The Bulldozer week/batch thread...

[BeepBeep2]

heat is problem every time...with Phenoms, Bloomfileds etc. IMC improves soemtimes memory stability and clocks (+performance of CPU with higher uncore and a bit CPU stability at high clocks), but chips are hotter and harder to cooling.

Guys, tomorow Il try my FX-8120 again under LN2, hope, chip will be stil live after that

I'm sorry, but I had no issues with heat on either my Phenom II 965BE, 565BE or 1100T. However I can hit 85c on my FX-8150 with a mere 1.5v vcore and 1.23v CPUNB voltage on water. All of my Phenom IIs loaded no more than ~60-65c with water no matter what voltage or clocks. I guess heat could become an issue once you pushed 1.7v or something on normal cooling and tried to load with Prime...but the voltage in itself would be an issue already.

CPUNB can improve memory stability (uncore stability) and make the IMC headroom/CPU performance higher yes, but when it takes you 1.35v from 1.25v to gain 100 MHz it's not really worth it imo...

[tbone8ty]

link to Software optimization guide and BKDG guide tech documents?

please

[charged3800z24]

I'm sorry, but I had no issues with heat on either my Phenom II 965BE, 565BE or 1100T. However I can hit 85c on my FX-8150 with a mere 1.5v vcore and 1.23v CPUNB voltage on water. All of my Phenom IIs loaded no more than ~60-65c with water no matter what voltage or clocks. I guess heat could become an issue once you pushed 1.7v or something on normal cooling and tried to load with Prime...but the voltage in itself would be an issue already.

CPUNB can improve memory stability (uncore stability) and make the IMC headroom/CPU performance higher yes, but when it takes you 1.35v from 1.25v to gain 100 MHz it's not really worth it imo...

What kind of water you using lol.. I never hit that high!

[BeepBeep2]

What kind of water you using lol.. I never hit that high!

My CPU VID is 1.275v. Temps are manageable around 1.4-1.45v...
Swiftech MCR320 with 1850RPM GTs and HeatKiller block. Nothing special but certainly not low end.

link to Software optimization guide and BKDG guide tech documents?

please

http://developer.amd.com/documentati...s/default.aspx

[Daveburt714]

wow, IMC looks good. Can u tets max CPU/NB for superpi test? Because I seen in Taiwan forum aircooled FX with superpi32M and NB over 3000 MHz!

That must be a Golden chip! Sometimes I feel a little out of place on XS. I love to push my system, but I also look for daily stability.
If I can't get it to PrimeBlend (Max Mem) for at least 30 minutes, I get failures when encoding video and thats pretty important to me.

I pushed the IMC tonight (took it all the way to 1.325v) and the best I could get was ~2.55 (nothing too impressive)....

As with all my FX chips IMC scaling slows drastically ~ 2.5Ghz reguardless of voltage, and it does add alot of heat when it gets close to the edge.
My original 8120 would do a little better (~2.6) @ 1.1875, but even then that wasn't the best configuration on water since the heat limited my core clocks...
Things are probably alot different on exteme cold.

Beep: You really pushed your rig all the way up to 85c...
The highest I've ever let mine get was 72c, and as far as I can tell temps come into play (while maintaing stability) ~65...

[wez]

Hmm tried the 1143DPM again yesterday and somehow it seem to be back from the dead we'll see how long it lasts..
Will try some cine/valid runs this weekend for you flanker

[blossa]

I'm thinking about booting up my 8150 on water cooling again. Probably only for one evening this current week due to lack of time... Maybe we can have a small CB-race?

[BeepBeep2]

That must be a Golden chip! Sometimes I feel a little out of place on XS. I love to push my system, but I also look for daily stability.
If I can't get it to PrimeBlend (Max Mem) for at least 30 minutes, I get failures when encoding video and thats pretty important to me.

I pushed the IMC tonight (took it all the way to 1.325v) and the best I could get was ~2.55 (nothing too impressive)....

