Ok who pulled the trigger & ordered the 1090-T ?

[choliscott]

Thanks for that. By chance is there a certain temp on here I should pay attention to? So far its saying 32c for each of the cores & TMPin0. Tmpin1 says 39c & Tmpin2 says 42c

You can try Hardware Monitor

 

[richierich1212]

Tmpin1 only, since the individual core temps are bugged.

 

[mv2devnull]

I thought that (previous) Phenom's don't even have per-core sensors?

 

[Markfw]

You got it. Damn, 2.5x > oc'd i7 920 at least in that specific application?!?!?!? :0

I think I now know what my next upgrade is gonna be.

Mark, what motherboard are you using?

Gigabyte MA770T-UD3P

I was wrong on the number. I have since found out that the new smp2 client does NOT want 2 client running at once. I will really use all the cores. So the 920 gets 8700, my Q9450 at 8500, and other quads, but the X6 gets 13k ppd. I find the most fantastic part, that @ 4.1 ghz, and full load on a $25 cooler, and 6 cores, its only at 44c !!!!

 

[BTRY B 529th FA BN]

45nm, 6 cores, 4.1Ghz, 1.4xx voltage and only 44*c at full load!?!?! WOW!!!

 

[Dark4ng3l]

Something has to be wrong with those temp readings or whatever AMD did to modify their process so it would have less leakage has had miraculous success. I can't wait to get mine for my new build... Going to hold out for a month though.

 

[Markfw]

Well, here is currently whats up. Up to 52.

Whats with the 1.6 vcore ? and the 82c temp ?

 

[ZimZum]

XP ?! ewwww

J/K

 

[richierich1212]

Something has to be wrong with those temp readings or whatever AMD did to modify their process so it would have less leakage has had miraculous success. I can't wait to get mine for my new build... Going to hold out for a month though.

There is hardly any leakage on the Thubans. I've been trying to see how high I can overclock while at the same time undervolting. I've managed to hit 3.5GHz / 2.5GHz stable with only 1.216vcore. Pretty amazing.

 

[lopri]

@Mark: 1.70V?! Am I seeing it correctly?

 

[Markfw]

@Mark: 1.70V?! Am I seeing it correctly?

Wow, thanks !

This motherboard sometime detects normal vcore at 1.325 and other times at 1.475. I set the bios at vid+0.175 and if it detected 1.325, then I get 1.5 vcore, 44c load, and all is well.

Is THIS reboot it detected 1,.475 (see coretemp) so I ended up high. So I rebooted it and not its running 42c load and seems fine. I guess that means if I want to try for 4.2 or 4.3 again, I can go that high (it ran 18 hours with 100% load the way you see the screen)

But I kind of like the temps, and that high of a vcore scares me, temps or not, so for now I am leaving it at 4.1 and 1.5 vcore.

13k ppd is fine with me.

 

[manimal]

My board has some serious quirks when it comes to reporting volts. I am hesitant to push it any more so for now its staying at 4.0 2.6 NB all at stock voltage all around. Didnt even have to push any more NB volts trough. When a new bios comes out and stuff seems legit I may go for more. Trying to decide if the Big Typhoon is it for this one or the shiny megahalems I have sitting collecting dust. Doesnt really seem like it needs the bigger cooler though. All in all very happy with my 1090.

Now setting up the htpc with the 1055. Wonder what It can do with the low speed fans and the zalman 160xt case airflow.

 

[Dadofamunky]

Gigabyte MA770T-UD3P

I was wrong on the number. I have since found out that the new smp2 client does NOT want 2 client running at once. I will really use all the cores. So the 920 gets 8700, my Q9450 at 8500, and other quads, but the X6 gets 13k ppd. I find the most fantastic part, that @ 4.1 ghz, and full load on a $25 cooler, and 6 cores, its only at 44c !!!!

That I just cannot believe. What an interesting result too, that shows the Q9450 and the 920 pretty much statistically even. I suspect that if I went i7, I wouldn't see that much of a boost over a Yorkfield for the things I do. (Which is not gaming, by the way.)

 

[richierich1212]

Now setting up the htpc with the 1055. Wonder what It can do with the low speed fans and the zalman 160xt case airflow.

