Bulldozer Engineering Sample CPU Overclocked to 4.63GHz

Why is everything blanked out? That looks ridiculous.
 
lol pxc, you're as reliable as a clock whenever there is AMD news. If I could figure out a way, I'd set my watch to your posts.
 
Damn, OC one of those and you could get rid of your furnace.
 
NDA. That's my first guess, and I'd assume yours too. Ruffling feathers? :p
Posting that wouldn't be sticking to any NDA in the first place.

The original source post looks real and this "leak" is just an incredibly odd one. I suspect these redacted "leaks" may not be so unofficial. I call monkeyshines that these are independent leaks. ;)
 
Sweet sauce! I hope it proves close to sandy bridge Clock for clock, Just imagine a 8 core sandy bridge.
 
Sweet sauce! I hope it proves close to sandy bridge Clock for clock, Just imagine a 8 core sandy bridge.

If ES were competing with Sandy Bridge AMD would be leaking the information themselves. I hope Bulldozer at least competes with Nehalem and AMD is allowed to re-enter the high end market.
 
It doesn't matter if the performance is crap. We need real performance information and a better idea of what the clocks will look like on retail CPUs.
 
I can't wait for this chip to come out so this retarded dust can settle and people can just buy the better performing CPU.
 
No, because it defies reality. See the CPU-Z thread just posted to understand why.

Pi time is probably edited cause of NDA, he probably doesnt disclose real numbers here, just showing, he has a working chip.

probably real, pi times, no real :p if they are WTF?
 
Sweet sauce! I hope it proves close to sandy bridge Clock for clock, Just imagine a 8 core sandy bridge.

When launched if it's anywhere near Sandy Bridge core for core, clock for clock, I will be scraping my jaw off of the floor.

I'll still probably be getting one, I just have more realistic expectations.

What makes these screenshots any different than the sketchy rumor/ engineering shots we've been seeing for months?

None of this is worthwhile until we have final silicon for a controlled test.
 
if those super pi times are correct which I doubt as the whole thing seems fake that would have bulldozer not even keeping up clock for clock with a core i7 930 which would really suck for amd.
 
It doesn't matter if the performance is crap. We need real performance information and a better idea of what the clocks will look like on retail CPUs.

No kidding. People to often forget that MHz != performance. If it can pull 4+GHz clock/overclocks wonderful, that puts it clock speed wise around Sandy Bridge, and depending on what the stock clocks are might force Intel to bump their some.

However that really isn't the question, the question is performance/clock. That is where we've seen most of the gains on recent CPUs. The new Core i7s really aren't that much faster, clock wise, than the old ones but per clock they have some significant gains. So with BD what will matter is how it holds up to SB per clock. If it is close per clock, then it is good to go. If not, then the MHz won't help it all that much.

MHz are fun, but they aren't what really matters.
 
if those super pi times are correct which I doubt as the whole thing seems fake that would have bulldozer not even keeping up clock for clock with a core i7 930 which would really suck for amd.

I would be surprised if Bulldozer keeps up - clock for clock - with the Core i7-9xx series processors. I highly doubt this will happen.

At best I think we are looking at clock-for-clock and core-for-core almost keeping up with a pre-sandy bridge Core i5, or a Core i7-8xx
 
Zarathustra[H];1037440106 said:
I would be surprised if Bulldozer keeps up - clock for clock - with the Core i7-9xx series processors. I highly doubt this will happen.

At best I think we are looking at clock-for-clock and core-for-core almost keeping up with a pre-sandy bridge Core i5, or a Core i7-8xx

To clarify, based on the information AMD has provided themselves (and a few educated guesses) Bulldozer will likely see a - at best - 12.5% IPC improvement compared to K10.

If it does, it will just barely catch up with pre-SB Core i5's clock for clock.

They are trying to offset this by adding more cores. Long term this will likely work, but right now, its not likely going to do a whole lot for most applications.

Multithreaded Cinebench will look pretty good though :p

That being said, there are some applications and games that are starting to take good advantage of multi-threaded performance. The latest builds of Civ V, combined with the latest nvidia drivers that have added full DX 11 multi-threading support actually load all 6 cores on my Phenom II X6 pretty evenly.

If more software moves in this direction - and long term it most definitely will - Bulldozer will look pretty good. Out of the box - however - with most current and past titles it's not likely to be pretty.
 
That would have alot of the fan boys crying if Bulldozer cannot keep up with Nehalem!

Only those that have been in denial to date.

It's perfectly clear from what AMD have publicly released about Bulldozer that there will not be a huge IPC increase, only a marginal one. Any added performance will have to come from added core count and higher clocks.
 
Zarathustra[H];1037440149 said:
Only those that have been in denial to date.

It's perfectly clear from what AMD have publicly released about Bulldozer that there will not be a huge IPC increase, only a marginal one. Any added performance will have to come from added core count and higher clocks.

Which is telling in itself. If it clocks only about as well as Sandy Bridge or Gulftown then it will still be slower. In cases where more cores come in handy, well things will look different but on the desktop side that will rarely happen outside of encoding.
 
Well...4.5GHz on Sandy Bridge isn't exactly the rarest thing on earth, and I doubt that many that can reach that need over 1.5V to get there. As for 186W... :eek:

Bulldozer's SMT better be absolutely amazing.
 
Which is telling in itself. If it clocks only about as well as Sandy Bridge or Gulftown then it will still be slower. In cases where more cores come in handy, well things will look different but on the desktop side that will rarely happen outside of encoding.

