More AMD Phenom X4 news inside

[wiak]

We know that in Nov.15 HD 3800 will begin to sell. But we still don't know that RD790 and Phenom X4 will come out earlier(than the AMD plan) or not.

Here is some news that every AMD fanboy would like:

AMD is developing a software called "AMD overdrive". Just like nTune, it can adjust different system values. It's interesting that we find out the software can adjust multiple frequencies of different cores independently.

Personally I think that's good news for people that cares about the total power consumptions of their computers. They can just turn off some cores while they are watching DVDrips.

http://www.expreview.com/news/hard/2007-10...13419d6698.html

In yesterday's “QX6850 VS QX9650 VS PhenomX4” we've already understand what situation Phenom may face. Some of you suggested that maybe let Phenom run some "normal banchmark" to see it's real protential. So now we are sharing the latest Superpi score of Phenom X4.

The test bed is the same one in yesterday, also the same Phenom X4 OC to 3GHz, same 790FX ( Gigabyte MA790-DQ6 is already selling in Guangzhou about CNY2588,about 346USD ), 1Gx2 dual channel memory in 375MHz x2, WinXP. But we adjusted the memory to 3-3-3-9-1T.

At last, the Superpi 1M is 27.531s, and the 30M is 26m28.531s. It is of course better than Athlon 64 x2, but we are using x4 now..

seems Phenom X4 is facing a serious situation.

Anyway, we managed to find out some interesting info about Phenom X4's new funtion. I think after all this bad news about Phenom X4, AMD fanboy finally can get some happy info about it. We are now writing some story about it, pleaese keep Expreview tuned.

http://www.expreview.com/news/hard/2007-10...16099d6667.html

 

[mandelore]

the overdrive looks awesome, but im sorry, 27 seconds is SHIT. seriously... something is not right here at all. my shitty opty 185 and old sandy 4000+ beatsthat score, hell... on ddr1 and my sandy @ 3ghz i got 26 seconds

Edit: why on earth didnt they include ati voltage adjustments also!!! ffs

also, independant voltage and frequencyt adjestments will be sweet for getting the best oc you can.

 

[[I.R.A]_FBi]

it sticky fi di fan bois ...

 

[Deleted member 3]

At last, the Superpi 1M is 27.531s, and the 30M is 26m28.531s. It is of course better than Athlon 64 x2, but we are using x4 now..

Indeed, the 2 extra cores should improve SuperPI 1M scores by at least a million seconds so it'll be calculated before you even clicked the button. Besides that, last time I checked SuperPI is single threaded.

 

[Jimmy 2004]

Lol, even my rig does 36 seconds. I know this has four times as many cores, but given it's at 3GHz and mine is at 2.6, something seems very wrong with that - almost as thought there has been no actual advances other than shoving a few CPU cores onto one processor.

Edit: and that's with Firefox, Firewall, Antivirus, Speedfan, ATITool, CCC, Logitech Software and all of the default Windows services running in the background.

 

[mandelore]

not asking for a ban here or anything, but Dan, alot of your posts in AMD-like threads has been somewhat sarcastic and cheeky? just out of curiosity how old are you??

*runs and hides under a bush

 

[Deleted member 3]

not asking for a ban here or anything, but Dan, alot of your posts in AMD-like threads has been somewhat sarcastic and cheeky? just out of curiosity how old are you??

*runs and hides under a bush

Sarcasm actually is quite an adult kind of humor, how many sarcastic infants do you know? If you read my post you get my point, the news clearly suggests that performance should improve due to the extra cores(see quoted part), Jimmy seems to think it influences the score as well. Fact is SuperPI is single threaded, The Phenom X4 is built from the ground up as quad core CPU, it's a waste of time to run this test. Sure single core performance seems bad here, it might scale very well though, that's the thing they need to test. It's like testing a Ferrari in the hills or a bulldozer on a racetrack.

Plus, if besides my post you'd read my profile you'd know my age as well.

 

[mandelore]

Sorry Dan you misunderstood, I wasnt asking your age for believing you infantile, just as your a senior moderator I was curious as im often surprised that alot of mods here are quite your for how I picture them lol and yes, sarcasm is quite an adult trait.

I understood your post but yes, I still think you come across as a bit cheeky, but hey, im not complaining really, as its not exactly a crime now is it?

Ahh, same age as me then, well, im soon to turn 24 on the 27th December

 

[Deleted member 3]

I don't consider cheeky negative when it's justified.

I believe that when writing about stuff like this you should at least know what you're talking about, specially when writing for a site called "expreview". The "but we are using x4 now.." comment on a single threaded and nearly obsolete benchmark doesn't show that the writer knows what he's writing about. Hence the "cheeky" response.

Besides, the comment was in defense of AMD, so what's with the "alot of your posts in AMD-like threads"? How is AMD relevant here?

