0x29A

That's exactly right. There were people replacing factory thermal pads on 580Ms in m18xes with ICD and the results were quite stunning. The point is to remember to use a non-conductive non-bleeding and thick enough paste to prevent short circuiting and proper application.

Like hell you did. Right. Unwilling to post results? You just don't have any, do you? Proper knowledge is empirically proven knowledge. You made a claim and you cannot support it with proofs. All you are is just talk. Well... I think we are done here :-/ You are clearly unable to prove your point.

"Indeed" what? There is no IC Diamond featured in that 'roundup'. How can this prove anything? IC Perihelion is a different product. If we are talking ICD, let's talk ICD. Edit: Note that ICP is a "cost effective" product as IC describes it. It's price is much lower than ICD's too. Clearly, it's not the same quality there. ICP is meant for mass, cost effective applications. And how exactly did you observe that? Let's face it, you didn't. You never conducted in-depth tests nor used that many computers in very heavy stress environments. You're just trolling. ICD's performance became already common knowledge and indeed can be proven in many test. The thing is you just decided you know better and keep talking shit. Your 'expert' does not even seem to notice that the digit stands for volume of the tube with paste. 7 or 24 is not the name of the compound itself. If we are talking proper knowledge, let's talk proper knowledge and not just anything found on any blog anywhere. This is the problem with kids of the wikipedia era. You think that if you googled it it must be right. Again, this is common knowledge mobile CPUs are lidless. So what exactly? The fragment you quoted does not prove anything else apart the fact that questionable expert 'recommends' things (well, in the same way like you do). You haven't really proven anything, have you? So far we established that ICD scratches dies' surface, but apart from that nothing really. Are scratches bad and therefore make ICD perform worse than competition? Prove it, properly this time. BTW, I really can't be your "bro". You're troll. I am human. Two different species. Are you sure this place is for you? You seem more like an NBR type to me. ADMIN NOTE: Highlights in red are the only reason we ban people here at T|I. We don't ask much from visitors here except to be respectful towards other members. If you can't do that, this forum is not for you.

Your "core" is supposed to work properly and be cooled so, not look good, GeoCake. Nobody can see it anyway. The goal of the 'race' is obvious: thermal paste that conducts best wins.

Why risk disappointment, bleeding or anything else? Just get IC Diamond. Well known, performs well and won't disappoint.

I cannot say I've had many Logitech mice, since if you get one it last for 3-5 years easily. G500 and G700 at my end at the moment, both perform really well. If I were you I'd go for G500. It features the Logitech classic form factor (which I find very comfortable) plus you can adjust mouse's weight what makes the device very accurate. G500 also offers 2x DPI and it only some 10 € more expensive. Anyhow, Logitech is good brand. I am pretty sure G300 will perform well too if you choose to get that one.

AFAIR, talking mobile GPUs, the point of the whole 40nm -> 28nm switch is first and foremost getting more performance per Watt. This means that indeed a card like 675 may be expected to have similar performance to 580, yet it will require way less power to achieve it, say 60W instead of 100W+ 580m likes. If you look at the specs of desktop line you can easily see that the Wattage requirements remain generally the same but GFLOPS number is at least tripled, so this in syn what Nvidia said some times about getting ~x3 more performance per Watt. However, just as well we may see 680M still running at 100W and reaching 3x the performance of 580M. Who knows. It feels undeniably good to be sporting the best mobile GPU SLI setup on the planet. It feels less funny being aware it will soon be over Hmm... Well, as longs as my 580M SLI runs any game maxed out I am fine with it.

Don't give up, daveh98. Remember that at the other end of the line there is just a corporate drone sitting there. No point arguing. Be patient, if the person you're talking to starts to be annoying and/or inept, simply and patiently ask for another person or perhaps a manager. I've had my share of technical "support" from Dell and apart of the fact most of those people were annoyingly slow thinkers it all ended well. The point is to be patient. In the end Dell offers the best warranty conditions out there.

Very nice, good for both of you!

The heatsink finally came. In total I spent 31 EUR -> the cost of the HS in the US plus shipping to Europe. If I got it from Dell EU as it was originally intended, I'd pay 36 inc. shipping, so in the end I saved myself more Dell's bullshit and those extra 5 EUR.

Was that a rhetoric questions or perhaps nobody really knows what's new in A08? A07 is supposed to limit Sata speeds at 3Gb/s, hence my question is this limitations are also present in A08. Thanks for help.

Still running A05. Does A08 plain/modded also lock SATA speeds at SATA II? My SSD works just fine at 6Gbi/s, so no reason to mess that up.

Great work, svl7!

Finally, the 2-piped HS is on its way. A friend ordered it for me in the States for $18 + another $35 porto to Europe = $53 total -> ~40 EUR, so only 4 EUR more than if ordered from Dell Europe. Case closed \o/

The best policy is to flash the stock clocks anti-throttle vbios where P0 = P1 (that's the AW_GTX580M_06_P1_P0_matched_087.ROM from the packages above) and then do further overclocking using Nvidia Inspector (see here for Nvidia Inspector details). Sound popping is caused by latency problem of the A04 BIOS (that's the proper bios, not vbios). Just downgrade to A03 and you will be fine. Simply flash the old bios. It is safe and it will work. I attached the original A03 bios for you. M18xR1A03.zip