Memtest86+ V1.30 Released !

[Technoid]

I just released a new version of Memtest86+. This new version is the "phase 1" of many internals architecturals changes on Memtest86+ in order to provide some stronger tests. These changes will be published in three phases (Thanks to Eric Auer for his valuable help).



In this version, I rewrited the "cache-off" function in order to really disable all kind of caching. MTRR are now disabled if previously activated. That can result in a memtest86+ pass two times slower (but stronger) on Pentium 4. I also added a "fast" mode, with caching always on. Faster, but a bit less effective. So, the standard test in Memtest86+ v1.30 is now stronger on CPU with MTRR enabled by BIOS (All Pentium 4). On these CPU, the new "Fast mode" is what we called Standard in Memtest86+ V1.27 or less.



Memtest86+ V1.30 now includes an experimental support for on-fly timings change (See Menu). The idea and base code was provided by Eric & Wee, two talented engineers from a well-known memory maker. Their code was checked, optimized and I added support for i852/855 chipsets. Memory makers contributions are welcome, but memtest86+ will *NEVER* become a vendor-specific tool and all addition will be carrefuly checked before additions. Some guys already tried to offer money to disable or "modify" a test which was accused to produce "too many" errors on their memory modules...



So, no more blah-blah and here is the complete changelog :



** Memtest86+ V1.30 changelog **



- Added on-fly timings change on some chipsets

(Thanks to Eric and Wee for this part)

- Added i852/855 on-fly timing change

- Added support for VIA K8T890

- Added a "Fast" Mode (with MTRR & L1/L2 Caches)

- Correct a bug (Disable MTRR for stronger test)

- Some others bug fixes



Download is available at http://www.memtest.org

Official Forum : http://forum.memtest.org

Please report all bugs and issues in the forum

[Ribbit]

Re: Memtest86+ V1.30 Released !

Originally posted by Technoid
Some guys already tried to offer money to disable or "modify" a test which was accused to produce "too many" errors on their memory modules...

Disgusting, that's all I can say. If I find out who this is, they go straight into my blacklist.

Also, see this forum thread. Apparently Corsair test their memory with ECC turned on, and recommend you do the same if you're getting memory errors. Hmmm....

On a higher note, the on-the-fly timing stuff apparently came from OCZ.