Rumour: Bulldozer 50% Faster than Core i7 and Phenom II.

[hv43082]

What do you do with your computer that you would notice so much of a difference to call it a "spanking"...honestly? Encoding? Rendering? Something else? Benchmarks like superpi or something like that don't count.

I think more people would notice a bigger difference between a SSD and a regular HDD before they notice the difference between 2 modern CPUs.

There is still no arguing at $200, the 2500K is better than the 1100T.

 

[thilanliyan]

There is still no arguing at $200, the 2500K is better than the 1100T.

For the same price yeah you're right but I maintain that most people probably wouldn't notice the difference day to day.

 

[996GT2]

For the same price yeah you're right but I maintain that most people probably wouldn't notice the difference day to day.

If you just use your computer to surf the web and type documents, then your statement is certainly true. However, once you get into more intensive tasks I think it wouldn't be too had to notice the performance difference. For example, in Civ5 the difference between a 2500K and 1100T could mean the difference between playable and laggy performance.

41.2 vs 26.4 FPS. The i5 is 56% faster here.

Or, if you work in Photoshop a lot (which I do), I'm sure you can appreaciate the performance difference between the 2500K and X6 1100T. To me, this is a very realistic measure of performance.

All of this is without overclocking in the equation. You can probably squeeze 4 GHz out of an 1100T with extra voltage, whereas you can easily get a 2500K up to 4.7 or 4.8 at reasonable voltages. That 7-800 MHz clock speed difference would only further swing things in favor of SB.

 

[drizek]

Yes.

Now consider the fact that the 2500K is not $200 but rather $180 at MC. The PII X6 is just ridiculously overpriced for the performance it offers.

 

[Anexate]

Or, if you work in Photoshop a lot (which I do), I'm sure you can appreaciate the performance difference between the 2500K and X6 1100T. To me, this is a very realistic measure of performance.

I'm really hardly awaiting few real world tests with the new Bulldozer.
What worries me is AMD talking about an 8core as a performance part for desktops; I somehow got the feeling that an 8 or 6 core are being put "against" 4 core SB.
Such a situation sounds good only for some "benchmarks", certainly not for games or other broadly used consumer applications.

Drizek:
Amd Phenom II X6 1100T € 186,78
Intel Core i5-2500K € 194,27
(European prices)

AMD has lost ground even at the affordable builds; such builds had an AMD CPU for every single buyers guide (in memory) until very recently. Now the Intel G6950 is just better than the Athlon II X3 when overclocked, and the Gigabyte H55 board is excellent and cheaper than ever.

One does not know what to tell when one sees such videos:
http://www.youtube.com/watch?v=kJ5abEXkZdI

 

[thilanliyan]

If you just use your computer to surf the web and type documents, then your statement is certainly true. However, once you get into more intensive tasks I think it wouldn't be too had to notice the performance difference. For example, in Civ5 the difference between a 2500K and 1100T could mean the difference between playable and laggy performance.

41.2 vs 26.4 FPS. The i5 is 56% faster here.

Or, if you work in Photoshop a lot (which I do), I'm sure you can appreaciate the performance difference between the 2500K and X6 1100T. To me, this is a very realistic measure of performance.

In terms of games, that was ONE test (out of several that didn't show a lot of difference) at 1680x1050 using a GTX580...how likely is that scenario (ie. someone using a GTX580 for that res)?

Working with media (and the x6 is not so bad with video encoding) is where you will likely see a noticeable improvement and I fully agree with that.

I'm not trying to be thick here...I fully understand, if you are building a brand new system it is probably better to go for a SB (that's the route I'd take). If you already have maybe an AM2+/AM3 system, and looking for an upgrade, one of the X6 CPUs might be a better bet. I just think in most circumstances most people wouldn't be able to tell the difference.

 

[drizek]

That civ 5 benchmark, comparing the 1100 to the 980X, shows that Intel has a 75% advantage clock for clock.

AMD better hope they can deliver 50% better performance per clock per core in BD.

When a 6-core AMD is having a hard time keeping up with a dual core i3-2100K running 200MHz slower, they've got a pretty serious problem.

Edit: thinlaniyan, look at the actual FPS. It's only getting 26fps. You're going to be CPU-limited in Civ5 regardless of video card if all it can do is 26FPS. Sure, maybe you can OC it to hit 30FPS, but should that really be necessary?

 

[wahdangun]

wtf people we are not discussing intel here. please don't make it amd vs intel thread.


btw

 

[Anexate]

wtf people we are not discussing intel here. please don't make it amd vs intel thread.

We are not going into Intel/AMD battle.
We are just worried about Bulldozer. If it does not deliver, we are all going to be in huge problems; Intel is already a monopolist and the prices are going to spiral even further into stratosphere.
Not to mention that without stimuli, the tech progress gets slowed down.

 

[Riek]

We are not going into Intel/AMD battle.
We are just worried about Bulldozer. If it does not deliver, we are all going to be in huge problems; Intel is already a monopolist and the prices are going to spiral even further into stratosphere.
Not to mention that without stimuli, the tech progress gets slowed down.

and? Like others said, AMD atm does not get the best performance, but the difference is in 90% of the cases not noticeable. So if AMD will need to sell its cpu's at extremely low prices, that is also good for 90% of the customers who don't need the 20 time gain on programs they never use.

