we've found THE GOLDEN ONEs : AMD Athlon64 World Record Thread!!

@stunned_guy Yes it does, a terrible cold bug :( At -25 I managed 4144mhz, at -28 it was down to 4050mhz and at -35 I couldn't boot. It's a pity really as it was scaling very well with cold, at one point I could 8m at 4ghz and it looked as if I may challenge the K8 frequency record. Oh well, it was loads of fun to be clocking that and setting records while my team mates were next to me getting all the attention with their 4.6ghz i7 965. Man that was a hideous Intel... :)

@knopflerbruce That's exactly what happens to me once the Vapochill takes it below 7c. I can bench fine all the way down to -30 with single channel but I can't even post with more than one stick of memory. I'll try another board, coincidently I've been trying with the Ultra D that got me 3.9ghz with single channel and it seems to have gotten stuck :p:

By the way, it's midnight here in South Africa and I'm freezing my bum off in front of an open window trying to get my RAM to do 300mhz c2 and my TCCD to do 351mhz or higher. So far my BH-5 won't pass 288mhz which I think is because of the cpu. I've never really abused them and another chip took them over 293mhz. Oh the joys of being a hardcore 939er, tomorrow I'll be sick in bed but I don't care :D 939 is my favourite sex position, it's hard to find a partner but for those who do it never gets old :)

Crazy :D

My beloved TCCD wouldn't even do 330mhz last night, I'm positive the cpu isn't up for it. I didn't break any personal records and I can already feel I have a scratchy throat but it was tons of fun :D Do you think I'll get a higher overclock using Win2k?

Oh, my TCCD can boot to Memtest at 360mhz but the errors I get... I should've taken pictures, I've never seen them before. Windows was no go though, I wonder if TCCD will appreciate a suicide run at 4v :rolleyes:

you guys are crazy :toast:
soon i'll bench a few old 939 cpus too ;)

Jonathan, you can give up to 3.5 volts on your TCCD with a 120er fan above. That should works. But with 2x256 mb you can´t get any fast time in superpi because the dims didn´t have the "dram idle timer".

Quote:

---

Originally Posted by stunned_guy

Nice to see u back here on socket 939 ;)
The YBBLE couldt go nice under cascade...but the CPBW is crap...no chance to go high.
These are all venice E3 so I guess that there will not be some high scores because the best of the 3000+ are all venice E6. But give your best it could be one of the good scores.

---

Héhé , i love the Old School , thk for your information :up:

Yes it's no very good the E3, i have just test before my departed a Taipei , my better pcs in 3000+ do 2950mhz on air (i think isn't good) and my 3200+ do 3.1ghz :)

But i have buy the good price for the lot of A64 :) i don't choose the stepping :)

I back my home tommorw , i need test during a few week in waterchiller and my caskade :) wait & see

Quote:

---

Originally Posted by Oj101

Thanks for the link boblemagnifique :)

---

No pb , you're welcome :)

Hey stunned_guy, why do you know my name but I don't know yours? :p: I know that TCCD often degrades or dies above 3.2v but has anyone ever tried to really kick them to the ground with high voltages? You might find that they have a "voltage hole" where say 3.3v to 3.8v doesn't help but 3.9v suddenly makes a difference again. My CH-5s are like that, 3.3v to 3.6v doesn't help anything but stability while 3.8v allows them to bench higher.

Anyway, I'm not pushing the TCCD for spi, just for high memory validations so performance doesn't matter :)

My best TCCD kit was G.Skill GBLA 2x256 mb with that I hold the wr for maximum frequency. I cant find the cpuz validation but 370 mhz + with dual channel was available. voltage was 3.46 volts with 120er fan above.
The older tccd were killed above 3.2 voltage but the newer ones could get up to 3.6 voltage samsung told somebody.

I know your name because here in the thread you told this to us :D My name should also here in the pages, try to find it out :D

Edit:
here is the validation with 1x512 mb geil one tccd: http://valid.canardpc.com/show_oc.php?id=120581
and thats the ultimate frequency with 2x256 mb g.skill gbla: http://valid.canardpc.com/show_oc.php?id=141947

enjoy :D

This is still my best, http://valid.canardpc.com/show_oc.php?id=288119 at 3.1v. So much trying, no improvement :(

Keep pushing, Jo ;)
Was sind das für G.Skill ? GBLA like mines ? PC4800 ?