 

[Kreij]

I agree with Dan. SuperPi is almost worthless as a benchmark other than for setting OC world records or system stability tests.

I will wait for reviews using real world apps, utilities and games before I make any judgements on the Phenom line of processors.

 

[DRDNA]

lol....you two are the same age
things are not always as they seem! never judge based on vision or hearing...alweays best to close eyes and listen to the inner ears and inner eye but mostly with the heart that fights for purity!

 

[mandelore]

I don't consider cheeky negative when it's justified.

I believe that when writing about stuff like this you should at least know what you're talking about, specially when writing for a site called "expreview". The "but we are using x4 now.." comment on a single threaded and nearly obsolete benchmark doesn't show that the writer knows what he's writing about. Hence the "cheeky" response.

Besides, the comment was in defense of AMD, so what's with the "alot of your posts in AMD-like threads"? How is AMD relevant here?

Point taken, but "alot of your posts in AMD-like threads" comment was refering to just a general feel i got from your posts in other threads. Usually short and slightly amusing, but this is going waaay off topic.

I agree with the single core bench comment. Why dont they use a multicore specific bench like Cinebench or Wprime, thats multithreaded aint it?

 

[DaMulta]

I wanna see some quad Crossfire X and Liquid nitrogen Phenom 3DMark 06 test!!!!


This is sweet

 

[TUngsten]

Pretty cool if you can adjust the multi on individual cores. I'm not sure why one would want to, but it's interesting.

 

[DaMulta]

Pretty cool if you can adjust the multi on individual cores. I'm not sure why one would want to, but it's interesting.

Because not every core will clock the same.

 

[mandelore]

core3 speed is nice, 3.3ghz

 

[Deleted member 3]

Because not every core will clock the same.

Plus nearly no games can use 4 cores, so when gaming you can keep power usage lower by lowering clock of 2 cores. Could be handy if you could set such things in profiles. ie speedstep per core.

 

[DaMulta]

Plus nearly no games can use 4 cores, so when gaming you can keep power usage lower by lowering clock of 2 cores. Could be handy if you could set such things in profiles. ie speedstep per core.

True you could use it for that, but there should be a auto tool for that. With the new 790 boards there should be a setting that will turn cores off and on as you need them.

 

[TUngsten]

Ah yes, I can see the possibilities. Profiles for various tasks could be very interesting indeed.

 

[Mussels]

Plus nearly no games can use 4 cores, so when gaming you can keep power usage lower by lowering clock of 2 cores. Could be handy if you could set such things in profiles. ie speedstep per core.

its more for cool and quiet.

game that uses one core? 3 turn off.

dual/tri core game? the only 2/3 of them stay at full speed. Good power savings, really.


Oh and as for quad core games, lost planet, supreme commander (and expansion), bioshock, hellgate, unreal3, and crysis do - i beleive CoD4 and timeshift will as well.

thems just hte ones on my desktop that i know can use quad, but there certainly is a lot more quad capable games out on the market these days (and while they may not use all 4 100%, it leaves a nice amount of room for multitasking at the same time)

 

[Dandel]

I think this will be a neat CPU to use, although i wonder if AMD will release a Good linux tool that will handle the same as the windows tool on tweaking each core's speed on the fly.

 

[mandelore]

good to let u customise ur cpu in real time for the first time i think. if you got 2 cores that oc insanely compared to the others you could make a dual core with massive oc and access to a quads worth of cache?, esp good if there were indeed profiles you could set

lols, if intel copy this it aint gonna be fun for AMD, imagine a native intel quad with independant voltage control

 

[Kreij]

Besides power savings, there may be other benefits to individual core settings.
For instance, if you are an OC'er, you could potentially shut down two cores to reduce heat output and get a superior OC on the other two (with the same cooling) than you could not if all four were in operation.

Just a thought.

 

[DaMulta]

Or if a program used one croe more than the other core you could just oc that one core that needed it.

 

[Jimmy 2004]

Fact is SuperPI is single threaded, The Phenom X4 is built from the ground up as quad core CPU, it's a waste of time to run this test. Sure single core performance seems bad here, it might scale very well though, that's the thing they need to test. It's like testing a Ferrari in the hills or a bulldozer on a racetrack.

I know SuperPi is single core btw, sorry if my post didn't make that clear

I understand that Phenom may be better at other things and SuperPi might not mean much, but I'm still surprised by a score that low. Like I say, that's less than ten seconds faster than a CPU that's been around for a few years (core vs core) despite my CPU running a few hundred MHz lower, which is somewhat unimpressive to me. We'll have to see some more conclusive benchmarks to make a final decision, but I would still have hoped the architecture would have moved forwards more than this. Surely the gap between a few-year-old P4 and a C2D is bigger than the gap between my s754 CPU and this?