t0mz0r

np which processor are you having this issue with tho? Just wondering if its an old c0 like my one, which also does it..

Hmm, so far I've tried a few different things that seem to be working - however shortlived they may be: Total power down at reboot, not simply the normal restart of the system Complete board and BIOS shutdown (with battery removed) before trying again Removing one DIMM, setting easier timings, reboot and re-install missing DIMM Raising "vQPI" as well as manual input for Uncore Having a look at DIMM REF voltages at the moment to see if there's something in there affecting this issue.

Hey boys Lately I've been testing a couple high-speed DDR3 kits, and at somewhat random - on a wide variety of X58 boards - one or more memory channels are disappearing, i.e. dropped from both BIOS and effective use. Has this happened to anyone else? At first I noticed this incorrect memory detection behaviour on an Asus Rampage II Extreme when using specific settings for CPU VTT (QPI/Uncore voltage) and memory speed, which led to one of the three channels being dropped, leaving a 3x2 GB kit to boot with only two of the three modules active. The issues were most apparent when using a low QPI/Uncore voltage or too tight timings - or at speeds in excess of 800 MHz (DDR3-1600), anything less ran without issues. It should be noted, that CPU-Z would display all installed 3x2 GB, but testing memory bandwidth reveals that only 4 (2x2 GB) is being used. I tested every possible aspect, even ran checked all modules with MemTest. The board was then RMA'ed, and the replacement board has so far not shown similar issues with +1866 MHz tri-channel kits. Initially I thought this to be a BIOS issue, but nowadays I'm also testing the W35xx series Xeon's which are doing just fine. No problems regardless of selected memory speed and/or tight timings. This got me thinking that loosing memory channels might be related to a partially degraded internal memory controller (IMC) on my old and used 965 ES. So, has anyone else had similar experiences with disappearing memory? And is the missing memory channel caused by the IMC or the need for a combo of ideal memory settings? (as well as finding the optimal QPI voltage of this particular sample) - Tom :ws:

Going along the lines of what gets posted here, that would render the entire server processor results on wprime1024 invalid too? I dont get what the fuzz is about

Rofl

Kein problem =)

login required.. for those without:

looks like 21x258 or summat..

Not really, no - but googling that handle turns up quite a few entries/postings everywhere..

To me it only looks like the same installation, O/S wise. Thats about it? Might be mobo too, but there's a diff. in bios revision - and they were submitted four days apart, so.. If anything this would be called "Software sharing", lol Going on a cheating rant in this case is a bit over the top for me.

+1..

Oh thats only like 1 hour by car for me then..

Ok, I'll bite.. So how do I "hide" a result?

Would be nice to meet ya and cool if Massmate showed up as well! Me neither, still looking into room reservations

Cleaning out my old listings that aren't showing points, and came across a few results I cannot delete - could one of the mods help out, please? http://www.hwbot.org/result.do?resultId=722657 http://www.hwbot.org/result.do?resultId=761713 http://www.hwbot.org/result.do?resultId=761716 Thank you! - Tom :ws:

Just thought it'd be nice to hear who else is thinking about going? :woot: Count down: http://www.cebit.de/bin/cebit_logo_countdown/countdown_logo_en.swf Site URL (in english): http://www.cebit.de/homepage_e?typ=e

Any particular stand you'll be at?

There you go. Sounds a lot better than ..don't ya think?

Job well done turning this into a rant tho.

SeaSonic M12-700 for 24/7 and Silverstone ZM1200 for benches

So basically this says that hwbot is incapable of successfully validating scores to the satisfaction of its top10 users or? I.e. if somebody does an extraordinary result, which the top10 weren't expecting, it now warrants additional proof? Or am I missing something here?

Well, maybe someone has just outgrown the whole kindergarden accusations. I would too.

Great idea, SF3D With all respect to who you are:

npaa

Press that lil blue book on each of your records, m8 "06 Nov 2008 01:40 - insufficient verification for rankings (please use only 1.55 version)" - Tom