As with all my FX chips IMC scaling slows drastically ~ 2.5Ghz reguardless of voltage, and it does add alot of heat when it gets close to the edge.
My original 8120 would do a little better (~2.6) @ 1.1875, but even then that wasn't the best configuration on water since the heat limited my core clocks...
Things are probably alot different on exteme cold.

Beep: You really pushed your rig all the way up to 85c...
The highest I've ever let mine get was 72c, and as far as I can tell temps come into play (while maintaing stability) ~65...

The throttling trip is around 90-95c so why not?
IMO Temps always come in to play. Drop your temps 10c and gain 100 MHz. Well, to be honest...about 75-80c is where things start getting wonky and more vcore just means more temps meaning no benefit.

Flank3r had CoreTemp values into the 70s if I remember correctly for Cinebench runs, meaning real CPU temp was probably in 90c range.

@wez
That is hilarious

@blossa
Lets start up the Cinebench 11.5 thread again

[demonkevy666]

The throttling trip is around 90-95c so why not?
IMO Temps always come in to play. Drop your temps 10c and gain 100 MHz. Well, to be honest...about 75-80c is where things start getting wonky and more vcore just means more temps meaning no benefit.

Flank3r had CoreTemp values into the 70s if I remember correctly for Cinebench runs, meaning real CPU temp was probably in 90c range.

@wez
That is hilarious

@blossa
Lets start up the Cinebench 11.5 thread again

drop your temps 10C! >.>
I just did that with My 1090T! and I did get 100mhz
that 4.8ghz bulldozer was only 1% faster then my 4.1ghz 1090T, LOL

[FlanK3r]

Hmm tried the 1143DPM again yesterday and somehow it seem to be back from the dead we'll see how long it lasts..
Will try some cine/valid runs this weekend for you flanker

wtf? How is it possible? I have one my FX dead, so it is here a change get back to life?

[wez]

wtf? How is it possible? I have one my FX dead, so it is here a change get back to life?

All hardware and software is the same, so I really dont know


Blossa, I'll post whatever results I get in the cine thread

[Daveburt714]

The throttling trip is around 90-95c so why not?
IMO Temps always come in to play. Drop your temps 10c and gain 100 MHz. Well, to be honest...about 75-80c is where things start getting wonky and more vcore just means more temps meaning no benefit.

Like I said Beep, I tend to tune things for 24/7 use...
Personally, I wouldn't feel comfortable pushing a chip all the way to the thermal trip.

I like to keep load temps ~65c because (like you said) adding more v's from that point causes temps to shoot up quickly.
Higher temps are fine for pushing benchies, but not so much for everyday use...

That being said...
Next time I try to push a bench, I may grow a pair and let it go a little higher than 72c.


EDIT:

wez: Gratz on your "undead" chip! Believe it or not, I pulled out my dead 1090T the other day and after 5 months it miraculously came back to life too...

ZOMBIE CHIPS.... FTW!

[BeepBeep2]

Like I said Beep, I tend to tune things for 24/7 use...
Personally, I wouldn't feel comfortable pushing a chip all the way to the thermal trip.

I like to keep load temps ~65c because (like you said) adding more v's from that point causes temps to shoot up quickly.
Higher temps are fine for pushing benchies, but not so much for everyday use...

That being said...
Next time I try to push a bench, I may grow a pair and let it go a little higher than 72c.

I'd say 70c is perfectly fine for normal use. I wouldn't totally be comfortable myself running 75-80c on a regular basis, but I don't have many problems pushing limits.

FX is a lot more voltage tolerant that I thought it would be, temps are the big issue on air/water. If it was possible to have load temps in the 50c range I could see running 1.5-1.55v 24/7

[tbone8ty]

Any word on 2012 batches?

[bynext]

hi, my batch is 1137EPM/0037 1.248V first run, with win7 x64.

I need some help to make stable this cpu/nb in 3000mhz, any advice?
And why 2800mhz is kind a hard to get on air ?

thanx

[BeepBeep2]

I need some help to make stable this cpu/nb in 3000mhz, any advice?
And why 2800mhz is kind a hard to get on air ?

thanx

You are joking right?
CPU/NB for Bulldozer is normal in 2400-2600 range. If you've gotten 2800, then congratulations...you've done better than the majority. It isn't the same as Thuban or Deneb.

