For all features requests or Bug fixes for Memtest86+ (www.memtest.org). NO HELP FOR HARDWARE ISSUE !
To help getting memtest recognising more chipsets i made this thread, so that people could leave a cpu-z dump everytime they found a chipset that memtest wasn't detecting.
Original Thread by Sven Bent.
Update : Official Memtest86+ USB Installer released
Select Download - Auto-installer for USB Key from this link
Run boot Memtest86+ from USB flash drive
Useful for Notebook they don't have DVD drive to boot.
Backup USB flash drive before run installer
After More than 2 years since 4.20, 5.01 is finally out and the development of Memtest86+ is now fully resumed. Regular updates will follow soon. We still have some issues to solve with embedded menu items and we must add support for IVB-E, HSW-E & BDW.
Tips : F1 will enable fail-safe mode, F2 will enable SMP Mode (disabled by default) and F3 will trigger an hidden alternative SMP Mode.
Memtest86+ V5.01 changelog
- Added support for up to 2 TB of RAM...
A simple primer on how to get Memtest86+ to boot from a USB Drive or Hard Drive (without having to install an entire distro or DOS).
note: drive used here is a usb drive located on /dev/sda, so make changes accordingly (eg a hard drive install would use /dev/hdc instead)
First Step - format the drive
optional, only suggested if you're drive needs a good cleaning
(this may take a long time given the size of the drive)
dd if=/dev/zero of=/dev/sda Second Step - create FAT partitions...
Memtest86+ FAQ v1.21
By R.R. van der Vossen (Wichetael)
Latest update 2007-02-28
- What is memtest86+, what do I use it for?
I extended my RAM from 8GB to 16GB.
My board has 4 slots and I just bought 2 of the exact same RAM-pieces and out them in the remaining slots.
So there shouldn't be anything I could do wrong right?
It seemed to be working, and I started playing The Witcher 3, but 30min later I had a bluescreen (00000050)
(this reminded me of the time I built my pc and the origin of ongoing bluescreens was that I used the wrong RAM slots)
Has anyone else have problems with the test software, while using xeon 4core cpu's
and more than 4 gb ram ????
6 gb locks up after 3 minutes running test 7 (block mode)
8 gb locks up after same
We have run previous versions of Memtest, while being able to use in AHCI mode. Is there anyway, to run in AHCI mode rather than Enhanced. We have run it in our own home built Linux kernel. Although, testing would not work at all when in AHCI mode. I am using an Intel S2600WT2 motherboard.
First and foremost, you need to edit FAQ:: Please Read before posting of 13/09/2008 sticky because http://forum.memtest.org/ forwards to this forum.
Not realizing I needed a USB flash drive to use memtest86+, I ran the installer and upon seeing the prompt to point to a USB drive, I canceled the operation.
Today upon booting my computer, it stopped at the boot loader manager and displayed an option under tools: Memory test.
Since I cannot use uninstall because memtest86+ was not...
When running memtest86+ 5.01 on a new R630 (G13) Dell server with forced SMP mode (F2) all tests pass besides 32 errors in Test 7 (Block Move).
This looks more like a problem with the (experimental) SMP feature than with the machine (since all other tests pass, and the test 7 also passes in single core mode and the Dell pre-boot system assessment tool in throughough mode does not show DRAM problems).
See attached screenshot of the failures (it also shows that the chipset is unknown, I...
I signed up just to ask this question. I've tried Googling it, but haven't found an answer. I've read release notes and their FAQ, but no answer.
What does "failsafe mode" actually do?
I'm running memtest on a PC that sporadically freezes and when I didn't choose an option (I just left it to start by itself), it come back with 10,000 errors in the first 2 minutes.
I then did failsafe mode, which comes up saying "No detection, same reliability" and it got no errors at all...
Hi, I have a number of new (to me) servers, and I'd like to memtest all of those.
I have Memtest86+ 5.01 booting OK from PXE, but the standard single-processor mode is very slow compared to the SMP/parallel mode. I'd also like to use the SMP/parallel mode to check power and CPU temp are OK in all the servers.
