AMD Breaks 8GHz Overclock with Upcoming FX Processor, Sets World Record.

[OBLAMA2009]

michael jackson just died too:

http://www.tmz.com/2009/06/25/michael-jackson-dies-death-dead-cardiac-arrest/

 

[Rifter]

hey man did you know JFK has been SHOT, HOLY CRAP!!!!!!!!!

P.S. the challanger crashed to.

 

[skipsneeky2]

I think the only legit world record AMD has made with the FX processors that no one can argue about is....

The longest delayed and the biggest piece of pass vaporware to date and we are still waiting on some legit desktop benchies which refuse to pop up anywhere.

I'm sure some smartass is gonna post below me that i'm late to the party and they have some link to make me look like a ass but no one can say this isn't the longest delayed computer product in the history of electronics....

I know Duke Nukem Forver prob holds that title but that was like waiting for a turd to push out of a____ 's ass while being constipated <use your imagination with the blank.

 

[Arzachel]

That, also, has been reported before. But it bears repeating, think.

I just wish they mentioned voltage and how many chips they had to try out before they came across one that'd do 5GHz.

None. Sami said that almost every chip they tried hit 5ghz on air, best was 5.5ghz.

 

[TuxDave]

Poor Netburst. Its last bastion of decency has been overtaken.

I have a feeling that someone, somewhere at Intel is trying to dig up the golden P4 chip from the archives to try and reclaim this record.

 

[Khato]

I have a feeling that someone, somewhere at Intel is trying to dig up the golden P4 chip from the archives to try and reclaim this record.

Eh, until AMD proves that they were running more than just the integer core at that frequency, then their 'record' is quite meaningless. The previous 8.3 GHz netburst record was for the entire core logic save for the double-pumped integer pipeline that would've been at about 16.6 GHz.

 

[Idontcare]

Eh, until AMD proves that they were running more than just the integer core at that frequency, then their 'record' is quite meaningless. The previous 8.3 GHz netburst record was for the entire core logic save for the double-pumped integer pipeline that would've been at about 16.6 GHz.

You only get to write the history books so long as you are the reigning champion.

You might be surprised how the definition of "world record clockspeed" comes to be defined if AMD holds onto this one for a year or more.

 

[Nemesis 1]

56x currently...

but pushing 1.7Vcore on the cpu and having it draw almost 300W translates into a TON of amps per pin on 1155pins.

SB-E has 2011 pins to spread that out... so im figuring they will clock better then the 1155, and be more durable.

But performance wise, its not going to be something ur going to lose sleep over not having.

IN this world record . LOL only 1 module was used. None of the AMD fanbois will except this as 8 cores 4 modules is the all everthing . Unless these AMD guys swing back and forth. Really this was only 2 threads . What bench marks were run . I have a $100 says my 2500k would beat it @ 4.5 ghz and my 2600k@5ghz would cream it in Cineabench 11.5 . Any takers?

 

[NostaSeronx]

IN this world record . LOL only 1 module was used. None of the AMD fanbois will except this as 8 cores 4 modules is the all everthing . Unless these AMD guys swing back and forth.

Then point out that H50/H60s/Antec 620s and $50 Performance air Coolers can get you to 5GHz on Bulldozer with the code word "simple"

and then with H70/H80/H100/Antec 920 and $75ish High-end Air Coolers can get you beyond 5GHz

No idea on custom water loops though

But, No facts till release :whiste:

Really this was only 2 threads . What bench marks were run . I have a $100 says my 2500k would beat it @ 4.5 ghz and my 2600k@5ghz would cream it in Cineabench 11.5 . Any takers?

It was done on a dual-core module, no benchmarks were run the OC was a suicide run.

