AMD FX-8130P pushed to 5.1GHz on air

[Djinni]

Bulldozer in sixth gear
A Czech bloke has managed to get hold of a Bulldozer engineering sample and run it through its paces.

He ran a couple of tests and pushed the 3.8GHz FX-8130P to 5.1GHz, on air and with at a core voltage of 1.524V.

The sample was apparently perfectly stable in SuperPi, while it was downcloked to 4.85GHz in Cinebench.

You can check out the screens here.


Source: http://www.fudzilla.com/processors/i...o-51ghz-on-air

[Deleted]

YES!



10char

[dicertification]

Very nice to see.

Thanks for the post.

[BicycleMafioso]

Saw it more than an hour ago (spoiled with computernews by living in Sweden? Mhm)
Still, looks great. And I'm reeeeeeeeaaally happy it does.

I'm also reaaaally looking forward to the "real" bulldozers now, seeing as they will phaze out the first round of models just a few models later. Allegedly because they can't get high stable clocks yet. :P

[Djinni]

Saw it more than an hour ago (spoiled with computernews by living in Sweden? Mhm)
Still, looks great. And I'm reeeeeeeeaaally happy it does.

I'm also reaaaally looking forward to the "real" bulldozers now, seeing as they will phaze out the first round of models just a few models later. Allegedly because they can't get high stable clocks yet. :P

Wasn't awake an hour ago! (At least hadn't got out of bed yet.)

[Jeleh]

I dont know much about clocking, but that's beastly right?

[Deleted]

http://www.youtube.com/watch?v=haV93vh20Q0

youtube vid of it, nothing special on the vid itself. It just shows an asus overclocking tool (i think it is, never seen it before) But the tool shows 8 cores.

Not sure if it's all real or just another fake...
I think there will be more info on july 16th

[OmnitronDefenseSystem]

Imagine their twelve core model being clocked this high... Jesus, that would be crazy awesome.

[Deleted]

I'd be interested in some benchmark scores.

[Asera]

Imagine their twelve core model being clocked this high... Jesus, that would be crazy awesome.

Imagine the TDP...

[Wries]

Lift the NDA and release them already!

[Shaqur]

Be noted, its still enginnering sample. Next revision it will be able to get more on lower voltage
.Cant wait to see it versus i5 2500k, I bet it will fail.

[OmnitronDefenseSystem]

Imagine the TDP...

As long as it isn't over 200 watts i'm fine with it!

[Saithes]

Also keep in mind this is an engineering sample meaning it's essentially a hand picked chip. Also just an FYI... Judging by the length of the bar and the positioning of the preset scores in Cinebench it scored JUST above a Core i7 960... So umm...it's still slow even at 4.8GHz? That means it doesn't come close to Sandybridge

[haxartus]

Also keep in mind this is an engineering sample meaning it's essentially a hand picked chip. Also just an FYI... Judging by the length of the bar and the positioning of the preset scores in Cinebench it scored JUST above a Core i7 960... So umm...it's still slow even at 4.8GHz? That means it doesn't come close to Sandybridge

This looks like the multithreaded benchmark, which means that it's impossible for an 8 core 4.8Ghz CPU to lose from i7-960.
Though the SuperPI result is obviously more than 10 sec, which is bad if the sample is working correctly.

[Saithes]

This looks like the multithreaded benchmark, which means that it's impossible for an 8 core CPU to lose from i7-960.

Yes, it is multithreaded. It still doesn't touch Sandybridge though judging by that heh. Even the i5 2500K clock for clock is still faster =\ As stated judging by the bar length it puts it just over the i7 960 but below AMD's Opteron 12 core. A Core i5 2500K gets 7.75-8.25 @ 4.8GHz while a 2600K @ 4.8GHz gets 9.8-9.9 in it. AMD's Opteron 12 Core gets a 7.95 and since Bulldozer seems to be below that, it means it doesn't touch the 2500K even when they're evenly clocked.

But I guess I should have expected this from AMD since their track record has been nothing short of not being able to match Intel ever at clock for clock or core for core.

Here's my 2600K @ 5.2GHz.

[haxartus]

An 8 core 4.8Ghz CPU can't lose from 4 core CPU on multithreading. If it loses from Sandy either the sample is broken and it's not really 4.8Ghz or the Bulldozer architecture is significantly slower than the current K architecture.

[Saithes]

An 8 core 4.8Ghz CPU can't lose from 4 core CPU on multithreading. If it loses from Sandy either the sample is broken and it's not really 4.8Ghz or the Bulldozer architecture is significantly slower than the current K architecture.

That's what I'm thinking lol I guess it could be that AMD modularized and divided up full cores on the die to each act as a dedicate core (Thus an 8 Core is really a 4 Core with each of its cores given an Int Scheduler). Which could result in lower performance if each core is maxed out processing wise?

So either that screenshot is fake or AMD flopped again. I guess they did also claim that they would beat last gen's i7's if anything.

[vesseblah]

Worthless 'leak' is worthless.

Half of the screenshots in the thread linked are blanked out and censored making it impossible to see what kind of performance you're actually getting with that 5.1GHz, or even score out of Cinebench. Or did people already forget both processor manufacturers stopped the GHz competition nearly decade ago because it was pointless?

[Saithes]

Worthless 'leak' is worthless.

Half of the screenshots in the thread linked are blanked out and censored making it impossible to see what kind of performance you're actually getting with that 5.1GHz, or even score out of Cinebench. Or did people already forget both processor manufacturers stopped the GHz competition nearly decade ago because it was pointless?

You should know its about squeezing performance out of the CPU :P