Can I add something to PXE config to automatically enable SMP/parallel mode?
Or are there some special images that have SMP/parallel mode as default?
I am getting quite alot of BSOD like PFSN_LIST_CORRUPT and MEMORY_MANANGEMENT. When I run Memtest86+ there are no errors other than the hammer test. I get maybe like 12 or 18 errors from that test alone. So could those errors be a reason I am getting bsod?
In version 5.* and 4.*
I believe the problem is in smp.c.
I changed MAX_CPUS in smp.h to 1 to disable SMP, then memtest86 works fine.
otherwise the PC will reboot when running into somewhere of smp.c
I just don't know where is the problem.
The CPU is L5408, on a 775 board, it runs Linux and Windows 7 without problem.
With one motherboard I have no throughput displayed by Memtest86+.
2 x AMD 6272 (2 x 16 cores)
2 x 8 Gib ECC memory PC3-12800
Tested with Memtest86+ v5.01 and v4.20 (v4.20 incorrectly reported memory timmings).
Tested with F2 at boot, and without.
I try to perform some memory test on the Mohon Peak CRB with memtest86+.
I downloaded memtest86+ v5.01 and compile it. So far, so good.
Following the thread: http://www.coreboot.org/pipermail/coreboot/2015-January/079143.html concerning this topic, I modified the given values, so that now:
# readelf -e memtest1modified
Magic: 7f 45 4c 46 01 01 01 00 00 00 00 00 00 00 00 00
what is the current status/plans of ECC detection&support? I know the older versions DID show the info. The v5.01 DOES NOT show any info about it. So I assume, the suport was removed. Am I right?
When you select the Last (11th) test, let it run for a little while and then try to skip it. Errors occur.
I've tried this on two different machines, running 2 different CPU's, motherboards (only gigabyte though) and memory types (DDRII & DDRIII).
Can anybody else confirm this ?
I have a Supermicro X9DAi MB (which Memtest+ recognized) which has 2 Xeon e5-2670s (16 cores each). I can't seem to enable multicore (I hit F2 when Memtest starts).
How can I enable multicore?
Thanks in advance.
I've tried the 4.2 version on Windows 8. It doesn't even notice Memtest86+. So I changed the order CD first HDD second in the EFI Boot Manager and still Memtest86+ is skipped and Windows 8 carries on to the Desktop.
Only when I resort to "converting" to a virtual CSM to mimic the BIOS Boot Manager can I get Memtest86+ to run.
So my question is how difficult would it be to have Memtest86+ determine for itself what Boot Manager to boot up with? Would make life a little less painful...
The following articles shows a problem in numerous recent DRAMs (more than 3/4 of DIMMs ?). Repeated reading of two rows may induce errors in other rows. The problem is usually prevented by CPU caches, but some programs do cause the issue (at least: using the flush instruction for this purpose, perhaps others involving thread synchronization)
Memtest86+ V5.00b6 changelog
- Added support for up to 2 TB of RAM on X64 CPUs
- Added experimental SMT support up to 32 cores
- Added complete detection for memory controllers.
- Added Motherboard Manufacturer & Model reporting
- Added CPU temperature reporting
- Added enhanced Fail Safe Mode (Press F1 at startup)
- Added support for Intel "Sandy Bridge-E" CPUs
i have errors on same module ddr2,
first motherboard - errors (read spd information -OK)
second motherboard - no errors (can not read spd information)
this is my first time using Memtest86 and I have the following questions:
1. If I boot it from a USB drive, I get a message saying that this will erase all data on the selected drive and possibly result in loss of capacity. I don't know what this message means because I have never done this before, and I don't want to take the chance that the USB drive will become unusable and never work properly again.
2. I have Windows 8.1 and I am considering the option of booting from a disk on my disk...
I am trying to check out a new server:
Supermicro H8QG6F with 64 cores (AMD) and 512GB of RAM (32*16GB). Running 5.01
The board has 4*AMD 6376 processors.