You can undervolt it like crazy. And it'll still run very, very cool. When I was trying to see how low I could run at 3.5GHz, I put it on 1.1975vcore for fun. When posting the system only drew 95w (When using 1.216v, it draws 125v). Pretty crazy, 6 cores only drawing that much at idle (six fans, 2 gentle typhoon 1850rpm, 3 120mm Yate Loon Mediums, 1 140mm Yate Loon Medium, and high vdimm).

 

[manimal]

Running 1055 at 3.2 passively on the scythe ninja out of the case. Temps get a tad uncomfortable priming-low 60s but its stable. May drop a few hundred mhz since its not really necessary for this htpc since its the bedroom unit and only plays 1080P m2ts rips which it does at stock.

 

[choliscott]

Were you able to install this w/o removing the motherboard to install a bracket on the bottom of the MB?

BTW, I have a sunbeam core-contact that you MIGHT be able to do without removing the motherboard. Let me use it Friday, and watch for my temps report. $30 last I saw (usually 40)

 

[Psyrecx]

Ok so since my mobo doesnt have a bios to support this thing yet ive just been looking at and polishing this thing.. its new, what else am I gonna do with it. Is it just mine or do they all come with this wierd layer of white crust arround the sides of it? Kinda looks like somebody got a lil too excited >_>

*edit* nm thats just the metal .. just looks like white crust...

 

[richierich1212]

Yeah sometimes CPUs come with crust.

 

[Markfw]

Were you able to install this w/o removing the motherboard to install a bracket on the bottom of the MB?

I was changing cases, so I didn't try. You MIGHT be able to, depending on your case.

 

[lopri]

That I just cannot believe. What an interesting result too, that shows the Q9450 and the 920 pretty much statistically even. I suspect that if I went i7, I wouldn't see that much of a boost over a Yorkfield for the things I do. (Which is not gaming, by the way.)

I couldn't believe my eyes, either. I first saw some strange Linpack numbers @xbitlabs and thought there was something wrong. It's now more believable with Mark's folding numbers, but I am still wondering what exactly AMD has done.



For comparison, Nehalem @4.0GHz puts out ~60GFlops, and Deneb @4.0GHz does ~50GFlops.

 

[jihe]

I couldn't believe my eyes, either. I first saw some strange Linpack numbers @xbitlabs and thought there was something wrong. It's now more believable with Mark's folding numbers, but I am still wondering what exactly AMD has done.

For comparison, Nehalem @4.0GHz puts out ~60GFlops, and Deneb @4.0GHz does ~50GFlops.

Well 50GFlops with 4 cores scales to 75 GFlops with 6 cores. The Linpack numbers are no surprise.

 

[Accord99]

I couldn't believe my eyes, either. I first saw some strange Linpack numbers @xbitlabs and thought there was something wrong. It's now more believable with Mark's folding numbers, but I am still wondering what exactly AMD has done.

LINPACK is extremely well optimized and since Core 2, K10 and i7 all have a limit of 4 double precision operations per cycle per core, it's just a matter of who ever has the most total clocks.

 

[Accord99]

That I just cannot believe. What an interesting result too, that shows the Q9450 and the 920 pretty much statistically even. I suspect that if I went i7, I wouldn't see that much of a boost over a Yorkfield for the things I do. (Which is not gaming, by the way.)

Those scores aren't representative of what the i7 should be getting in Folding. Later responses in this thread suggest 3.7GHz i7 should be getting well over 15K PPD.

http://forums.anandtech.com/showthread.php?t=2071527

 

[nyker96]

I couldn't believe my eyes, either. I first saw some strange Linpack numbers @xbitlabs and thought there was something wrong. It's now more believable with Mark's folding numbers, but I am still wondering what exactly AMD has done.



For comparison, Nehalem @4.0GHz puts out ~60GFlops, and Deneb @4.0GHz does ~50GFlops.

Well 50GFlops with 4 cores scales to 75 GFlops with 6 cores. The Linpack numbers are no surprise.

and by the same scaling 980X should be at about 90GFlops at about 4ghz vs 75 from Thuban@4ghz.

 

[jihe]

and by the same scaling 980X should be at about 90GFlops at about 4ghz vs 75 from Thuban@4ghz.

I don't think anyone is disputing that the 980X is a stronger cpu.