AMD Shanghai/Deneb Review Thread

[Oliverda]

Well, you're running 64bit Cinebench. Check.

Unfortunately I can't run 64bit Cinebench on my 32bit Win XP SP3.

[Zucker2k]

Unfortunately I can't run 64bit Cinebench on my 32bit Win XP SP3.

Oh that screenie isn't yours? Post a screenie of your 32bit XP SP3 run.

[demonkevy666]

@Ghostbuster

I just checked some Q9xxx 3dmark06 CPU scores at the ORB.It looks like Phenom II and Q9x50 are pretty much even at the same clocks in this test
PhII 4Ghz : 5846
Q9550 @4Ghz : 5766 (Phenom wins by a small margin-80pts which is within margin of error-so they are the same per clock)
Q9650 @ 4.05Ghz : 6033 (Q9650 per clock wins by 1.1%,within margin of error).

I say again all these Phenom II previews are just that-previews.We should wait for final reviews with proper bios versions and final platform to make up our mind where it sits exactly vs Q9xxx series.By the looks of the latest info it's quite competitive ,both per clock and in OCing abilities.

1.8ghz nb lol 1.1% 200mhz Nb/L3 = what like % ? 0.7% ( Idk)

[informal]

1.8ghz nb lol 1.1% 200mhz Nb/L3 = what like % ? 0.7% ( Idk)

"Q9650 per clock wins by 1.1%,within margin of error" means that instead of 4.05Ghz ,the score for Q9650 at 4Ghz would be ~5958pts which is 5958/5846=~1.9% better than Phenom II at the same 4Ghz clock (I made a mistake,wrote 1.1% vs 1.9% -still practically the same score). The scaling with clock is 97% for Phenom II,going from 3Ghz to 4Ghz in CPU subtest of 3dmar06.The L3 speed seems to play less of a role here.

[Caveman787]

The one site with benches supposedly had near perfect scaling with clock throughout there tests.

That might sound odd but AMD did have overclockers in mind when making these chips.

[demonkevy666]

"Q9650 per clock wins by 1.1%,within margin of error" means that instead of 4.05Ghz ,the score for Q9650 at 4Ghz would be ~5958pts which is 5958/5846=~1.9% better than Phenom II at the same 4Ghz clock (I made a mistake,wrote 1.1% vs 1.9% -still practically the same score). The scaling with clock is 97% for Phenom II,going from 3Ghz to 4Ghz in CPU subtest of 3dmar06.The L3 speed seems to play less of a role here.

http://www.xbitlabs.com/articles/cpu...x4-9850_4.html

In there 0.40-3% with 200mhz NB. so going up on NB from 1.8...2.6ghz max for AM+...3.2ghz for AM3...

[charged3800z24]

Oh that screenie isn't yours? Post a screenie of your 32bit XP SP3 run.

well here is my 3.3ghz run on 64bit and 32bit ...system is tweaked a lil and still lower then the Phenom II @ 3.2ghz.
(64)

(32)

[Ghostbuster]

I just checked some Q9xxx 3dmark06 CPU scores at the ORB.It looks like Phenom II and Q9x50 are pretty much even at the same clocks in this test
PhII 4Ghz : 5846
Q9550 @4Ghz : 5766 (Phenom wins by a small margin-80pts which is within margin of error-so they are the same per clock)
Q9650 @ 4.05Ghz : 6033 (Q9650 per clock wins by 1.1%,within margin of error).

Yours is slightly flawed because that was on Windows XP 64-bit. The reference 4GHz Phenom was on Windows XP 32-bit. I can find better scores then that, for example Q9550 @ 3.739GHz scores 6013 on Windows XP 32-bit

I say again all these Phenom II previews are just that-previews.We should wait for final reviews with proper bios versions and final platform to make up our mind where it sits exactly vs Q9xxx series.By the looks of the latest info it's quite competitive ,both per clock and in OCing abilities.

Anyway, I've made my prediction many moons ago.. As for the finale vs Q9xxx, I've already saw early on that it will be behind Q9550 as all clues are pointing towards that. I compare with stock clocks, and again all results pointing to that conclusion again. Well, now you can predict the pricing...

[roofsniper]

there are no clues since there are no benchmarks that can be trusted. this entire thing is still up in there air. no one really knows how it will perform against intel cpus but we know its at least better than the first phenom. i would like to keep the whole intel thing out of this until official benchmarks are released since it will only start an argument that can't be ended since there is no proof of anything. and you really can't compare benchmarks from different sources together since all the parts aren't the same and the systems are not tweaked the same.

[Ghostbuster]

Update... Q9550 @4.038GHz CPU score = 6316 on Windows XP 32-bit This is the nearest score I can find, almost the same GPU scores, also its a HD4870...

[AbelJemka]

Yours is slightly flawed because that was on Windows XP 64-bit. The reference 4GHz Phenom was on Windows XP 32-bit. I can find better scores then that, for example Q9550 @ 3.739GHz scores 6013 on Windows XP 32-bit

Anyway, I've made my prediction many moons ago.. As for the finale vs Q9xxx, I've already saw early on that it will be behind Q9550 as all clues are pointing towards that. I compare with stock clocks, and again all results pointing to that conclusion again. Well, now you can predict the pricing...

Phenom II already scoring 5846 at 4GHzwith a beta bios on non optimized system (400MHz memory cas 5) is for me on the same range as Q9xxx scoring around 6000 at +3,8GHz

Update... Q9550 @4.038GHz CPU score = 6316 on Windows XP 32-bit This is the nearest score I can find, almost the same GPU scores, also its a HD4870...