The most important thing for BD is that it is a capable future proof platform which they can use for their future generations. (just like Bobcat is intented).

 

[cusideabelincoln]

He wrote, quite litterally that the socket 1156 "surpassed" socket 1366.
That is a lie...
(Hell, even for socket 1155 it would be stretch)
Thanks for playing...
You are the weakest link....
Good bye.

You take a single statement literally and fail to put in context with everything else he said. That is a huge fail. The point went entirely over your head, but you only saw what you wanted to see so you had to swoop in and defend, and make something out of nothing (red herring).

I'm not playing a game here, you are. And it's making you look like a fool. A bitter, arrogant, and incorrect fool. But keep it up, because it's also humorous how blind and hypocritical you are.

Now literal is not how you should read his post. His criticism of 1366 is valid. 1366 is a fine platform, but I'm sure people were expecting better support and some kind of processor upgrade path. But the only upgrade path is a super-expensive Gulftown. No die shrink, higher clocked, added cache, tweaked Bloomfield chips. No mildly affordable six-core Gulftown processors. Instead Intel released speed-bumped Bloomfield processors (930 and 950), which aren't really upgrade paths from a 920 and focused their efforts on Lynnfield and Sandy Bridge.

 

[Martimus]

There is still no arguing at $200, the 2500K is better than the 1100T.

I don't know about that. The 890FX MBs are both cheaper, and have better features than the P67 chipset motherboards. The cost of ownership of the 2500K is still quite a bit higher than the 1100T, just because Intel charges more for their chipset, which is lacking in features in comparison to the AMD equivalent.

 

[996GT2]

I don't know about that. The 890FX MBs are both cheaper, and have better features than the P67 chipset motherboards. The cost of ownership of the 2500K is still quite a bit higher than the 1100T, just because Intel charges more for their chipset, which is lacking in features in comparison to the AMD equivalent.

What features are the P67 boards lacking?

 

[Castiel]

Thuban would have been killer if Amd actually had a decent IPC. Thats why bulldozer doesn't impress me at all. Yeah 8 cores are sweet but the IPC is still putrid if the 12% improvement over Thuban is accurate.

 

[Martimus]

What features are the P67 boards lacking?

The P67 chipset has 2 SATA 3.0 ports
The 890FX chipset has 6 SATA 3.0 Ports
The P67 chipset has 8 PCIe 2.0 lanes
The 890FX chipset has 44 PCIe 2.0 lanes

USB ports are the same between the two.

P67 features:http://www.anandtech.com/show/4080
890FX features:http://www.anandtech.com/show/3674/amds-sixcore-phenom-ii-x6-1090t-1055t-reviewed/4

What gets to me is that the AMD chipset has more features, but the intel chipset is more expensive. In fact it is much more expensive, which is the biggest reason I didn't buy a SB computer before the bug. The bug is forcing me to wait even longer, although I am chomping on the bit a little to upgrade. I still want the additional performance of the SB processors, but I don't want to have to eat the lower features of the board, especially at higher prices.

 

[Anexate]

You take a single statement literally and fail to put in context with everything else he said. That is a huge fail. The point went entirely over your head, but you only saw what you wanted to see so you had to swoop in and defend, and make something out of nothing (red herring).

But keep it up, because it's also humorous how blind and hypocritical you are.

Actually, I think I discovered what is afoot. He probaly uses Google translate.
Look at the outcome:
:whiste: :whiste: :whiste:


Socket 1366 was the high-end, socket 1156, the main stream.
Socket 155 is mainstream, socket 2001 Will be high end.

Not possible there is something for Socket 1156 (1336 socket a killer) ... just like you should Socket 1155 inn is not something then it is not (high-end socket is 2011)

So yes, his arguments and Ditt innlegg forferdelig ... and not part of the real world.

He wrote literally to 1156 "overgått" socket 1366th touch
It is situated on a ...
(Hell, even for socket 1155 would be prank)
Takk for a play ...
You are svakeste stage ....
Good bye.

You have no idea. but chooses to have this law?
. oO (Madre Mia)

* presses ignore Botton *

 

[grimpr]

If you just use your computer to surf the web and type documents, then your statement is certainly true. However, once you get into more intensive tasks I think it wouldn't be too had to notice the performance difference. For example, in Civ5 the difference between a 2500K and 1100T could mean the difference between playable and laggy performance.

41.2 vs 26.4 FPS. The i5 is 56% faster here.

Or, if you work in Photoshop a lot (which I do), I'm sure you can appreaciate the performance difference between the 2500K and X6 1100T. To me, this is a very realistic measure of performance.

All of this is without overclocking in the equation. You can probably squeeze 4 GHz out of an 1100T with extra voltage, whereas you can easily get a 2500K up to 4.7 or 4.8 at reasonable voltages. That 7-800 MHz clock speed difference would only further swing things in favor of SB.

