Thread View: alt.comp.hardware.pc-homebuilt
8 messages
8 total messages
Started by zz
Tue, 23 Aug 2005 12:53
ram test query
Author: zz
Date: Tue, 23 Aug 2005 12:53
Date: Tue, 23 Aug 2005 12:53
16 lines
1858 bytes
1858 bytes
my usual practise is to test a new sim with Memtest 3.1a if it fails [usually test 5] i return it outright if it is on a friends machine ill try other utils such as windows memory diagnostic [not a microsoft util] or docmemory ram diagnostic 2.1 if it fails all tests the ram is discarded if it fails memtest test 5 and passes all others i will leave the sim in the board and recomend that it be replaced..... am i being too stringent on these tests?? seems like every 3 rd chip fails the memtest test #5 interested to know if you guys bother to test ram at all or what you test with and at what point you replace a chip thanks heaps for any replies........ zz
Re: ram test query
Author: spodosaurus
Date: Tue, 23 Aug 2005 12:58
Date: Tue, 23 Aug 2005 12:58
34 lines
2616 bytes
2616 bytes
zz wrote: > my usual practise is to test a new sim with Memtest 3.1a > if it fails [usually test 5] i return it outright > > if it is on a friends machine ill try other utils such as windows > memory diagnostic [not a microsoft util] or docmemory ram diagnostic > 2.1 if it fails all tests the ram is discarded > > if it fails memtest test 5 and passes all others i will leave the sim > in the board and recomend that it be replaced..... > > am i being too stringent on these tests?? > seems like every 3 rd chip fails the memtest test #5 > interested to know if you guys bother to test ram at all or what you > test with and at what point you replace a chip > thanks heaps for any replies........ > zz I've only ever had four memory modules fail memtest 86. I've tested almost 100 computers with it over the years. I let it run for at least a full pass. In one system it was the motherboard that was bad, not the RAM. I also try and use the latest versions (you're using an older version, but not much older). -- spammage trappage: remove the underscores to reply I'm going to die rather sooner than I'd like. I tried to protect my neighbours from crime, and became the victim of it. Complications in hospital following this resulted in a serious illness. I now need a bone marrow transplant. Many people around the world are waiting for a marrow transplant, too. Please volunteer to be a marrow donor: http://www.abmdr.org.au/ http://www.marrow.org/
Re: ram test query
Author: Bob Day
Date: Tue, 23 Aug 2005 13:13
Date: Tue, 23 Aug 2005 13:13
23 lines
2180 bytes
2180 bytes
"zz" <jhgfjh@hfksdfse.com> wrote in message news:430b1a14.3916181@news-server... > my usual practise is to test a new sim with Memtest 3.1a > if it fails [usually test 5] i return it outright > > if it is on a friends machine ill try other utils such as windows > memory diagnostic [not a microsoft util] or docmemory ram diagnostic > 2.1 if it fails all tests the ram is discarded > > if it fails memtest test 5 and passes all others i will leave the sim > in the board and recomend that it be replaced..... > > am i being too stringent on these tests?? Maybe just a tad. You might try reseating the module in question or putting it into another slot and re-running the tests. Doing that can sometimes solve the problem. One thing for sure, if one of the decent memory diagnostics such as Memtest86 or Memtest86+ fails, there is very definitely a problem that needs to be solved. Personally, I would recommend avoiding non-ecc memory. -- Bob Day http://bobday.vze.com
Re: ram test query
Author: BnH
Date: Tue, 23 Aug 2005 13:27
Date: Tue, 23 Aug 2005 13:27
24 lines
2298 bytes
2298 bytes
DDR or SDRAM ? on DDR they almost never fail .. while on SDRAM on CUV4X-E board, somehow there is alway a fail somewhere .. esp 1x 512 stick. =bob= "zz" <jhgfjh@hfksdfse.com> wrote in message news:430b1a14.3916181@news-server... > my usual practise is to test a new sim with Memtest 3.1a > if it fails [usually test 5] i return it outright > > if it is on a friends machine ill try other utils such as windows > memory diagnostic [not a microsoft util] or docmemory ram diagnostic > 2.1 if it fails all tests the ram is discarded > > if it fails memtest test 5 and passes all others i will leave the sim > in the board and recomend that it be replaced..... > > am i being too stringent on these tests?? > seems like every 3 rd chip fails the memtest test #5 > interested to know if you guys bother to test ram at all or what you > test with and at what point you replace a chip > thanks heaps for any replies........ > zz
Re: ram test query
Author: kony
Date: Tue, 23 Aug 2005 18:23
Date: Tue, 23 Aug 2005 18:23
37 lines
2595 bytes
2595 bytes
On Tue, 23 Aug 2005 12:53:08 GMT, jhgfjh@hfksdfse.com (zz) wrote: >my usual practise is to test a new sim with Memtest 3.1a >if it fails [usually test 5] i return it outright > >if it is on a friends machine ill try other utils such as windows >memory diagnostic [not a microsoft util] or docmemory ram diagnostic >2.1 if it fails all tests the ram is discarded > >if it fails memtest test 5 and passes all others i will leave the sim >in the board and recomend that it be replaced..... > >am i being too stringent on these tests?? >seems like every 3 rd chip fails the memtest test #5 >interested to know if you guys bother to test ram at all or what you >test with and at what point you replace a chip >thanks heaps for any replies........ >zz You might be jumping to the wrong conclusion. With so many modules failing it seems more likely the motherboard is not stable running at certain timings, with modules that only meet JEDEC standards (which is all that is actually required but some modules have more margin than others). Generally speaking, what this means is that no matter what paper-spec, rating, the memory has, the board may need to run any particular module at slightly lower than it's "true", physical capability. It's certainly not too stringent to not allow the combination of memory and board to run when memtest has shown errors, there should be none and whatever-is-needed should be done. In some cases that could mean manuallly setting more conservative timings in the bios.
