I.nfraR.ed

Yes, sure - 1242PNS. It's just a matter of time someone with big funds recognizes this category as a target. This is just my 5th cpu for the past 2-3 years.

4760MHz 1.5V boot from bios with Megahalems put on top. Never tried max validation on air. 6800+ possible on LN2, but I was stuck at "cpu overclock fail" after reboot and non-detectable 3 keyboards on any port at the end of the last session.

Got it working on my smartphone. First I tried Debian, but couldn't solve the java command symlink. Although there was a symlink in /usr/bin, it was still reporting that there's not. So I've installed the Raspbian. Here's my first run: http://hwbot.org/submission/2412106_i.nfrar.ed_hwbot_prime_4x_arm1176_s_1360.94_primes_per_second Phone is Lenovo A820 with Quad-core ARMv7 MT6589 chipset + SGX544MP video. What is the correct category to submit? Unfortunately MTK-based phones can't be overclocked yet, so I'm stuck at 1.2GHz. BTW, it's using 2 threads only, don't know why - maybe it's because I'm running it chrooted (as a slave OS). root@infrared:/# java -jar hwbotprime.jar java -jar hwbotprime.jar No native cpu speed library for linux yet... falling back to OS tools. --------- HWBOT Prime 0.8.3 ---------- Processor detected: ARMv7 Processor rev 2 (v7l) MT6589 Estimating speed... 2x 1,209MHz @ 37.347 C 976 MB memory Running benchmark using 2 threads. Starting benchmark... Warm up phase: ..................................................................................................... done! Benchmark phase: ..................................................................................................... done! All done! Current CPU temperature: 37.971 C Score: 1360.94. Hit enter to compare online, enter a filename to save to file, or type q to quit.

Опа

Do you have load line callibration enabled? If not - enable LLC for CPU and disable LLC for CPU-NB, these worked best for me when I was using that board. You can't compare any air cpu cooler with DI no matter how great it is. Dry ice is -78°C. Even 1.65V is too much for this cpu on air, better stay in the 1.55V range.

All 1090T have unlocked multiplier and that's what BE (Black Edition) means, so yes - your CPU is BE. 1.74V on air is very high for these. You're basically running in the dry ice territory. I have M4A89GTD Pro/USB3 at home and as far as I remember, voltage readings were close to what is set in bios. This might have changed in latest bioses, I don't know.

If only we had Asrock Z87 OC boards here...

I get a warning in bios (red color), sometimes even a message, but I always disable those kind of things and can't remember clearly

I've always wanted to ask, please don't blame me if my question is stupid, because I've never ran XTU. What happens if a user chooses to try Andre's settings or any kind of dangerous settings on his aircooled Haswell? Does he get a warning of any kind? If I click on the button it just downloads me a profile, but if I had XTU with the registered .xtu extension I might be able to open it directly in XTU and apply the profile?

Since the final product is already in production, I can show you the very early prototype developed in my secret laboratory long time ago . EK Dominator plate + polystyrene

Don't think it will make such a difference. Perhaps try thq second module, it might be better. And definitely learn how to use pscheck, if you still don't.

Still no ROG boards in Bulgaria. There was Gene for a while, but no other models even listed in the sellers' websites Seems like I will have to import it from another country without warranty and higher price it I want a ROG board. No ASrock OCF, too. There are some Gigabyte boards only and nothing else...

Chilled or normal water? I can boot ~4.8 with 1.55V air on mine, but haven't tried max validation.

CL4 is a bug, you're probably running something like CL8, but it depends on the board. Minimum CAS for DDR3 is 5.

I get lockups and sometimes error on first start.

1.55: OK - http://hwbot.org/submission/1060504_i.nfrar.ed_cpu_frequency_athlon_xp_2500_barton_2767.85_mhz 1.57: NOK - http://hwbot.org/submission/2185244_i.nfrar.ed_cpu_frequency_athlon_xp_m_2800__3094.62_mhz 1.58: NOK - http://hwbot.org/submission/2220806_i.nfrar.ed_cpu_frequency_duron_1.6ghz_2750.85_mhz 1.60: NOK - http://hwbot.org/submission/2295507_i.nfrar.ed_cpu_frequency_athlon_xp_2000_thorton_3182.53_mhz 1.61: OK - http://hwbot.org/submission/2324603_i.nfrar.ed_cpu_frequency_athlon_xp_m_3000__2923.81_mhz 1.62: OK - http://hwbot.org/submission/2354919_i.nfrar.ed_cpu_frequency_athlon_xp_3200_2779.22_mhz 1.63: OK - http://hwbot.org/submission/2368188_i.nfrar.ed_cpu_frequency_athlon_xp_m_2500_3288.84_mhz Guess it might be setup dependent, but that's on my Abit NF7s. As I said, the problem was related to wrong FSB detection.

cpuz was fine with my Abit boards until 1.55 or 1.57, don't remember exactly. Then it was fixed in some of 1.6x versions. Haven't tried recently, but it seems to be broken again., however the issue in the past was wrong FSB detection.

Just don't make it past 7012MHz, ok?

Yep. How would you determine if I'm a skilled race-driver if I'm the only one in a race without any competition? What is the base you measure my skills against? I may even run with stock tires, without any suspension and aeordynamics tweaks, without even using 20% of the power I have in my hands, but will still be №1, because there's no competition.

You seem to miss the fact that the change is not applied on the production server and the missing points issue for some users is not connected in any way with the new algorythm. But on the topic, I agree with the things jjjc said. There are a lot of people benching old stuff that awards them 2pts or close to that. How running an obscure server setup is any different from this? If they are not doing it for the global points, then they will continue to run these setups. Otherwise it's just an exploit/backdoor in the point algorythm that these people are absuing. This change is not made to reflect time spent on a certain score, it reflects the competition in a certain ranking. You can't ever measure the time spent. I might have spent 6 hours on a single s.939 score that awards me 2pts. Should I get 15pts for that, or 20pts? I do it for the passion and for fun. Why should I have 60 global points in a category without any competition, while I'm getting 2 HW points in a non-popular hardware category. IMO it's the same, so the change is the right thing to do.

Well, it's almost always like that. If you want to maintain your position, you need to rebench all cards (which are still not dead ) with the new CPUs. It depends on benchmarks, ofcourse.

I don't have many exotic categories with a lot of points. Actually it might be Sempron 145 Cinebench only where I have full Global points and that's a rather popular category. Although I have many subs in low-points rankings with old chips/vgas, it won't hurt my ranking. There was definitely something wrong, because I had not global and hw points, but now they are back to normal level, though I expect when everything is recalculated I will move back to 30-40th place (16th now).

Think again OS4105WLU4DGN

I still have to test that, but just wow... it's 3:30 minutes better than my time!

Cinebench was crashing at 6.5GHz, but no problem with wprime, however score was awful - not worth posting it