[XS]Just got Phenom II x6 1055T (Retail)

Is there a Thuban 1055T in here somewhere? It was interesting for awhile.
No offense.

Quote:

---

Originally Posted by ajaidev

But if you compare a OCed i5 750 to a i7 860 one can notice how well HT does. It does not behave like real cores the added performance varies a lot, in some case HT can boost pretty high numbers and in others the added performance is little or even negative.

---

But without a 6-core Nehalem not running HT to compare, how can you be sure the performance variations are not simply due to the software? And given that HT in general only benefits between 10-25% in well multi-threaded applications; whose to say it won't be Thuban that suffers more in real world desktop applications which in general aren't that well multi-threaded.

Quote:

---

Originally Posted by daseto

Efficiency and heat output are not really comparable.

---

But they are.
A product with the same performance and lower efficiency has higher heat output.

Quote:

---

Originally Posted by daseto

And when you say energy lost through leakage, there is no such thing as lost energy, energy doesnt disappear, it changes state.

---

We never said it disappeared. we said it is lost as heat (i.e. transformed from electrical energy to heat energy)

Quote:

---

Originally Posted by accord99

But without a 6-core Nehalem not running HT to compare, how can you be sure the performance variations are not simply due to the software?

---

We have 980X which is 6-core.
But it's also at a way higher price point. And that's what's appealing about Thuban.

Quote:

---

Originally Posted by accord99

And given that HT in general only benefits between 10-25% in well multi-threaded applications;

---

Yes, HT is less efficient than actually having more cores.

Quote:

---

Originally Posted by accord99

whose to say it won't be Thuban that suffers more in real world desktop applications which in general aren't that well multi-threaded.

---

Saying that more cores or HT won't benefit single threaded programs is just stating the obvious. People won't be buying 6 core CPU's expecting higher single threaded performance.

In programs that are well threaded, Thuban can close the gap or beat 4 core i7. In single threaded programs, Phenom II was already close enough anyway (except for things like SuperPi)

Seriously why is there so much Intel talk here?

Quote:

---

Originally Posted by audioave10

Is there a Thuban 1055T in here somewhere? It was interesting for awhile.
No offense.

---

now look at what happened to your thread imamage :ROTF:

Quote:

---

Originally Posted by accord99

But without a 6-core Nehalem not running HT to compare, how can you be sure the performance variations are not simply due to the software? And given that HT in general only benefits between 10-25% in well multi-threaded applications; whose to say it won't be Thuban that suffers more in real world desktop applications which in general aren't that well multi-threaded.

---

A 6-core Nehalem would totally break Thubans back even if its not running HT, that comparison is well brutal :D

As of now it seems that Thuban is a equal or bit better than a equally clocked bloomfield with HT on.

As per a measure of scale i think that AMD needs 1.5 cores for 1 of Intel's and if we use this scale you will notice that in Turbo mode AMD has 3 cores at 3.3Ghz that means it can rival 2 nehalem class cores.

But the weak point for AMD's Thuban is single thread execution i expect it to lag behind bloomfields in app's that use just a single thread. Lets hope AMD's turbo mode does good enough for a single core at 3.6Ghz beating a nehalem class single core at 3Ghz "1090T vs i7 930"

Thubans strongest point is three thread execution...

Quote:

---

Originally Posted by OhNoes!

You're talking about two different things. A high leakage chip (leaks power) and therefore runs cooler, whereas the opposite is true for a low leakage chip. It is about efficiency.

Also, low heat dissipation = means a hotter chip; no? The heat generated from current has to go somewhere.

Edit: Did you realize I highlighted a particular portion of your post? Don't try to throw blinds here.

---

no... a high leakage chips loses current.... so it needs more volt to stabilise it ... so it gets it hotter... lower leakage = less heat because they need less voltage to stabilise thus running cooler .....

Quote:

---

Originally Posted by BeepBeep2

High leakage means more power is turned into heat.

High leakage chips are more suited for LN2 because they are leaky transistors and like to use more voltage.

Low leakage is more efficient but less tolerant to voltage and puts out less heat in general.





Will you guys stop argueing and babbling like idiots and acting like you know everything now? I mean everyone.

---

high leakage or low leakage dont matter when you use LN2 or LHE



http://en.wikipedia.org/wiki/Superconductivity low leakage or high leakage wont matter ..... they all have the same resistance ... well if you reach that extreme cold point from what i could understand ....

Quote:

---

Originally Posted by OhNoes!

Contradictions galore! Since you insist on me spelling it out in layman terms, let me oblige.

A high leakage chip is very inefficient at utilizing power (because of the leaks) so at say 1.20v, a high leakage chip will generally run cooler (than the more efficient, low-leakage chip). The low-leakage chip, because it utilizes power more efficiently also generates more heat because you have to realize that: input power = output heat. What you're describing is a process improvement, not a high leakage chip characteristic. The reason why high lekage chips are suitable for ln2 is simply because they can take more voltage; more voltage=more power=more heat <- needs to be cooled by ln2. Makes sense?

