websmile

On AMD - on AMD it is easy to run gmh at 400c3-3-3 at 2,1v and below as well, Intel is different on low latency. When I tested the legendary cellshock from stunned_guy, which did above 500c3 on amd 32M, I needed 2,2v for 400Mhz 3-3-3-8 on Intel P965 and on P45 even more volts, he was able to run these 434Mhz at less volts 3-3-3-5 on AMD.... Please do not underestimate the influence of chipset and mch vs IMC, old Intel vs old AMD when you judge results

http://abload.de/image.php?img=ballistixgbhz2rffe.jpg The mems you linked, they are rated for 2,2v like all Crucial Ballistix with Micron D9G Fatbodies need looots of volts, I am not sure they help you at 2,1v limit, you need very good binned GMH I think and luck. Without voltage limit the topic would be much easier to solve - mod it

Voltage of 2,1v might spoil your fun, no idea how good MCH is on asrock board. But 667 crucials were also at c3 2x1gb available.

Fixed

I agree on Geniebens proposal, good Micron D9G and also some Elpidas used for example on OCZ 800C3 kits could do this. Team 800C3, 1200 and 1300c6 are very well binned as well as Crucial Ballistix 1GB dualsided sticks, if you are not shy on voltage there are more Micron based kits that can do this.

It is reasonable, if you can test oc you might even ask a bit more - 2400c9 TridentX are often able to run 2800C9-12-12-18 around 1,98-2,1v depending on quality on 1150 at tight subtimings, kits that can do this are searched for and very useful for benchers. 80 US-Dollars shipped out of the box without oc results is price that seems to be OK for me

Hi, up for sale are a couple of CPUs and PSC, all items are working 1. Intel Core I5 4670K boxed batch L315B388 Perfect condition, box and cooler can be part of the deal. VID 0,976v on ImpactVII, the cpu runs extremely hot because of my boxed-like cooler, anyway I managed to run 5ghz 32m 1,35v 4/4. Due to the scaling batch and low VID this cpu seems to be a good option on cold, for normal use you better delid it or use strong watercooling . Good IMC, 2800c9 tight with Samsung and 2667C8 tight psc work easily 2. Intel Core i7 4770K batch L315B407 CPU was delidded and treated with liquid metal, it is now reworked and uses Gelid Extreme as TP. VID on Asrock and Impact 0,976v, on air CPU did 5ghz 4/8 32m 1,21v on Asrock Z97, 5ghz core 5ghz cache at 2800C9 tight 1,20vcore. Tested after apllication of Gelid for function and works flawless 3. 9x2GB X-PSC (GSkill Flare, Ripjaws, ExceleramVX, Dolphin) The bundle contains the following kit: 2x2GB GSkill Flare DDR3-2000 9-9-9-24, 2x2GB GSkill Ripjaws DDR3-1600 7-8-7-24 (2x); 2x2GB Exceleram VX DDR3-2000 C7-10-9-24. On top comes a Dolphin DDR3-2133 C9-10-9-31 stick with 2286 xmp . The first eight sticks do 2400 C7 on Z77 between 1,57-1,72v at single tests(2x1,57v, 1x1,63v, 2x1,65v, 1x1,69v, 1x1,71v and 1x1,72v), on Z97 I tested the flare and the Exceleram for 2600 C8 around 1,88-1,9v and the ripjaw kits at 1,8v 2667 8-13-10-28 twcl6 4-5-4 tertiaries and the secons ripjaw kit for 2600 8-12-8-28 twcl6 4-5-4 1,8v and 2667 8-13-8-28 1,85v 32m. The Dolphin stick is only tested to work. All of the sticks are useful air clockers, but especially the ripjaws could be real interesting on cold because they can take tight subs and beat nearly all 2400C8 and C9 kits for example I know of on voltage at 2667 on air. Prices 1. 245 Euros+shipping now 240 Euros shipped sold 2. sold 3. 180 Euros shipped Preferred payment options are paypal and western union, buyer is responsible for fees. If I shall ship outside of EU, please ask me for shipping quote, on some locations I am not so fond anymore to ship to after I learned that certain forums prohibit sales outside of their country while benchers from these forums loot stocks worldwide, there are enough black hardware holes already ^^ Disclaimer: no returns, no warranty provided by me because I sell as a private person who had hardware and benching as a hobby. On oc ymmv, but I test on air and with non cherry picked components most of the time