Memtests brings up its blue screen, beeps once, and then prints the SPD info for 8 slots. No tests start, and the keyboard is nonresponsive.
Am I running up against a limit of Memtest+, or is this some sort of hardware error?
I'll probably pull a lot of the RAM and try testing smaller batches.
I came across with this issue yesterday:
0. Start Memtest86+ V5.01 (27/09/2013), with the Auto-installer for USB Key (Win 9x/2k/xp/7)
1. Spam the 'c' key to get into configure menu immediately (tests are 0% complete at this point)
2. Press '1' '1' '0' keys for select tests, default tests, cancel
3. Memtest instantly jumps to test #2 and displays a never ending list of errors.
hi i ran memtest after getting BSOD randomly every now and then.Is 3 errors bad?I have 16GB of ram in 4 sticks and im just wondering if anyone can see from this log what might help find the bad ram quicker as i have limited time to run memtest,otherwise it will take a few days once i have time to do it.
The only error in this log i see is down near the end and it says test 10.should i just put each stick in one at a time and only run test 10 or do you have to run the full thing with 4...
Greetings, I wanted to ask what is the reason to get the same error (exactly the same failing address) when I ran the memtest86+ and I have switched every RAM into different DIMM slot. Is it possible for the error to be because of the big amount of memory (256G) ?
I've experienced several problems with memtest86+, in particular booting fails (the PC reboots when selecting memtest.bin during boot using grub 0.97). Looking into the code, I found what I think is a problem concerning stack setup. I attach a patch (that I had to rename as .txt to be able to attach) that fixes the issue for me, and I describe this issue better below.
In setup.S, the stack setup seems really weird to me. Register SP is initialized as DX, but DX has no particular value...
Hi there, I built a new PC 3 weeks ago, and I have been experience weird ass random crashes as well as clock_watchdog_timeout blue screens.
I ran memtest86 today and got the following result:
All these errors happened during the hammer test, I stoped the test because it basically stuck there for about 3 hours.
Should I change my memory sticks?
I have a 32 gb Corsair 2133 mhz kit on a Asus Rampage IV Black edition and I wanted to test it. Turns out that if I run memtest86+ using all the cores, the memory is not detected correctly (frequencywise) and the computer will shut down after a few minutes. If I run it in fail safe mode, everything will run fine and complete a pass without errors. Why could this be happening? Should I be concerned? Thank you
Please can someone tell me how to output/save the test results of memtest86+ version 4.2 and 5.01. Many thanks Edit - I've just seen the sticky so apologies for the question/request doh!
I used Memtest now a few times on my Xeron E5-2620 on a Gigabyte GA-X79S-UP5-Wifi and sometimes Memtest starts with a black interface wich looks more graphical and sometimes in a Blue interface.
It seems like some Bios settings are the reason, which interface starts, but I couldn't figure out, which.
When does it start in which mode and how do you call those two different modes in the Memtest interface?
When I did a test on 3 4GB modules (12GB, DDR3, 1333), the first pass took about 1:45 to finish (using the F2 SMT mode.) The second pass, however, took about 8:15 to finally finish. No errors were found, but this seems like an incredibly long time for a single pass of only 12GB of 1333 RAM. Is this normal?
The vast majority of the time was spent on Test #8. I believe that the rest of the tests took less than an hour, while Test #8 took the rest. On top of this, the percentages counters...
So you ran memtest, found some bad bits, but don't feel like throwing away a 4GB stick? Here's how to tell Windows to ignore those bits.
I'm paraphrasing some of the information from this thread:
First run memtest, and write down the memory addresses that it says are bad. For example, I had a test return that 0x0020B9BF07C was bad.
There is a Bad Memory List feature in windows, but it uses Page numbers instead of absolute addresses. Pages are usually 4096 bytes in size. To turn...
Use this control to limit the display of threads to those newer than the specified time frame.
Allows you to choose the data by which the thread list will be sorted.
Order threads in...
Note: when sorting by date, 'descending order' will show the newest results first.