I.nfraR.ed

All stages are open, except first 2 and we have just 5 days left Other subcompetitions open next year, but the final ends in 5 days too

But on the other hand..it is like the winner of the World cup decides to erase their scores, because FIFA is bad. These scores are nothing personal and can't compromise the person in any way. It's not private pictures on facebook. What will happen if someone decides to remove all his posts in a forum? Everyone's else posts related to the deleted ones will just stay there in the void without logical connection. Keeping that in mind, it is a public history and in most cases you can't "undo" it. So on a second thought I believe these scores should stay at least under the generic "anonymous/deleted user" name.

Unicode seems totally broken for me. I can't send a PM in Cyrillic. ***1058;***1077;***1089;***1090;, ***1090;***1077;***1089;***1090;. <-- Test, test in Cyrillic

Same with update/edit score.

I think 282MHz 1:1 would not be possible at tight timings, at least for my setup. Maybe with low ambient on DFI Ultra-B it would be possible, but it is summer here and ambient is too high. I did some test with my AN7 with my custom bios, but 277 is the max I can validate at the moment with 1:1 DC and tight timings: http://hwbot.org/submission/2915144_i.nfrar.ed_reference_clock_an7_276.7_mhz/ 1M benchable about 270-272MHz, 1:1 DC tight. With DFI I could probably do some MHz more, but still not 282. It's not a walk in the park even with nforce4 chipset. Now what left is to reinstall windows and get a proper score, but the AN7 is a little faster than the DFI.

Yes, I think they should stay like that. It was his call and although it is sad we will never see the scores again, we should respect his decision. Same applies to Sam's scores, unless he gives a green light. It is sad, but we have to accept it.

Exactly. It was always motivating trying to beat those results. I wanted to go through my s.A collection and bench every single model I have, but after some of the best scores are now gone, it's not so rewarding if I get the 1st place. There are still some old scores left, like tiborrr's and from older generation overclockers, but these rankings lost a big part of their value. Same applies to old intel sockets, but I'm concentrated at AMD only. I would be glad if there's a way those scores can be made visible again (both Sam's and Karl's) without points. Unfortunately we can't ask Karl for permission

That's what I meant, yes. He is well known and recognized in the community. Sam, personally I don't care about your other points with more recent hardware, but you should have left your s.A, s.939 and s.AM2 scores. Now these rankings lost another big % of value, after Karl's s.A scores were gone

Yeah, I guess he sees no sense in trying to get higher and stay on top of the ladder when this does not mean anything at all (none or close to no recognition).

You can find good D9 even in low-binned stuff. For example this is Crucial value ram without heatsinks and rated 800 6-6-6 1.8V: http://hwbot.org/submission/2306681_i.nfrar.ed_3dmark03_geforce_fx_5900_xt_9438_marks Less than 2.5V if I recall. Chips are D9GKX. 2.1V is your limiting factor though, 4-3-3 would be more realistic, unless you bin a lot of D9.

Ok, then is it really limited to AM2 only or AM2+ processors are also allowed? Such as Agena (first generation Phenom).

No, you can't. You can run AM3 CPU in AM2/AM2+ motherboards, because AM3 CPUs have DDR2 controllers. AM3+ CPUs can be used only in AM3+ boards and older AM3 with proper bios support. For example Asus Crosshair IV Formula is AM3, but supports AM3+ CPUs with certain beta bioses. AM3 CPUs are up to Phenom II/Thuban. Pure AM2 boards support AM2 CPUs only, which means Athlon 64/Athlon 64 FX and these are similar to s.939 CPUs. Socket AM2+ boards support both AM2 and AM2+ and with bios update - the newer AM3 CPUs, but limited to X4. Thuban is not supported on AM2+ boards or I haven't seen such a board with bios support for X6. This is incorrect, some AM2+ with updated bios support all AM3 CPUs, including Thuban. AM2 is the successor of s.939 and this round should be limited to X2 6400+ Windsor F3 being the fastest of the stack. Or basically K8 micro-architecture. Looking at hwbot database though, X2 7750 and X2 7850 (and others from this family) are incorrectly listed as AM2, while they are AM2+ (K10, Kuma core). Also the infamous "newer" X2 5000+ and X2 5200+ which were in fact locked Denebs and some of them even unlocked to X4. They are listed under the AM2 category in hwbot database, which is again not correct.

And I will start earlier than some hours before the end when I found my 9800 Pro does not work Good job everyone.

Several hardware-related stores sell Dimastech stuff (caseking for example). Maybe contact them and ask about some cooperation. They should have a way to contact him. HWBOT should be able to do something too, since they had Dimastech as a sponsor for the World Tour. Not sure what would be the response, but worth trying, I think.

This one has the same stepping as my XP-M 2500+. With that low vcore doing 3.2GHz+, it should be very decent once pushed to the max. I would say it is rare to see XP-M 2800+ that good...

This CPU has great potential!

cpc on, cpu on agressive, no wazza, totally broken windows.

Mine worked at 3.7V VDIMM for some time without problems. And my NF7-S boards have always worked at 3.6 - 3.7V, but I think NF7-S is more reliable than DFI Ultra-B.

I have maybe 10 dead s.A boards, but some of them were DOA, so I haven't killed them all on my own Don't even remember how I killed some of them, but... NF7-S due to leak from the waterblock, one or two from dead SB/NB and one from vtt=vref mod (v0.51 I think). I have 3 NF7-S still working though and they all are v0.53. One DFI Ultra-B due to high VIO, but it had problem with VDIMM mod from the beginning. One Soltek SL-75FNR2 for unknown reason. Have 2 AN7, but one was DOA, the other one still going strong. I use this AN7 to install windows from USB stick, then move to NF7-S for benching

That LCD shows voltage on 3.3V rail.

The problem with my Ultra B was that conventional vdimm mod didn't work either.

I don't see a way for 282 1:1 and tight timings (BH-5) though It's not so easy even with s.939. Will fix windows later, however I think the other board is dead (the one without capacitor between 2 FETs to the mem slots) from the high 3.3V rail. It was behaving the same with just vdimm mod. So maybe it is good to be extra careful if your infinity is the same.

Leave just 2 cores on and crank that NB up (disabling cores helps for higher NB clock).

No, he wants the scores on the front page, then disables points after they are not considered as new top scores anymore and get dropped from that table. Nice trick

All my 3 DFI Ultra-B can do at least 268-270MHz DC 1:1. 282 will be hard though.