Roman, these are 2400c9-11-11, what you offered at Luxx, if we talk about same kits, were crappy 2400c10-12-12^^ - correct me if I am wrong

Already under investigation, ticket was created http://forum.hwbot.org/showthread.php?t=138852

We already discussed the wallpaper idea, it might come in the future

4,8ghz 1,4v air is awful - and 4690K is devils canyon, which on average do better on air and water than Haswell (4670K and 4770K) but have very bad cold scaling as far as I know. On Haswell, there are a few so called "scaling batches, around L310B-L320B rawly sketched, if these do 5ghz at 1,4v on air people told me these have good chance to do 6ghz+. I depends on scaling and coldbug, time for experts to take over^^

If you can prove that result was submitted earlier than this to hwbot, it is invalid for competition I think, but will be hard to do. Time shown at benchmarks is not always reliable, my bench xp show 2009 after each CMOS reset

Strong run

Wow

Of course I can understand your frustration. But if you have problems, ask at forum or with PM for solution - the XTU not showing frequency bug is long solved, newest XTU and win7 for example work easily including showing frequency for me, which was last years problem. There was shown nothing at your XTU, try a new clean win7 sp1 install and use newest XTU version and test again, this should work I have removed results that show 45x100 4500 MHz because XTU showed spike of 4,61ghz as max, there is no way I can tolerate above limit at a low clock, for noone On the GPUPI, you cannot hide bench settings, even if you work for them. And others can´t as well, this is simply not possible because you need settings to check plausibility. If we would not understand that you worked hard for your bench results and LN2 and hardware is costly the GPUPIs would have been removed instantly.

I can delete your GPUPI if you want, I will discuss with head of moderation if we tolerate unsportmanlike conduct like this and bending the rules or not:) - I am working on comps and not on other results, you show an XTU without frequency and temp reading, CPUZ at 4506 at limit of 4503 low clock and start bunnying around after I was generous enough to not remove your gpupis as well? Maybe you should think twice about your attitude and logic http://hwbot.org/submission/2900765

Your XTU screens, all of them, were not showing temps or, what is more important, frequency during XTU benchmark - invalid for the competition. On your GPUPIs, you did hide benchmark settings, we discussed this, and because we are quite generous on tthese competitions at the moment with new benches, I approved the scores once! - if this ever happens again, it will not be tolerated again

Hot result

Here is link to the official GPUPI support thread, driver and open CL are linked there as well https://www.overclockers.at/news/gpupi-international-support-thread

This "tweak" was allowed for HOT qualifier 2014 iirc.... it is a bit late to be shocked now, though I agree on conclusions drawn from what you might have experienced at testing.

Final bump before ebay

Thanks for the clarification, maybe next time you bother to update the rules before I check more than 800 results^^

Upload your fastest one and best one with two screenshots if the issue isn´t clarified before competition is ended. What is really funny, everyone benches this and noone seems to look at the rules, maybe I shouldn´t bother as well

Until the rules are not cleared I can not make further comment - you see rules at hwbot, and on example screenshot the pi calculation is done tab is mandatory. Edit Rules have been cleared, result is unblocked - btw, there are 2,1.1 submissions 3 days old that are using the tab, so that some people use and others don´t is never a hint on which is OK

The problem is that this has to be cleared with head of moderation because someone failed either at tutorial or at verification rules on which picture is clear. Not my mistake but I would like this to get cleared before we have drama again and I am flooded with reported results 5 minutes after comps end