Quote:

---

Originally Posted by knopflerbruce

It's idling in BIOS now at 2.08v to raise the temps so I can remove the pot.

---

Why don't you use a mapp gas mate ? It makes the whole process a LOT easier. De-freezing GoRiLLakoS's pot from -186*C (Cedarmill benching :D) to -20*C was a matter of a minute or two :yepp:

Quote:

---

Originally Posted by George_o/c

Why don't you use a mapp gas mate ? It makes the whole process a LOT easier. De-freezing GoRiLLakoS's pot from -186*C (Cedarmill benching :D) to -20*C was a matter of a minute or two :yepp:

---

...because I have NO clue what it is, and therefore don't have one:p:

I guess that was German for "Is your TCCD G.Skill GBLA PC4800 like mine?" :D Yes, it is ;) As far as I know, LA is rated at DDR400 2-2-2-5 and DDR600 3-4-4-8 while LC is rated at DDR400 2-2-2-5 and DDR550 2.5-4-4-8. Mine's the former.

Quote:

---

Originally Posted by knopflerbruce

...because I have NO clue what it is, and therefore don't have one:p:

---

Mapp gas is basically liquefied petroleum gas mixed with methylacetylene-propadiene. It can be used in combination with oxygen for heating, soldering, and brazing due to its high combustion temperature in oxygen.

Wanna see how it looks like ? Here you go :) First 3 photos are print screened from Windows Media Player. I was benching with the rest of the guys in Athens, and noticed that Stelaras, instead of patiently waiting for the pot to heat up, took the mapp gas and a lighter and went crazy :D :yepp: The minute I saw him doing it, I pressed "Record" button on my camera :rofl:

That was my first time on LN2, that's why I was THAT excited - using cameras and everything :ROTF:

De-freezing pot with mapp gas (4th picture shows what mapp gas looks like):

http://img5.imageshack.us/img5/29/mapp1.png
http://img8.imageshack.us/img8/1896/mapp2.png

http://valid.canardpc.com/show_oc.php?id=60175 my best TCCD run that I can find.

I did hit 340 at 3-4-4-8 timings, and 320 at 2.5-3-3-7 once. i think the voltage was around 3v, with a fan.

I took screenshots and even those appear to be gone. Cant find any of my old AMD rig screenshots :(

http://i42.tinypic.com/23ra5x1.jpg

FOUND ONE

nice tccds you got there:up:

Over 9000MBps :D

@gianni:

I have a few "double entries" on my account now, as I updated a few scores. I guess the best thing would be to delete the bad ones:p: Also, I posted that 3200+ NC WR in the 3000+ category:rolleyes:

AM2 is rocking:
http://www.abload.de/image.php?img=x...3825_1dgt4.jpg
http://www.abload.de/image.php?img=x...3840_1m17e.jpg

@ Knut
I have fixed some things ;)

Quote:

---

Originally Posted by stunned_guy

AM2 is rocking:
http://www.abload.de/image.php?img=x...3825_1dgt4.jpg
http://www.abload.de/image.php?img=x...3840_1m17e.jpg

@ Knut
I have fixed some things ;)

---

Not so bad, I think:D

I got 21 CPUs today... Including an Opty 1224 and a 6000+;) The 1224 can do 3.6 quite stable on water cooling (orthos ran for 5 minutes on 1.52v and 26c ambients). Dunno about the 6000+ yet:)

Nice 6400+ you have there stunned_guy. I see you got yourself a decent 6000+ as well, you planning to challenge my record? :D Please give me motivation to kick butt again :D

I've also just bought myself 2x512MB BH-5s which do 270MHz, time for some more FX-55/Pi action. I kinda lost control and also took 2x1GB OCZ UCCC DDR500 and 2x1GB TG UCCC DDR500. I'm a hopeless 939 addict :(

EDIT: Stunned I can't find your name anywhere :(

New 6000+ challenge;) Check this out:

http://www.xtremesystems.org/forums/...9&postcount=34

(is this a good way of notifying you of scores, or is it good enough to post in my thread only?):p:

Quote:

---

Originally Posted by Oj101

Over 9000MBps :D

---

Let me rephrase:

IT'S OVER NINE THOUSAND!!!