Try turning off HT and Cores and do this you get a burnt chip, supposedly not sure if it was a rumor or fact

And since none of the current gen Intel CPUs have ever busted the Pentium 4 WR clocks having AMD just waltz right over it with a processor with half the stage pipelines(31-stage pipeline vs 16-17) of the Celerons while also having double? the transistors

 

[Nemesis 1]

Ivy Bridge will raise the max multiplier to 63! I wonder if 6Ghz would be possible on air.

http://www.anandtech.com/Gallery/Album/1374#10

Think average 5.3 ghz on air water 5.5 5.6 ghz average. The gaming performance should be very good . If you recall When I told befor SB was released the gaming performance was the best improvement I seen . Yet read ATs first review and what they say about gaming . I almost fell over when I read it was very little improvements . Yet I was seeing great improvement . The other thing I couldn't help but notice was the claim of only 15% improvement over last generation. I still see people making that same false claim . Its better than 20% improvement period .

Now I see this IB will have little to no IPC improvement at same clock . What do you want to bet its more like 15-20% improvement . Less in some benchies and more in others.

 

[NostaSeronx]

Now I see this IB will have little to no IPC improvement at same clock . What do you want to bet its more like 15-20&#37; improvement . Less in some benchies and more in others.

Ivy Bridge can increase the efficiency of IPC possible do to improved sharing algorithms(I think these are what they are called)

Ivy Bridge with the 22nm Tri-gate can allow for a improvement of IPS with improved clock rates and efficiency of IPC^(There are more items)

Ivy Bridge might not show an improvement but it will be a improvement in Performance/Watt/mm^2

 

[Nemesis 1]

Read ATs article on what was revealed about IB . Also AT didn't get all the dope either. Some real surprizes coming.

 

[Nemesis 1]

Then point out that H50/H60s/Antec 620s and $50 Performance air Coolers can get you to 5GHz on Bulldozer with the code word "simple"

and then with H70/H80/H100/Antec 920 and $75ish High-end Air Coolers can get you beyond 5GHz

No idea on custom water loops though

But, No facts till release :whiste:



It was done on a dual-core module, no benchmarks were run the OC was a suicide run.

Try turning off HT and Cores and do this you get a burnt chip, supposedly not sure if it was a rumor or fact

And since none of the current gen Intel CPUs have ever busted the Pentium 4 WR clocks having AMD just waltz right over it with a processor with half the stage pipelines(31-stage pipeline vs 16-17) of the Celerons while also having double? the transistors

As was pointed out many times in the pre 2006 days by AMD fanbois . frequency is nothing and IPC is everthing . My these guys jump around alot! Remember when super pi was the AMD standard for excellance befor 2006 . Than it became a nothing benchmark overnight . All those old topics should be here. Go take a look see . Its so ding dang funny its almost comedy. Just like the 3 stooges its almost comedy!

 

[Idontcare]

As was pointed out many times in the pre 2006 days by AMD fanbois . frequency is nothing and IPC is everthing . My these guys jump around alot! Remember when super pi was the AMD standard for excellance befor 2006 . Than it became a nothing benchmark overnight . All those old topics should be here. Go take a look see . Its so ding dang funny its almost comedy. Just like the 3 stooges its almost comedy!

It is comical. Hype the benches that hype your product, FUD the benches that don't. Be it super-pi or sysmark.

 

[BlueBlazer]

It is comical. Hype the benches that hype your product, FUD the benches that don't. Be it super-pi or sysmark.

wPrime, Cinebench and SPEC may soon join the list, that leaves 3dmark next...

 

[podspi]

Price, performance, power usage.

All that matters.


As I said in another thread a while back, Anandtech should open a dialog with its readers on what apps they actually use, and benchmark those. The only way I could care less about SuperPI runs is if I could finally stop hearing about them

 

[drizek]

I wish someone would benchmark Lightroom already.

 

[sm625]

As long as I know why my chip wins or loses a particular benchmark I dont care. But any new uArch should outperform the old one by 10&#37;, clock for clock per year, else it is a failure. So if you release a new uArch every 6 years, it better beat the old one by 60%. AMD has always done this in the past. Their stock would be a lot lower if BD really had problems more serious than crippled Engineering samples.