Linux - HardwareThis forum is for Hardware issues.
Having trouble installing a piece of hardware? Want to know if that peripheral is compatible with Linux?
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
After installing new RAM, i've been having a nightmare making my computer stable. The full story:
I recently installed 1 Gb (2x512Mb) RAM in my abit KD7-G m/b. At first, it had problems seeing more than 512 Mb of it but this I overcame by actually plugging the 2nd module in properly (bad start, i know)...
However, once the bios was finally aware how much memory the machine had, booting windows invariably resulted in restarts/crashes whereas linux (ubuntu hoary) was kind enough to tell me that it was segmentation errors. With the side off the box, i was able to get into linux before programs would randomly start crashing and shutting themselves down.
These problems, however, were fairly sporadic and, occassionally, i was able to get into either OS, if only to check all my data was still fine. The benchmarking software SiSoftSandra in windows was telling me the motherboard was too warm so i've now installed aluminium heat spreaders on the new RAM modules.
This seems to have solved the restart problem and both OS's *seem* stable without random program shutdowns etc. XP is still trying to tell me it keeps recovering from serious system errors, however. I also noticed that 2 of the case fans were sucking air out of the box, so i've now reversed them to blow onto components which will surely help...
However, last night, i tried to watch an avi (or possibly mpeg) in Totem in ubuntu and about 10 mins in, the whole system freezes. Windows, meanwhile, is complaining about bad sound drivers. I've also noticed that the linux messenger program Gaim shuts down randomly, without any kind of notification. it's entirely possible that other programs will do this too, they have in the past. i haven't run the machine for long with the heatspreaders in place though.
I've tried running Memtestx86+ and (admittedly only after 1 pass), the RAM modules seem fine. they also seem happy enough when only 1 module is installed.
Is it likely that, with all the random restarts, i've corrupted the data on my system disc?
cpu: AMD Athlon XP 1500+ (running at 1333 MHz)
ram: generic 'value' ram modules from ebuyer: 2x512Mb PC3200
m/b: abit kd7-g
video: Nvidia Geforce4 64Mb
Hdd: 1x40Gb ata, 1x160Gb sata, 1x80Gb external (usb connection, ata disc)
OS: winxp 'professional', ubuntu hoary hedgehog 5.04
Based on your description the new RAM very well could be defective. As Thoreau suggested, run memtest on your RAM to determine its reliability. Let it run at least 2 or 3 fully cycles. If it fails any test, or even if you still encounter stability problems, return it for a replacement.
At the same time I advise very strongly against the comment to "fry" it and then try to return it - I can virtually guarantee that damage that is deliberately caused will void any warranty. In other words warranties protect you against manufacturing defects, and aren't applicable to vandalism. Put yourself in the shop owner's shoes - if someone bought (for example) a TV from you and then tried to return it after they had thrown a brick through the screen, you would consider that to be a defective unit? -- J.W.
JW may have a point with the static frying and warranty invalidation. In either case, the point is probably moot.
You don't have a warranty on the RAM. The online store that sold it to you probably has a 20-30 day return policy with a 15% restocking fee. I have noticed alot of new generic RAM coming out of asia that has been pure untested crap. Even in Lenova/IBM notebooks. They are saturating the market with this untested stuff.
My suggestion, buy OCZ, Corsair, Mushkin, or even on the cheaper side- Patriot RAM. Patriot has their headquarters in California, so you can grab a throat if they mess you over. Good luck.
Well, regardless of whether or not a warranty exists, I think it would just be unethical to deliberately break something that you wish to return.
In any event, in terms of quality RAM, I personally think that Kingston is the only way to go, due to their 100% testing policy on the products they sell. As they say, "Kingston has always had an unwavering practice of 100% production testing of all finished goods. This means testing every cell on every chip on every module. On a 128MB module that is 1,024 million cells." In other words, if Kingston is selling it, it has been 100% tested.