This means that AMDs CPUs are not suitable for Intel optimized games and Adobe programs, well, who could have guessed....:biggrin:

http://software.intel.com/sites/billboard/va-magazine/issue-07/articles/civ5/
http://software.intel.com/en-us/videos/adobe-creative-suite-4-and-intel-core-i7-processors/

 

[Voo]

This means that AMDs CPUs are not suitable for Intel optimized games and Adobe programs, well, who could have guessed....:biggrin:

No they're just worse for everything that needs high single threaded performance and doesn't profit from lots of cores.

 

[Idontcare]

Thuban would have been killer if Amd actually had a decent IPC. Thats why bulldozer doesn't impress me at all. Yeah 8 cores are sweet but the IPC is still putrid if the 12% improvement over Thuban is accurate.

Like P4 vs. K8 all over again, only the roles are reversed.

BD will be sweet even if it has putrid IPC provided the clockspeeds are demonic while power-consumption and prices are angelic.

Price/performance and performance/watt...and performance is app-dependent so what is good for one person and their apps will not be good for another person.

Now what AMD really needs is an even higher turnover rate in their executive offices...only then will my confidence in BD start to rise. I mean look at how awesome Larrabee turned out to be, and Pat Gelsinger only left 6 months before it was to be released...oh wait, I got that backwards don't I...

 

[Stoneburner]

I mean look at how awesome Larrabee turned out to be, and Pat Gelsinger only left 6 months before it was to be released...oh wait, I got that backwards don't I...

HA! That's just mean

 

[Voo]

I mean look at how awesome Larrabee turned out to be, and Pat Gelsinger only left 6 months before it was to be released...oh wait, I got that backwards don't I...

Can you think of even one successful high profile launch where several executives left a company a few months earlier? That's an honest question, I've got no idea, but it sounds pretty strange from a layman's perspective

 

[Anexate]

Can you think of even one successful high profile launch where several executives left a company a few months earlier? That's an honest question, I've got no idea, but it sounds pretty strange from a layman's perspective

Of course we can:

year 2007; AMD

AMD Faces Exodus of Executives Ahead of Major Product Launch
AMD Loses Executives as Quad-Core Processor Launch Looms

August AMD announced that its chief sales and marketing officer Henri Richard will leave the company after spending five years with it. Mr. Richard joined AMD on the 10th of April, 2002, months ahead of the company’s failure to release its highly-anticipated 64-bit processors on time and a period dramatically gloomy sales of AMD Athlon XP processors, which at the time could not compete against Intel’s Pentium 4.

On the last day of Summer several media outlets reported that Rick Hegberg will also resign from AMD shortly. Mr. Hegberg came to AMD from ATI Technologies, where he has been serving as senior vice president of world wide sales since 2003. The executive was responsible for the overall management of ATI’s world wide sales organization, as well as for achieving the company’s global sales revenue and customer satisfaction objectives among large customers and partners.

The resignation of Henri Richard and Rick Hegberg comes one month after former ATI chief executive David Orton left the company in July, weeks after Chris Evenden, the former public relations chief at ATI, stepped down.

David Orton, is among numerous of ex-ATI executives as well as specialists, who have left Advanced Micro Devices in the recent quarters. Richard R. Heye, a vice president of ATI, left AMD late last year, ironically, just three years after he left AMD for ATI; Paul Dal Santo also left AMD as it absorbed ATI; whereas Peter Edinger, another vice president of ATI, departed from AMD early this year. In addition, a number of public relations persons have left AMD since the two companies became one.

According to official statements, all the leaving AMD employees resign “because of the personal reasons” and it is a natural process since “people come and go”. However, there are no such huge flee of personnel from companies like Intel Corp. and Nvidia.

 

[Castiel]

Like P4 vs. K8 all over again, only the roles are reversed.

BD will be sweet even if it has putrid IPC provided the clockspeeds are demonic while power-consumption and prices are angelic.

Price/performance and performance/watt...and performance is app-dependent so what is good for one person and their apps will not be good for another person.

Now what AMD really needs is an even higher turnover rate in their executive offices...only then will my confidence in BD start to rise. I mean look at how awesome Larrabee turned out to be, and Pat Gelsinger only left 6 months before it was to be released...oh wait, I got that backwards don't I...

I want BD to be awesome because i want lower 2011 prices 20 megs of L3 cache and quad channel memory....

 

[podspi]

It certainly doesn't instill a great amount of confidence... I am of the mind that they are getting ready for a sale and Meyer wasn't cooperating. Say what you will about the mobile segment, Meyer not only kept the company afloat, but the product pipeline full, (fuller than its ever really been).


On the GPU side, the last two major series were pretty much home-runs (especially given the process issues that were out of their control) so its difficult to see why Meyer would have been ousted at all.



Sure, AMD could have focused more on the Tablet/Smartphone, but at the expense of what? As Intel is finding out, this isn't an easy market to just break into, it would have been foolhardy to bet the farm on the mobile sector when there is a ton of growth potential in mainstream and server x86 CPUs.

 

[grimpr]

No they're just worse for everything that needs high single threaded performance and doesn't profit from lots of cores.

Lets say, like the new Internet Explore 9 RC; :biggrin:

http://www.beautyoftheweb.com/#/highlights/all-around-fast