Please give me some details on what you used for your 6000+ record? Voltage, temperature, stepping?

:(

Revenge will be mine :)

EDIT: That was really low, you ruined my weekend. I've just ordered 20l of LN2 :D

EDIT Part Dues: And I've have another 6000+ :D

Quote:

---

Originally Posted by Oj101

Please give me some details on what you used for your 6000+ record? Voltage, temperature, stepping?

:(

Revenge will be mine :)

EDIT: That was really low, you ruined my weekend. I've just ordered 20l of LN2 :D

EDIT Part Dues: And I've have another 6000+ :D

---

-20c, measured by ITE, and 1.65-ish volts. CCB8F 083x something. (pretty sure it was an xPMW, but x is def. not M). The SN ends with 80003.;)

Good luck with the hunting!:D

@ripping admin(s): please add opty 1224;) And Phenom 9900 if you want to:) . I've also got some low voltage AM2 chips, one 2000+, one 2600 and one 2650e. Here you have the 2650e: http://www.cpu-world.com/CPUs/K8/AMD...650DPBOX).html - the difference between 2600 and this one is that the 2600 is a G1. The 2000 is a G1 200x5 part:) Sempron LE1300 is not listed, either: http://www.cpu-world.com/CPUs/K8/AMD...300DPBOX).html - and same with Opteron 256, and the 130nm 1xx series for socket 940 - 140, 142, 144, 146, 148 and 150:p:

These AM2 chips really love lowish voltages, it's a pity because decent scaling would put K8 at about 5GHz :D

I think the chip I just received is from the same batch as yours, from what I've heard it's an 0831. Let the battle begin ;)

Vince and James can murder each other over 3DMark01, I've got my fun right here :)

Crazy scores mates ;)

@ knut
I will check this in ripping.org.

great work knut :toast:

Quote:

---

Originally Posted by Turrican

great work knut :toast:

---

Thanks. Got a new WR today, too;) Athlon pwnage - I got all 4 now (actually 5, but the 1640 lima isnt added yet):D

New board: M3A79T Deluxe
Chip: Old 6000+
LN2: Approximately 10L
Results: Coming later today :D

Knut beware :D

Quote:

---

Originally Posted by Oj101

New board: M3A79T Deluxe
Chip: Old 6000+
LN2: Approximately 10L
Results: Coming later today :D

Knut beware :D

---

I've got "backup scores" so bring it on:p:

...and 10L Ln2;)

You take the record :( I could Pi at 4.09ghz but not touch 4.1ghz at all. Vapo/DFI time again I guess :)

Quote:

---

Originally Posted by Oj101

You take the record :( I could Pi at 4.09ghz but not touch 4.1ghz at all. Vapo/DFI time again I guess :)

---

What's the clue behind running 4ghz+ 1m dual channel? I'm always limited to single channel at those speeds... happened on my 6000+, my 1224 and both my 6400+'s:(

Now that would be telling :D Honestly I've never had that issue before, but I do normally go a bit on the high side with NB I/O voltage :shrug:

Would be great to break that barrier... ran wprime32m at 4168mhz on my Opty 1224 last night (froze because of coldbug after, though). I wonder what it can do 1m at without those problems...:p: I smell sub 20s then!:D

Teaser for you s939 folks: http://valid.canardpc.com/show_oc.php?id=588169

Temp: 194 in ITE:p:

Amazing that 25c only gave an extra 30mhz or so:shakes:

Btw, it's an X2 4600+ Manchester (ES, as you can see).

Great clocks, Knut !!!! ;)
Stepping of that manchester ?

I could only do that score but with high htt under single stage... http://valid.canardpc.com/show_oc.php?id=413660

damn nice scores knut & phillip. :D

Quote:

---

Originally Posted by stunned_guy

Great clocks, Knut !!!! ;)
Stepping of that manchester ?