Agreed, for most software today, this is true...

But do me a favor and fire up the latest update of Civ V, with any 275 release Nvidia drivers and you'll see some pretty neat multi-core loading.

Now that the 275 family of drivers are out that supports all of the multithreaded components of DX11, the loading across cores is pretty impressive and leads up to what the developers were talking about as regards their multi-threaded design.

The question is ho long it will take others to follow Firaxis's pretty awesome lead on this front.
 
Did that FX processor just crunch 1 million digits of SuperPI in 1.29 seconds? Am I reading that right? Geez, and I was so proud of getting under 16 seconds with my 955BE. Wow.
 
I really do hope AMD can get back in the game, I miss the Athlon days when they were legitimate competition.(

Unfortunately I don't know that they will ever be a high end competitor. Those days were actually something of a fluke, when you look at it. Basically there were two things that happened that helped them have a competitive chip, neither of them their doing:

1) Intel was sandbagging on Pentium 3 speeds. The P3 could be clocked faster, but wasn't because they didn't need to. So when the Athlon launched, it was really good by comparison. Intel then launched higher clock P3s, of course, but it helped a lot.

2) The Netburst/P4 architecture didn't work out so well. It was a "slow per clock, but lots of clocks" architecture. Did less per MHz than P3 but was supposed to scale to extremely high speeds (they had ALUs testing at 10GHz). Well that didn't work out, meaning it was just a "slow per clock" architecture. That helped the Athlon out.

Basically, Intel stumbled. However they regained their footing with Core and have been strong ever since. The SB is an extremely strong showing.

So historically, AMD has been behind Intel except for the one time Intel tripped up badly. Thus it stands to reason they will still be behind Intel, particularly given Bulldozer's long and problematic development.

I don't think it'll be worthless or anything, but I think they are going to be stuck in the more budget segment they are in currently. That seems to be where they've always been and I don't see anything that makes me think that'll change.
 
I'm not a fanboy of either Intel or AMD, all I hope is that Bulldozer DOES succeed, because a successful Bulldozer will be good for us consumers and competition is good and healthy for the CPU industry, if Intel is the only game in town, then that creates a monopoly and we'll end up with monopoly prices.

AMD chips are great for their stated purpose, good enough performance at a lower cost, I'm honestly not expecting much more than that from Bulldozer in spite of the immense hype, however, I would be pleasantly surprised to find out if it turns out to be more than that.

That would be great , indeed.
 
Did that FX processor just crunch 1 million digits of SuperPI in 1.29 seconds? Am I reading that right? Geez, and I was so proud of getting under 16 seconds with my 955BE. Wow.

No the number behind the 1 is blanked out. so it reads 1x.29 seconds.
 
Zarathustra[H];1037440106 said:
I would be surprised if Bulldozer keeps up - clock for clock - with the Core i7-9xx series processors. I highly doubt this will happen.

At best I think we are looking at clock-for-clock and core-for-core almost keeping up with a pre-sandy bridge Core i5, or a Core i7-8xx

By clock-for-clock do you mean including the HT cores on the 900 series? Because the Lynnfield i5s are great chips, per anandtech's testing they are FASTER than the 900 i7s clock for clock in gaming, not slower:

http://www.anandtech.com/bench/Product/109?vs=47&i=47.48.49.50.59.60.61.62

I have 2 nehalem boxes at work I put together used to take advantage of 24gb of cheap ddr3, but other than that I can't say I've ever seen the hype. If bulldozer runs with lynnfield clock-for-clock (whether we are talking 700s in gaming, or 800s all-around) that means it runs with nehalem.
 
No the number behind the 1 is blanked out. so it reads 1x.29 seconds.

Considering that an overclocked i3 540 can do a sub 10 second 1m Super Pi, Bulldozer looks to be underwhelming (if the screenshot is legit).
 
By clock-for-clock do you mean including the HT cores on the 900 series? Because the Lynnfield i5s are great chips, per anandtech's testing they are FASTER than the 900 i7s clock for clock in gaming, not slower:

http://www.anandtech.com/bench/Product/109?vs=47&i=47.48.49.50.59.60.61.62

I have 2 nehalem boxes at work I put together used to take advantage of 24gb of cheap ddr3, but other than that I can't say I've ever seen the hype. If bulldozer runs with lynnfield clock-for-clock (whether we are talking 700s in gaming, or 800s all-around) that means it runs with nehalem.

Good point but Nehalem is now the "Old Processor" so keeping up with it doesn't look that great.

Most of the AMD guys wants to see something that can keep up with SB or even surpass it.

Regardless of how things turn out most of us are certain that SB-E will hold the current Highend crown and BD probably won't do anything to change that.
 
Guys its 1x.29 seconds not 1.29 seconds....

This just shows how BD sucks..my W3520 can get 11 seconds @ stock I think(Maybe it was @ 3.5ghz). Again its 1x.29 so anywhere from 10-19.29 seconds, not 1.29. Stock sandy gets 10 I think.
 
I call shens!!

It's real.

[ame="http://www.youtube.com/watch?v=utgbRLmKaZg"]YouTube - ‪AMD FX (Bulldozer) + AMD 990FX motherboards RoundUp‬‏[/ame]


Though I want to see real end result silicon, not eng samples.

I don't fanboy for either side, and I'm not holding my breath, but I hope dozer is good. AMD and Intel actually competing = better chips at lower prices for us all.
 
Back
Top