---

so fermi should dissipate less heat???? its an extreme high leakage chip.....


btw this doesnt compute.....

Quote:

---

Originally Posted by Titan7171

Seriously why is there so much Intel talk here?

---

maybe they want to steal amd's 15min of fame for the week....

Shame this thread has lost its way, enough has been said now should really just move on :up:

@ajaidev I agree with what you're saying, when the Core i3 2C/4T vs. Athlon II X3 benchmarks showed up it was clear 4C/8T i7 would be matched by Phenom II X6 :)

For some friend I need a test with mental ray or a 3dmax v-ray bench.

Okay folks
Back from work and trying to hit 4GHz
4GHz done
4.2GHz is the current max ( managed to run wPrime 32M and SuperPI 1M)
I guess it's good enough for today :)

I will get the screencap up ASAP, sorry for waiting

P.S. : for other poster asking me to run different bench / chess game with the rig
Please reply after this post
Thanks

Quote:

---

Originally Posted by audioave10

Is there a Thuban 1055T in here somewhere? It was interesting for awhile.
No offense.

---

Sorry
I wish I can post more bench but my work don't allow me to stay up all night
(I wish I can !)

4.2 is very nice, how much Vcore for that? And how does CPU-NB overclock? We've seen Mad222 pushing the CPU-NB to 3150MHz, is yours as good?

No need to apologize, i'm sure everybody will understand you have to go to work, like me and most members here. ;)

imamage: Cinebench 4 GHz please :) and maybe some test stability (LINX or prime???)

Thanks imamage seems 4Ghz air should be achievable by most since 4.2Ghz is max stable :D

This is indeed good news waiting for the screen's F5 getting pressed evey few secs :P

Thanks imamage seems 4Ghz air should be achievable by most since 4.2Ghz is max stable :D

This is indeed good news waiting for the screen's F5 getting pressed evey few secs :P

Eagerly waiting for ur benchmark screenshot:)

2.6ghz 12 opteron cores = 7.95
4.2ghz thunban cores =7.38

difference 0.57 speed does make up for less cores now lol

Quote:

---

Originally Posted by demonkevy666

2.6ghz 12 opteron cores = 7.95
4.2ghz thunban cores =7.38

difference 0.57 speed does make up for less cores now lol

---

More Cheaper too !!!!!

I didn't have much time
Only did these test
4GHz Fritz
http://www.hkepc.com/forum/attachmen...f0bb59e663.jpg

4GHz wPRIME 32M
http://www.hkepc.com/forum/attachmen...f6a605926b.jpg

4.22GHz SuperPI 1M
http://www.hkepc.com/forum/attachmen...0686a3162a.jpg

http://valid.canardpc.com/show_oc.php?id=1134111
4.22GHz CPU-Z Validated

I lost one screencap with 4.2GHz wPrime 32M
The VCore is 1.58V

This CPU surely can replaced my 955BE in my rig :)

Quote:

---

Originally Posted by audioave10

Is there a Thuban 1055T in here somewhere? It was interesting for awhile.
No offense.

---

Quote:

---

Originally Posted by ajaidev

Thanks imamage seems 4Ghz air should be achievable by most since 4.2Ghz is max stable :D

This is indeed good news waiting for the screen's F5 getting pressed evey few secs :P

---

3.9GHz is quite stable for daily usage
But personally I don't like running my rig at 1.5V+ Vcore daily :(

Up the on chip north bridge, wonder how much it would help in cinebench?

uffff, nicw winrar bench
http://img140.imageshack.us/img140/9548/winrar3593.jpg

Yes, waiting for that NB to be pushed higher ;)

Quote:

---

Originally Posted by RaV[666]

IMAMAGE, VERY important question for you.
Can u use higher than stock multipliers up to the end of the turbo range , or just x14 ?
Try setting it manually with and without turbo, use base 200mhz HTT in this test please :).Its very important question for all the peeps that wont buy BE version of this :-)

Try,bios and software like K10stat and AMD Overdrive.
Oh, and second thing, is your bios officialy certified with X6 ? Im wondering because there is high probablity that even the not certified ones could work with X6 no problem, just maybe without turbo feature.

---

I have tried , Max out Ratio in BIOS is x14 , I can't choose X16.5
Haven't try K10Stat / AMD OverDrive yet

Quote:

---

I have tried , Max out Ratio in BIOS is x14 , I can't choose X16.5
Haven't try K10Stat / AMD OverDrive yet

---

Thats sad to hear.However all will be lost when you will try with k10stat with both turbo ON and OFF in the bios.Thx for info!