I could only do that score but with high htt under single stage... http://valid.canardpc.com/show_oc.php?id=413660

---

No stepping, blank IHS:p:

It was pretty good on dryice, but going from -50 to -80 gave me nearly nothing:(

I know this chip isn't the best there is, if the price goes down i will look for a better one;)

A couple of scores from the last few days sessions:

http://valid.canardpc.com/show_oc.php?id=588669 6000+ @ 4233MHz
http://valid.canardpc.com/show_oc.php?id=588655 1222 @ 4091MHz

*cries and admits defeat* Jeez Knut, you killed me with that score!

Oh, and hi PHILIP :p:

Quote:

---

Originally Posted by Oj101

*cries and admits defeat* Jeez Knut, you killed me with that score!

Oh, and hi PHILIP :p:

---

4250 is on the way, but I want a faster pi board than my nf590 before I start pumping 1.7v+ through these chips. Not that I worry THAT much below 1.8v, but I'd rather not lose this golden one:D

It's pretty pointless with my chip, 1.72v gave me my max. Even 1.9v didn't help at all.

Quote:

---

Originally Posted by Oj101

Oh, and hi PHILIP :p:

---

Please Philipp ;) it takes so long to find it out but hey thats the way ;)

damn i also wrote your name false, sorry bro :p:

Quote:

---

Originally Posted by Oj101

It's pretty pointless with my chip, 1.72v gave me my max. Even 1.9v didn't help at all.

---

1.79v made a small difference for my LE1620, even at +15c. Not much though, just from stable enough to start validation to complete the validation required for WR:p:

No problem, bro ;)

+15c ? Thats hard for overclocking and single stage / dry ice. I have here one x2 be-2350 and be-2400 but both failed under single stage. very hard cold bug :(

Quote:

---

Originally Posted by stunned_guy

No problem, bro ;)

+15c ? Thats hard for overclocking and single stage / dry ice. I have here one x2 be-2350 and be-2400 but both failed under single stage. very hard cold bug :(

---

All brisbanes have a wall at 15-16c or something... the kind of wall you'll find on Intel CPUs, they just don't work below that temp:p: I hate it too, the key is to use a pot that has little mass so the temp changes more rapidly - then the waiting is not so long. SS is quite hard, but tro to fasten it on only one side, and use the other one for temp adjustments;) That's what I did when i used my phase unit.

Hey I just copied Turrican :yes:

@ Knut
Same thing to me. Most of my Brisbanes shut down if the cold bug was found. The PC do not freeze --> it shuts absolutely down :D

Quote:

---

Originally Posted by stunned_guy

@ Knut
Same thing to me. Most of my Brisbanes shut down if the cold bug was found. The PC do not freeze --> it shuts absolutely down :D

---

Yeah, the annoying part is that it takes a while to make it bootable again:rolleyes: I wonder why it shuts down instead of freeze...

Thats the question because socket 939 there is no shut down. only freezes and you can start the pc again but here its hard to wait for another run.

yeah guys, that also happend me. pretty bad tough.

speaking of cold bug on 939 - I bought two 3200's, one of them Venice, second Manchester, and I plan to do a little DI fun (there will be those 3200, x2 3800 Toledo and maybe my old 3000+, if it's still working) in few weeks time, just wondering where they can catch the cold bug :) of course there will be a warm-up with water cooling :D

I know that it's the matter of every single CPU, but what's the worst CB you encountered with those cores?

Quote:

---

Originally Posted by phobosq

speaking of cold bug on 939 - I bought two 3200's, one of them Venice, second Manchester, and I plan to do a little DI fun (there will be those 3200, x2 3800 Toledo and maybe my old 3000+, if it's still working) in few weeks time, just wondering where they can catch the cold bug :) of course there will be a warm-up with water cooling :D

I know that it's the matter of every single CPU, but what's the worst CB you encountered with those cores?

---

Manchesters and Venice's should be able to stand -20c, or maybe even a bit more, before you hit the CB;)

sounds fun :)

Quote:

---

Originally Posted by phobosq

sounds fun :)

---

It varies a little bit though, and if you have one of the venice labeled San Diego's (for instance CACAE-steppings) you will face some problems quite a bit sooner than -20c:p: But manchesters... -20c should be no problem no matter what stepping you get.