Other manufacturers may have similar policies (I don't know for sure as I'm committed to always buying Kingston) but at least in my experience I've never had issues with this brand. Good luck with things regardless of your final choice. -- J.W.
i'm not actually in front of my own machine just now and probably won't be for at least a few days.
what i can say though, is that i quite agree that buying something like kingston or corsair would've been an extremely good idea. putting faith in 512Mb ram modules which only cost £25 was a little naive...
i did suspect that it was overheating that might've been the issue, which is why i got the heat spreaders. they've stabilised it a little bit but the fact that it now freezes is a little disconcerting. the mainboard temp is up around 39 degrees C, i think, although that's off the top of my head. is that suspiciously high? the SiSoft benchmarking software i mentioned seems to think it's too warm but it also classes 'too warm' as 50 degrees while showing my temps well below this so i've no idea what's going on there...
Incidentally, it also mentions that large memory (i.e., >=1Gb) should be registered Ram, whereas mine is unregistered. i'm not a complete novice but i don't have a huge amount of experience with this kind of thing so what does that mean?
There is also a bios setting which automatically disables unused RAM slots. The manual for my m/b mentions that using this setting can cause unstability with some devices, so that's my next plan of action.
another thought might be that there's insufficient voltage across the modules. any monitoring software i have tells me it's within ok limits but i'm not actually sure what those limits are or how i can find out. understandably, i've been a little hesitant to start messing about with that just yet...
incidentally, the reason i've not already sent it back is because the modules appear to function fine individually.
I can't really speak to most of your questions since I have no direct experience with them, however, system temps in the high 30's or anywhere in the 40's C are perfectly normal. In my opinion, once you get above mid-50's C, your system is running too hot. Good luck with things -- J.W.
This corsair DDR module has the lowest clock timings(2-2-2-5) and will be very stable under any load. It's the XMS Xpert TwinX1024-3200XL RAM paired module. It was also be faster than anything else out there for your motherboard type. I actually am using that RAM on the box I'm typing from.
Although you can get any with these configuration's:
About the RAM issue, also some OS's squack and bitch when the ram is rediculously high.
I had an old 400Mhz Celeron mainboard, that already had two 128MB chips in it, but when I replaced it with a new 256 Mb chip (just testing the system), it still ran exactly the same thinking it was 128Mb.
Maybe some motherboards have their limits, if you exceed that, they just truncate the amount and work just happy and fine with what they got.
Others might screw up.
Windows ME seems to run flawless on an old Celeron board (usually 256Mb max), but with my newer Athlon/P4 boards with DDR, it screws up and freezes, resets, fatal erros, etc, but then I heard that it can't handle over 512Mb, I simply truncated the amount of ram allowed by windows and it runs fine (dual boot with XP BTW).
I'm not sure what the numbers actually mean, but I imagine it could cause problems if they're not the same???
My one experience with bad RAM, was with the used HP Omnibook LapTop I bought last year, and it turned out that it was the 128Mb add-on that went bad, not the original 64Mb "softie" chips that were the original hardware.
Why good RAM would go bad is beyond me (save for overclocking, etc), but I had to pay an @$$-load for mine on my new dual-opteron workstation, because it is 4 1-Gig units of 3-3-3 REG ECC RAM (mission crit. server-grade components), and recommended by the motherboard manufacturer.
High-end systems are much more picky when it comes to that then desktops.
Can't go wrong there :-)
Last edited by I_M_Tired_Of_XP; 04-26-2005 at 08:47 PM.
At the time of writing this post OCZ site seem to be down.
Never buy generic memory to save money. Always buy name brand memory.
Steps to install memory:
1) Note the memory slots from the manual and the motherboard.
2) Turn off power supply or disconnect it from the AC outlet.
3) Wait about 30 secs to a minute, so stand-by current can discharge (never skip this step)
4) Insert or replace memory modules in their designated slots. Do not FORCE!
5) Double check all memory modules are correctly inserted by pressing firmly.
6) Check all cables just it case they popped out when you installed the memory modules.
7) Connect the power cable or flip the power supply switch to on.
8) New installed memory should then work.
BTW, I recommend using ECC memory instead of non-ECC memory.