Heavily optmized system versus unoptimized system with beta bios. I don't know what u search to prove?

[Ghostbuster]

Phenom II already scoring 5846 at 4GHzwith a beta bios on non optimized system (400MHz memory cas 5) is for me on the same range as Q9xxx scoring around 6000 at +3,8GHz

Now what's left is how well it can run in air or water cooling in normal setups, then we will see how competitive it can be...

[AbelJemka]

I see quite a bunch of overclocking with normal setup on aircooling already but i prefer to wait for final review.

[Ghostbuster]

Heavily optmized system versus unoptimized system with beta bios. I don't know what u search to prove?

The 4GHz Phenom II system was using ASUS HD4870 Toxic.. so that is pre-factory overclocked for sure. Thus I look for a similar card (thus similar GPU scores) and a similar clocked Q9550 processor.

[roofsniper]

Now what's left is how well it can run in air or water cooling in normal setups, then we will see how competitive it can be...

there have been plenty of times that it has been on an air or water cooling setup. it is a known fact that it can get 4ghz on air and 6 ghz on ln2 so most of the time when you are seeing a 4ghz phenom II it is on air.

[Caveman787]

So anybody know how well its gonna perform gaming wise?

Apparantly i7 is better for games then core2 because it gets better minimum framerates.

At one of the events in the video they said the minimum framerates were doubled.

Can anyone verify them being the good?

Like everyone else I can't wait for more benches and details other than ln2 since I'm not gonna be running more than air.

[roofsniper]

So anybody know how well its gonna perform gaming wise?

Apparantly i7 is better for games then core2 because it gets better minimum framerates.

At one of the events in the video they said the minimum framerates were doubled.

Can anyone verify them being the good?

Like everyone else I can't wait for more benches and details other than ln2 since I'm not gonna be running more than air.

well i have always thought that a phenom was smoother for gaming than a core 2 was in the first place even tho the core 2 got higher frame rates. this could be the fact that the phenom keeps the fps being about the same instead of having it go from high to medium and from high to medium. i can't say how it will be for gaming for sure but im sure it will be amazing.

[Morais]

oh no, not again

[mongoled]

Chad Boga is temporarily banned. This is not the place to troll and flamebait. Keep AMD vs Intel discussions or "AMD can't beat Intel" out of this thread.

Please stay on topic.

How hard is it for peeps to stay on topic????

Its quite clear what is happening, AMD peeps are trying to keep on topic, then someone posts info regarding Intel CPU v the Phenom II CPU's and the AMD peeps have to counter.

Now.......the peeps posting the Intel info are consistant, one of them has already been banned, I will be contacting Ashraf directly if the other peeps (you know who you are, the usual suspects) continue in trying to derig this thread.

There are other threads making comparissons between PII and Intel CPU's

GO AND POST ON THOSE

Leave this thread to be as title says

'AMD Shanghai/Deneb Review Thread'

PI vs PII comparissons very welcome as they are both AMD and relevant

PI vs PII vs Intel counterparts

Unwelcome........

-- ON TOPIC --

Peeps posting info regarding PI vs PII, try to post comparissons using similar OS's, been many people posting 32bit vs 64bit scores and not initially realising this. In this way a more realistic picture can be gained of what to expect.

Also, this is Xtreme Systems, we are ofcourse happy to see default clocks vs default clocks. But its not like we are going to be running our CPU's at default clocks, LOL

I am very interested in scaling of NB clocks as are many with CPU speeds, this will be the rabbit in AMD tweakers hats to getting better scores. Hopefully the info regarding PII reviews will increase.....

Final words

PLEASE STAY ON TOPIC

[LC_Nab]

The 4GHz Phenom II system was using ASUS HD4870 Toxic.. so that is pre-factory overclocked for sure. Thus I look for a similar card (thus similar GPU scores) and a similar clocked Q9550 processor.

Like we are all saying the phenom II is runing on beta bios so im pretty sure it will perform even better with the true bios and mostly on the true am3 boards . Time will tell

[Pt1t]

Here is my 5.6 http://valid.canardpc.com/show_oc.php?id=466799

it 's -195° ?

[amdcian]

Another validation of our boy, this time is 6.1GHz , a tiny pity...2 core are disabled

Anyway, thank dread77 for this

[AbelJemka]

The 4GHz Phenom II system was using ASUS HD4870 Toxic.. so that is pre-factory overclocked for sure. Thus I look for a similar card (thus similar GPU scores) and a similar clocked Q9550 processor.

Time wasting for you!
On Orb it's too hard to know if videocard are overclocked, so busting such score will indicate a lot but so little in the same time
Same for CPU score with 06, so influenced by subsystem optimization is hard to compare.

[ajaidev]

Hey i was looking through coolices benchs and i saw something really nice. In 3dmark 06 the cpu score is 4483 and i know the cpu score for i7 920 was 4609. So not that much difference!!

Also at stock Wprimes was done in 13 secs on a Ph II while it took 10 on an i7

[SoF]

it 's -195° ?

no it's -155°C 1.85V - A79A-S can't run HT Link 1.0 yet but with 6x multi and 1320 MHZ I was doing not too bad I think. with 1.0 and 5x there should be room for maybe 6gig validation my friends from benchbrothers could run both of their chips -195°C with 1.0 on Destroyer and Gigabyte board.