Venice is NBBWE 0617DAD
Manchester - CDBWE 0543UPMW

Sorry to say that but both cpu will sucks :(
The first is an chinese product and the second one is too old...

well I hope you're wrong :P
at least I'll have some fun, then I'll find another ones and those will be sold if they suck, I bought them dirt cheap:)

EDIT:

I dismantled the computer from the case and assembled it on the table with the Venice - and it's not starting, even the LEDs don't light up when I turn on the PSU :(
I'll have to check the PSU first on my old NF2 Infinity before looking for another NF4 - had to go to work so I couldn't finish :D

Quote:

---

Originally Posted by phobosq

well I hope you're wrong :P
at least I'll have some fun, then I'll find another ones and those will be sold if they suck, I bought them dirt cheap:)

EDIT:

I dismantled the computer from the case and assembled it on the table with the Venice - and it's not starting, even the LEDs don't light up when I turn on the PSU :(
I'll have to check the PSU first on my old NF2 Infinity before looking for another NF4 - had to go to work so I couldn't finish :D

---

Sounds weird...

I lost a WR yesterday btw:p: Tapakah stole my 3500+ Venice WR:mad: Like 0.3MHz, so I think it's possible to take it back. I think the old WR was 36xx Mhz, so there is a good chance I didn't bother maxing the chip 100% since I already got it. If so, there should be at least 10Mhz more in it:D Still, 3.8 is possible with a really good chip, some non-bugged E6 revision should be able to do that.

Edit: lol @ post count

Now create a new account :)

Nice benching all !!!

Quote:

---

Originally Posted by Gamer

Now create a new account :)

Nice benching all !!!

---

New account?;)

I think I'm going to use my single-channel Venus board as wPrime/CPUZ benchboard in the future - I fear my Expert is a bit worse at sub-zero. Perhaps that's what I need to get back to the #1 spot of the 3500+:p: Yay! That means this :banana::banana::banana::banana:ty ASUS Crosshair will be my 24/7 mobo...

Quote:

---

Originally Posted by knopflerbruce

Sounds weird...

I lost a WR yesterday btw:p: Tapakah stole my 3500+ Venice WR:mad: Like 0.3MHz, so I think it's possible to take it back. I think the old WR was 36xx Mhz, so there is a good chance I didn't bother maxing the chip 100% since I already got it. If so, there should be at least 10Mhz more in it:D Still, 3.8 is possible with a really good chip, some non-bugged E6 revision should be able to do that.

Edit: lol @ post count

---

Yup, the PSU was checked on four mobos and it doesn't show any signs of life. It's 5 year old Enermax 465, still 20-pin, so it really had a right to die :) looking for new (and cheap :D) one - right now OCX ModXStream 500 is on my mind

I saw the 3500+ WR on OCX - the thread also contains one interesting score - 389 MHz with TCCD (A-Data Vitesta 566) - keep pushing Knut, it's just a matter of 1 HT :D

What was going about post count? :D

Quote:

---

Originally Posted by phobosq

What was going about post count? :D

---

I guess it was:

posts 939 :welcome:

Quote:

---

Originally Posted by MaSell

I guess it was:

posts 939 :welcome:

---

Indeed:D

And I just passed 754 :)
up to 775.

Quote:

---

Originally Posted by phobosq

Yup, the PSU was checked on four mobos and it doesn't show any signs of life. It's 5 year old Enermax 465, still 20-pin, so it really had a right to die :) looking for new (and cheap :D) one - right now OCX ModXStream 500 is on my mind

I saw the 3500+ WR on OCX - the thread also contains one interesting score - 389 MHz with TCCD (A-Data Vitesta 566) - keep pushing Knut, it's just a matter of 1 HT :D

What was going about post count? :D

---

Forgot to ask... can you please link me to that thread?:p:

Ygpm

EDIT - found interesting SD 3700 - link - link leads to Polish auction site, but I guess you can read the pictures :)

Briefly - SD 3700+ KA5CF, LCBBE0640CPMW, but one pin missing (last pic) - in AMD docs I found it's DBREQ_L - description says 'Debug Request', maybe the CPU can run without the pin (reattaching it shoudn't be hard, anyhow :D)

Quote:

---

Originally Posted by phobosq

Venice is NBBWE 0617DAD
Manchester - CDBWE 0543UPMW

---

New PSU arrived, got Topower SilentEZ 400.
Venice is total crap, walls around 2,7G at 1,6V - more voltage doesn't help at all, I'd have to freeze it (currently it's watercooled). Max suicide screen was around 2.82G - didn't even bother to capture it.
I'll start fighting with Manchester in a while.

