by date / thread => ]
On 01/07/17 19:29, Adrian Midgley wrote: > It went flaky on Jessie, alas. I'd definitely memtest it heavily then, and keep an eye on temperatures via lm-sensors or similar to see if your system has started running really hot [80-100C under sustained load is normally pretty bad news] for some reason. Perhaps the heatsink has slipped on its mounting and isn't flush against the CPU or the thermal paste has 'evaporated' over time? Then it's purely down to the usual money/time issue - presumably it's not worth hours of your life fiddling with a broken machine when you should be doctoring instead (unless you want to just for fun, but that's different). In your situation I'd be minded to try the obvious and quick fault-testing scenarios first and would then most likely demote the PC to my spare/testing/fixer-upper pile and simply treat myself to a brand new one. Come back and fix it later during that mystical period of "free time" if it ever comes. When something critical I need for work breaks I usually take it as an excuse to buy something fun and bang up to date - as you said earlier, bleeding edge isn't strictly necessary but I do really like to at least keep up the with the big picture. I don't know how much firepower you may need for computational loads in your line of work but I'm personally eyeing up the latest Ryzen Pro and Intel Core i9 based systems (x299 + x399 platforms) currently: Threadripper is just around the corner though. Cheers -- The Mailing List for the Devon & Cornwall LUG https://mailman.dclug.org.uk/listinfo/list FAQ: http://www.dcglug.org.uk/listfaq