[bynext]

I just put 1.4v on cpu/nb !! my old 1090t I was usually able to hit 3100mhz easily.
are there some problem to put >1.4v on cpu/nb on air ?
I'm still not used to this cpu..

[BeepBeep2]

I just put 1.4v on cpu/nb !! my old 1090t I was usually able to hit 3100mhz easily.
are there some problem to put >1.4v on cpu/nb on air ?
I'm still not used to this cpu..

Bulldozer has a totally redesigned uncore and IMC. They are not the same, and any MHz comparison is not apples to apples. Also, the CPU is manufactured on 32nm, not 45nm like Phenom II.
1.4v+ long term is dangerous, especially on air/water. Most people run 1.15-1.25v.

Trust me, I could run 3 GHz CPUNB all day on 1100T, however these are not Phenom II based CPUs

[bynext]

Bulldozer has a totally redesigned uncore and IMC. They are not the same, and any MHz comparison is not apples to apples. Also, the CPU is manufactured on 32nm, not 45nm like Phenom II.
1.4v+ long term is dangerous, especially on air/water. Most people run 1.15-1.25v.

Trust me, I could run 3 GHz CPUNB all day on 1100T, however these are not Phenom II based CPUs

thanks !!! next time i'll try less voltage in cpu/nb and try to get max mhz i could.

Just look my new set of memory 1600mhz 1.65v ...
Exceleram vx saturn.



What kind of IC's this set use ?

[bynext]

second try with this set, i don't know about ic this kit use, BUT I LIKE IT.
Yes don't compare with hyper's ic, but it is great.



i've try put lower value on trcd but don't work.

[AMDnord]

my 8120 (2012) died without a reason by just idling on ln2 with 1.75v.
wasnt good piece tho, on air (noctua D14) couldnt pass 49XX cinebench run so i think its better dead :p

i put my 8150 on then and tried again to reach 8GHz, and was successful!

i think i ll wait for PD and go back on denebs for some fun

[BeepBeep2]

second try with this set, i don't know about ic this kit use, BUT I LIKE IT.
Yes don't compare with hyper's ic, but it is great.



i've try put lower value on trcd but don't work.

How about 32M? 1M shows nothing but "I can barely save a screenshot" stability.

ICs are PowerChip (PSC) but what voltage are you using?

[bynext]

I forgot to put asus probe with ''dram voltage'' but is 2556mhz 7-12-8 1.77v.
I took off heatspread and is a PSC ic ....
---------------------------------------------------------------------------------

New problem !!!

why my hyper's kit is kinda worst than my psc ic with BD ?
i have 2 kits of hypers: adata 2200mhz cas8 + patriot sector 5 2000mhz cas8 and i'm getting some trouble with both of them.
2400mhz 9-9-9-27 dont work
2200mhz 7-7-7-21 dont work
2600Mhz 11-11-11 dont work

DOESN'T MATTER USE 1.65V ~ 1.85V..

[wez]

I forgot to put asus probe with ''dram voltage'' but is 2556mhz 7-12-8 1.77v.
I took off heatspread and is a PSC ic ....
---------------------------------------------------------------------------------

New problem !!!

why my hyper's kit is kinda worst than my psc ic with BD ?
i have 2 kits of hypers: adata 2200mhz cas8 + patriot sector 5 2000mhz cas8 and i'm getting some trouble with both of them.
2400mhz 9-9-9-27 dont work
2200mhz 7-7-7-21 dont work
2600Mhz 11-11-11 dont work

DOESN'T MATTER USE 1.65V ~ 1.85V..

You are expecting too much out of your hypers 2000 7-7-7 is already stretching it on most kits for full stability. Also depends on your CPU's IMC.

[tbone8ty]

Just got a FX-8120 with week/batch 1204PGT in today from the egg!

We'll get around to testing this weekend...

It's to nice out! 82F here in CT