Did you buy that 3700+ toledo E6 ?? 0640 was good but not good enough. the world record was made with an 0628 bpmw e6.

Quote:

---

Originally Posted by stunned_guy

Did you buy that 3700+ toledo E6 ?? 0640 was good but not good enough. the world record was made with an 0628 bpmw e6.

---

Not yet, 17 hours left. Didn't know that was Toledo, though. :)

Gonna freeze some ice now, maybe dropping temp will help the Venice reach 3G.

EDIT - the best I could do with Venice with just water (no ice so far) is 2940 - valid. However, it required 1,87V - with 1,75V I could do only 2880 - it seems that heat is the main issue here.

Too bad scores to play in the world league with us :D
But good thing will play a while. You have to search for good steppings not the chinese one.

Quote:

---

Originally Posted by stunned_guy

Too bad scores to play in the world league with us :D
But good thing will play a while. You have to search for good steppings not the chinese one.

---

I know, just started my search :)

EDIT - got that 3700+, hope it works :D

EDIT #2 - threw some ice to water tank - temp drop about 20C (water @13C) - gain some 100 Mhz - 3050 valid

EDIT #3 - the Manchester is similar or a bit better than Venice - did 2961 with 1,72V but it was already to hot for him so I didn't try higher :) did some benches @~2,8G, I will retry tomorrow (work work :P)

S.I.H.:p: (3000+ NC, just got to validate it first;) )

Btw, 2 freak accidents here. I saw someone who broke a GB socket at -192c, but 3h ago I did it at -15c... Not the entire socket, but behind the "CPU matrix" or what you'd call it, where the lever is. So... that mechanism is broken, and I must use all kinds of torture equipment to get these chips out of the board when i bench it:p:

The second one was quite a bit bigger... I was benching a 3rd 3000+ NC (this one even better than the WR one I think), and suddenly I hear a "swooosh"-like sound. It came from the water loop I use for testing when I dont have LN2. For no reason at all, the block and the tube on the outlet of the block decided they didnt like each other any more, so the Iwaki pumped 5L dirty, old water STRAIGHT at the benching PC through the block (which OF COURSE was pointing in the wrong direction:p:).

Looks like everything survived, but I haven't tried the HDD yet. It's in the oven at 70c as we speak:p:

Edit: S.I.H. score gone, HDD dead after the little bath it got. And seems as if the board died as well, so I must use either an a8n sli deluxe og the crippled venus in my main pc tomorrow :p

lol, looks like you are back in action. :D

Nice to see knut still in action !
i tested one 3200+ e6 last week but it was awesome...not good chip.

OMFG, I had 2 WRs opn my bench drive, and then after a BSOD the entire desktop (with all the validations, including 3000+ sempron 34xx mhz and 3000+ newcastle @ 328x mhz........).

My bad luck will never end:rolleyes:

Edit: looks like the scores got moved to a different part of the hdd, here are the two records:

http://valid.canardpc.com/show_oc.php?id=651048 (sempron 3000+)

http://valid.canardpc.com/show_oc.php?id=651044 (athlon 3000+ nc)

Too bad my rage killed the damn expert before I found the validations... ;) Was a TINY bit tired of the madness I've experienced lately.

Got my 3500+ Venice WR back:p:

http://valid.canardpc.com/show_oc.php?id=652802

I had a dump at 3771 too, but it was corrupted. Bad mount, which is why the vcore was a bit low. Unfortunately, I didn't have time (or ln2) to do a remount:( With 1.7v I think that 3771 one would've been valid.

Edit: Look what I just did with gianni & stunned_guy's old WR chip:

http://valid.canardpc.com/show_oc.php?id=651044 :D

Post 939 goes here :D

You had a DUMP at 3771? Do you have any idea how wrong that is???

Post your newest scores dude :D

Great results knut. Keep pushing.

Quote:

---

Originally Posted by Hampti

Great results knut. Keep pushing.

---

Thanks. Still pushing. Just got the best efficiency of K8 singlecores in wPrime 1024m:D 3500+ SD FTW!

i saw that, well done knut :D