Phenom II X4 @ +6 GHz by Team Finland

[dread77]

Really good to see 6GHz on retail CPU. Like I said, there shouldn't be any problems hitting those speeds. But of course, let's wait for more results.

What was your max benchable speed?

Well it's pifast stable, that was the only bench I tried yesterday. Today and during weekend I will run every possible test to check stability. We got pretty close to Sampa's record (6,3+) though system crushed and could not validate (4 cores as well).
It was a small session just to check if 6GHz is feasible.

PS. We need better and durable motherboards in order to push those cpu's off limits...

[macci]

Yesterday we made 6GHz+ as well. All four cores, no coldbug issues... Just pure fan!!!

More info here.

Awesome result there!

Yet another Phenom 2 sees the big 6G

[dread77]

Awesome result there!

Yet another Phenom 2 sees the big 6G

Thanks macci! You have been very helpful...

[dinos22]

Yesterday we made 6GHz+ as well. All four cores, no coldbug issues... Just pure fan!!!

More info here.

are you saying you experienced no cold bug of any kind

have you tried running with more than 1GHz HT

[dread77]

are you saying you experienced no cold bug of any kind

have you tried running with more than 1GHz HT

No cb at all!!!
Regarding the high HT Link and coldbug issues, I think that Sampsa already posted an official answer directly from AMD.

[Lastviking]

Well it's pifast stable, that was the only bench I tried yesterday. Today and during weekend I will run every possible test to check stability. We got pretty close to Sampa's record (6,3+) though system crushed and could not validate (4 cores as well).
It was a small session just to check if 6GHz is feasible.

PS. We need better and durable motherboards in order to push those cpu's off limits...

Gj there

I wonder if the Dfi boards can feed beter vcore than the Asus board. Anyone know what max vore is on the dfi boards?(SB750 boards)

[dread77]

Here is Sampsa's post.

I think that most boards (including Asus boards) give max vcore 1,7V. Only Gigabyte DS4H can give more than 1,9V.

[macci]

Apparently 06 is quite sensitive for HT Link frequency as well. Check the graph Tye sent me a few hours ago:

PIC
(All credits go to KTE)

Note, though that the graphic card's power is less than in my setup (HD3850 versus HD4870X2) and that he ran the entire 3DMark01.

the scaling of the entire 3d2k1 is quite different from Nature test only (=lower). Even then 3d2k1 scaled more than 06 when going from 800MHz to 1800MHz (1.1% vs 0.9%).
And yes its a different GFX card, in particular only 256MB Frame buffer (=more chance for PCIE/HT traffic during 06/Nature vs 1GB GFX card).

In general there is definitely a difference between 200MHz HT and 1000MHz HT. 200MHz is not enough while 1000MHz is enough in most cases (excluding IGP scenario). The difference between 1000MHz HT vs 2000MHz HT is much smaller (as HT above 1GHz is usually not a bottleneck).
And like I said we may not be limited to just 1.0GHz HT Link speed (just keep the HT Link multiplier at x5 and tweak the HT ref. clock).

He linked me these as well:

Pic1

pic2

I'd say the difference in GT2 is significant.

This is with Integrated GFX, right? Yes, this scenario definitely responds to HT Link bandwidth as the Video memory is behind HT Link. Would someone run 6GHz LN2 cooled CPU with IGP GFX though?

[dinos22]

No cb at all!!!
Regarding the high HT Link and coldbug issues, I think that Sampsa already posted an official answer directly from AMD.

of course i am aware of that
have you tried to test with higher HT to see where your cold bug is if any

[massman]

the scaling of the entire 3d2k1 is quite different from Nature test only (=lower). Even then 3d2k1 scaled more than 06 when going from 800MHz to 1800MHz (1.1% vs 0.9%).
And yes its a different GFX card, in particular only 256MB Frame buffer (=more chance for PCIE/HT traffic during 06/Nature vs 1GB GFX card).

In general there is definitely a difference between 200MHz HT and 1000MHz HT. 200MHz is not enough while 1000MHz is enough in most cases (excluding IGP scenario). The difference between 1000MHz HT vs 2000MHz HT is much smaller (as HT above 1GHz is usually not a bottleneck).
And like I said we may not be limited to just 1.0GHz HT Link speed (just keep the HT Link multiplier at x5 and tweak the HT ref. clock).

I see what you mean. To be honest, I failed to add a very important sidenote, which is: "Yes, 3D performance decreases when lowering the HT Link frequency, however, the performance gain you obtain because the lack of a coldbug, which makes it possible to clock your CPU frequency much higher, definitly outweights that loss."

