My last attempt - AMD 3000+ Week 41 just arrived!

[dnottis]

Like everyone else here - my AMD3200+ week 41 that topped out at 2.520Ghz just wasn't enough and with all this talk of the magical new 3000+s - well I just had to byte. This just arrived today - ordered from Monarch Computer on Monday. Week 41 3000+ UPMW as opposed to my 3200 which is a RPMW.



So far I'm at 2.6 GHz with it. The 3200 would refuse to SuperPI or Prime at anything about 280 x 9 regardless of volts. I'll update later tonight once I get some bench time in with it.

[cupra]

for my opinion a good process for oc is not depends only in the week of the process
the most important thing is a luck.

2.6g is good, week 41 in new , lets hope he have luke .

kept going

[Ferris23]

Which are the new chips at Monarch, the Retail or OEM?

Thanks

[Dumo]

for my opinion a good process for oc is not depends only in the week of the process
the most important thing is a luck.

2.6g is good, week 41 in new , lets hope he have luke .

kept going

It depends on cooling

[dnottis]

This was an OEM from Monarch. I managed to hit 290 x 9 2.6GHz however it wont prime stable with the memory 1:1 that high (XMS4400). I remember someone saying the Neo2 Plat had issues 1:1 past ~285 though. Running DDR480 it primed 290 x 9 for a few hours tonight.

[Ferris23]

Do you think it would be worth trying one of this if I currently have a 3500+ retail Winchester that is Prime stable at 2520 but not much over. While I'm happy with what I have, the OC is sub par so I wouldn't mind trying another chip if it had potential to hit 2.7 on water.

[dnottis]

Looks like max 1:1 @ 1T is 285 x 9, I can go up higher to 290 x 9 1:1 @ 2T. Trying to find Prime95 stability though and checking benches to see what is actually faster. I can go 290 x 9 and drop the memory with the divider seems the chipset is the limitation running 1:1 over 285 just crashes or reboots on me.

[cupra]

so this is the limit of your memory at t1 and its better form t2,
did you try higher vdd ?

might be the cpu can go higer then 2.85g, you try 265 x 10 to check it out ?

[Bungo]

3000+=9x multi.

Turn on a divider and se what the cpu can do instead.
Running with divider isent bad, u just gotta be lucky with the ratio so u can max both mem and cpu at one setting. lalalolo.

[L0$t Pr0PhEt]

2600Mhz @ 1.5v

[MaxxxRacer]

id almost be willing to trade you my 3200 that does 2.8ghz just to get the good memory controller on your chip.

[Hype]

2600Mhz @ 1.5v

That's nothing... my 3200+ will do 2600mhz @ 1.45volt stable!



and a 300 FSB shot (not stable though)

[L0$t Pr0PhEt]

Looks like you got yourself a good motherboard and cpu

I know for sure when I'm gonna buy a winchester and board I will have the badest overclocking cpu ever seen :p

[cupra]

3000+=9x multi.

yes yes my mistake

[skate2snow]

Cupra : You are right and wrong... The most important is luck, but the week and batch will meen something.... If 90% of the chips can do 2.6GHz stock volts, you just have to be lucky to not catch the 10%. But if in that batch 30% does 2.6GHz, members wont be attracted that much by that batch...

[dnottis]

UPDATE TIME!

I've tried my memory @ 2.85 and SuperPI crashed immediatly...no biggie check this...

Ok folks - here is the farthest I can go 1:1 1T (this seems to be a limit of the chipset not my memory)





So above 285 I either had to use the divider..or I decided to try 2T which let me go higher. At this speed the 2T performance drop is actually un-noticable since the gain in overall speed at 292 offsets it. Either 1:1 2T @ 292 (which I may add is still stock volts (giving it more does nothing) and timings on the memory - Crucial rocks!). Here's a shot with it running prime for 8 hours now.



So here are a few things to note - MAX on my chipset was 285 1:1 1T. Above this either use the divider or 2T which let me push it higher. I may be able to get more out of it even, I'm still playing but I want to maintain 100% stability so it may take some time. At 2630 I am MORE than happy! It took 5 winnies to get a great overclocker - 4 to get a decent one (3200 @ 2.5).

[dnottis]

so this is the limit of your memory at t1 and its better form t2,
did you try higher vdd ??

Nope, limitation of the chipset - many others are topping out about the same - search the boards here.

might be the cpu can go higer then 2.85g, you try 265 x 10 to check it out ?

can't it's a 3000+ mult is 9

[dnottis]

3000+=9x multi.

Turn on a divider and se what the cpu can do instead.
Running with divider isent bad, u just gotta be lucky with the ratio so u can max both mem and cpu at one setting. lalalolo.

True but some apps, especially Q3 based really show performance gains with higher memory speeds. I really want to keep memory and CPU at 1:1 even going to 2T and moving bus speed up 1:1 that way is just as fast.

[dnottis]

I know for sure when I'm gonna buy a winchester and board I will have the badest overclocking cpu ever seen :p

Good luck man - took me 5 Winnies to get a good overclocker. Oh yea and if you go with the Neo2 Platinum - pull the Chipset and apply artic silver it really helped me to break my ~ 2.450 GHz OC barrier

[cupra]

Good luck man - took me 5 Winnies to get a good overclocker. Oh yea and if you go with the Neo2 Platinum - pull the Chipset and apply artic silver it really helped me to break my ~ 2.450 GHz OC barrier

u have week 42 korea like me ( i have 43)
the taiwan chipset in much better for oc .

i pull the chipset with artic silver also and put some ocz coolers .

[clockedOut]

I just received a 3000+ 0442SPBW and a Neo2.... but I leave early Sat for Florida until the 31st. So no testing until next year.

[cupra]

I just received a 3000+ 0442SPBW and a Neo2.... but I leave early Sat for Florida until the 31st. So no testing until next year.

happy new year

[aonic]

nice, how can you people afford so many CPU's? lol

[Hombre]

Through RMA's I think dnottis did you got your chip straight from AMD?

[Super strokey]

how do you rma a good chip and get them to send you back a new one?