Re: ram test query
Author: Rod Speed
Date: Tue, 23 Aug 2005 18:42
Date: Tue, 23 Aug 2005 18:42
26 lines
1812 bytes
1812 bytes
zz <jhgfjh@hfksdfse.com> wrote: > my usual practise is to test a new sim with Memtest 3.1a > if it fails [usually test 5] i return it outright > > if it is on a friends machine ill try other utils such as windows > memory diagnostic [not a microsoft util] or docmemory ram diagnostic > 2.1 if it fails all tests the ram is discarded > > if it fails memtest test 5 and passes all others i will leave the sim > in the board and recomend that it be replaced..... > > am i being too stringent on these tests?? > seems like every 3 rd chip fails the memtest test #5 Dont see anything like that myself. > interested to know if you guys bother to test ram at all Yes. > or what you test with and at what point you replace a chip Basically do what you do, if memtest fails, time to use different ram. Not so much decide that its defective ram, just that the motherboard doesnt like that particular ram.
Re: ram test query
Author: zz
Date: Tue, 23 Aug 2005 22:39
Date: Tue, 23 Aug 2005 22:39
48 lines
3166 bytes
3166 bytes
On Tue, 23 Aug 2005 18:23:11 GMT, kony <spam@spam.com> wrote: >On Tue, 23 Aug 2005 12:53:08 GMT, jhgfjh@hfksdfse.com (zz) >wrote: > >>my usual practise is to test a new sim with Memtest 3.1a >>if it fails [usually test 5] i return it outright >> >>if it is on a friends machine ill try other utils such as windows >>memory diagnostic [not a microsoft util] or docmemory ram diagnostic >>2.1 if it fails all tests the ram is discarded >> >>if it fails memtest test 5 and passes all others i will leave the sim >>in the board and recomend that it be replaced..... >> >>am i being too stringent on these tests?? >>seems like every 3 rd chip fails the memtest test #5 >>interested to know if you guys bother to test ram at all or what you >>test with and at what point you replace a chip >>thanks heaps for any replies........ >>zz > > >You might be jumping to the wrong conclusion. With so many >modules failing it seems more likely the motherboard is not >stable running at certain timings, with modules that only >meet JEDEC standards (which is all that is actually required >but some modules have more margin than others). > >Generally speaking, what this means is that no matter what >paper-spec, rating, the memory has, the board may need to >run any particular module at slightly lower than it's >"true", physical capability. > >It's certainly not too stringent to not allow the >combination of memory and board to run when memtest has >shown errors, there should be none and whatever-is-needed >should be done. In some cases that could mean manuallly >setting more conservative timings in the bios. > how should i set the timing? i have the following options in bios dram clock by default is set to: auto by spd i have auto by spd, 100, 133, 166 and 200 dram timing turbo, auto by spd, turbo and ultra thanks zz
Re: ram test query
Author: kony
Date: Tue, 23 Aug 2005 23:17
Date: Tue, 23 Aug 2005 23:17
40 lines
2648 bytes
2648 bytes
On Tue, 23 Aug 2005 22:39:26 GMT, jhgfjh@hfksdfse.com (zz) wrote: <snip> >>...In some cases that could mean manuallly >>setting more conservative timings in the bios. >> >how should i set the timing? You haven't even mentioned the hardware yet, any further answer I provide will therefore be incomplete. >i have the following options in bios >dram clock by default is set to: auto by spd >i have auto by spd, 100, 133, 166 and 200 >dram timing >turbo, auto by spd, turbo and ultra >thanks >zz Since you make no mention of the specific hardware, I write of generalities. Generally you'd not use "auto by SPD", rather manually setting the memory to same speed as FSB, or +33 if the board and memory are still in spec at that speed. One exception is Athlon XP, "usually" it's good to set to same as FSB. Do not use "turbo" or "ultra". Manually set the actual timings. Some menus do not show these settings until you change the prior settings. Determine the spec'd timings for the memory as provided by manufacturer or reseller. Set it to those and test with Memtest again. If that doesn't work, use higher timings. For example, raise CAS another 0.5 tick. Raise voltage if it's at 2.5, up to 2.6 or 2.7. Instead of further speculation it would be best for you to Google some memory tweaking web articles but instead of tweaking for higher performance, you're essentially doing the opposite, tweaking for lower performance, as the slower timings are typically more stable.
Thread Navigation
This is a paginated view of messages in the thread with full content displayed inline.
Messages are displayed in chronological order, with the original post highlighted in green.
Use pagination controls to navigate through all messages in large threads.
Back to All Threads