I ran some performance scaling tests on the CPU frequency as well and you can see Nature scale quite impressively .

This is with Integrated GFX, right? Yes, this scenario definitely responds to HT Link bandwidth as the Video memory is behind HT Link. Would someone run 6GHz LN2 cooled CPU with IGP GFX though?

People run laptops under LN2, so why not . That's not your point, though.

[starledXpress]

Gj there

I wonder if the Dfi boards can feed beter vcore than the Asus board. Anyone know what max vore is on the dfi boards?(SB750 boards)

Hi...

On DFI LANparty JR 790GX-M2RS i have +0.775 Volts @BIOS, additional + 0.2Volts @AOD
I don't know exactly, but BIOS is OK for Overvolting
http://img143.imageshack.us/my.php?i...dfijr79fe7.jpg

BtW: Can´t wait playing with P!! 940BE...
Really Impressive Clocks!

[Overgloc]

Thanks for the Preview!! Cant wait for Retail!!

[Pt1t]

Here is Sampsa's post.

I think that most boards (including Asus boards) give max vcore 1,7V. Only Gigabyte DS4H can give more than 1,9V.

Biostar boards (TA790GX serie) give also + 0.787V to Vid.

[dread77]

Biostar boards (TA790GX serie) give also + 0.787V to Vid.

Thanks for the info man. It would be good if we could gather info regarding all 790FX & 790GX chipset motherboard which can supply more than 1,7V to the cpu.

[Lastviking]

Biostar boards (TA790GX serie) give also + 0.787V to Vid.

I´m new to AM2/AM2+ boards have only read info on forums...so how much do +0.787v + AOD 0.2v give? is that 1.9v? or hmmm 1.7v hmm?

[Lastviking]

Thanks for the info man. It would be good if we could gather info regarding all 790FX & 790GX chipset motherboard which can supply more than 1,7V to the cpu.

Yes a Faq about AM2+ Phenom II boards is needed....i will try to find info from all good boards and make a compair.... The best would be if evry brand could send a board to me

[Pt1t]

I´m new to AM2/AM2+ boards have only read info on forums...so how much do +0.787v + AOD 0.2v give? is that 1.9v? or hmmm 1.7v hmm?

more than 2V

[Lightman]

Just to let you know guys:
GA-MA790FX-DQ6 - CPU vCore +0.5V in BIOS (Bios F6b)
Memory +0.5V as well
Only downside is CPU NB voltage which can be only lowered form BIOS. To add more vNB I need to use AOD, then it goes all the way up to 1.55V.

Using 1.35V as default would give you 1.85V MAX.
Still decent

[knightwolf654]

dec 20th, were them results?!?!?!

[informal]

dec 20th, were them results?!?!?!

January the 8th-NDA.

[TiTON]

Ok... i am kinda confuse. I am seeing people in this thread saying that AOD gives an additional .2v to vCORE? Is that right?

For me, AOD stops at like 1.5 ~ 1.55.. not sure which.. but I give it much more volts via BIOS. If you look at CPUz.. it is reporting closer to what I set it at. So to hear about a .2v bump in AOD is kinda strange. Can anyone elaborate?

Ton

[charged3800z24]

Ok... i am kinda confuse. I am seeing people in this thread saying that AOD gives an additional .2v to vCORE? Is that right?

For me, AOD stops at like 1.5 ~ 1.55.. not sure which.. but I give it much more volts via BIOS. If you look at CPUz.. it is reporting closer to what I set it at. So to hear about a .2v bump in AOD is kinda strange. Can anyone elaborate?

Ton

On my Gigabyte MA790GP-DS4H, I can set 1.5v Vcore in bios and then in AOD, it says 1.3v VID. I can max it to 1.5v VID which adds .2v to my current 1.5v for a total of 1.7v. This in not the case for my Asus M3A79-T Deluxe board.

[TiTON]

On my Gigabyte MA790GP-DS4H, I can set 1.5v Vcore in bios and then in AOD, it says 1.3v VID. I can max it to 1.5v VID which adds .2v to my current 1.5v for a total of 1.7v. This in not the case for my Asus M3A79-T Deluxe board.

Hmm.. very interesting. I have both boards, so i have to check that out. Thanks for the info

Ton

[leoy]

Unfortunately the asus reads the value in the bios, and goes all the way up to 1.55. Nothing more.

[dread77]

6221....