Official Improvements of Piledriver Cores.

[Chiropteran]

when using the same number of cores.

Ding ding. That is where the logic falls apart, since the number of cores isn't the same in reality.

 

[LOL_Wut_Axel]

I will probably be scratching my head on this in the morning. I am not an engineer and I don't plan on being one because CPU architecture has always been my weak point of knowledge. I tried at least. D:


I am naturally a software guy, thats what I went to school for.

Don't worry, LOL. You're still two steps above the vast majority of computer users when it comes to hardware knowledge. Very few people know the very specifics of what makes CPU architectures faster.

 

[BallaTheFeared]

I will probably be scratching my head on this in the morning. I am not an engineer and I don't plan on being one because CPU architecture has always been my weak point of knowledge. I tried at least. D:


I am naturally a software guy, thats what I went to school for.

Don't worry some people who claimed to know things went over the docs for bulldozer up and down for months telling people IPC went up.

 

[LOL_Wut_Axel]

Ding ding. That is where the logic falls apart, since the number of cores isn't the same in reality.

You do realize that that makes absolutely no sense, right?

Both Core 2 Quad and Phenom II X4 have four Int and four FP units. Neither has SMT or CMT.

 

[386DX]

I think Rory is smart, and will do the company extremely well in the long run, but I don't believe for one bit that AMD isn't vigorously competing with Intel on the CPU design front at least internally. At the end of the day, they still have to design a CPU that's reasonably competitive with Intel. Although Intel has a long ways to go to catch up to AMD on the GPU front, they have mighty deep pockets, and given their track record in GPU design over the last 3 years, seem to have quite the focus as well.

I think in the next few years you might begin to see AMD's competitive advantage in the iGPU market begin to shrink as Intel improves on their design. Although Haswell is still a ways away, Excavator is going to have to be quite good in order to compete.

It's crazy scary how fast Intel has caught up to AMD on the iGPU front (mobile side, not desktop). As Anandtech's mobile Ivy review showed Intel is on par with Llano on the mobile side right now. So in a year an a half Intel went from the laughing stock of the integrated graphics to about half a generation behind AMD (based on mobile Trinity being out soon). While AMD's Trinity is about 4-5 generations behind Intel on the CPU side (Ivy, Sandy, Westmere, Nehalem, Yorkfield). I'm going to say by the time Haswell is out AMD is going to be on the losing end of both the mobile CPU and iGPU market. AMD is going to officially changed the name of Steamroller to Cleveland Steamer, and there new competitor is going to be VIA in the lucrative "what's the cheapest sh*t you can put behind a computer screen for a kiosk" market... okay maybe not that far but you get the point.

 

[LOL_Wut_Axel]

Don't worry some people who claimed to know things went over the docs for bulldozer up and down for months telling people IPC went up.

You're still arguing about this crap which went down over half a year ago? Here's a good idea: stop looking to incite arguments with members.

It was JF-AMD that was saying up and down that IPC went up. Naturally, everyone believed him because, guess what, he's the director of product marketing for servers at AMD. But hey, it sure is fun to get to attack someone else for his lies, right?

 

[BallaTheFeared]

I can't imagine how their brains work, it's really fascinating though.

 

[Chiropteran]

You do realize that that makes absolutely no sense, right?

Both Core 2 Quad and Phenom II X4 have four Int and four FP units. Neither has SMT or CMT.

Huh? I was responding to your post. You said, and I quote:

Originally Posted by LOL_Wut_Axel
This may caused bruised e-peens, but the reality is that AMD at the very best can only match Intel's CPU architecture from 6 years ago.

AMD's best certainly isn't quad core phenom 2, not while the FX-8150 and 6 core phenom 2 exist. Now, perhaps you were saying one thing while you actually meant something else, perhaps you meant "AMD's best ON THIS PARTICULAR CHART", however even in that case ignoring multithreaded performance is a flawed way to compare the CPU because that chart did indeed include a 6 core phenom 2, and several dual core Intel CPU.

So yeah, I stand by what I said. You pointing out that AMD and Intel both have quad core CPU is irrelevant when the original statement I responded to was not limited to quad core CPU in any way.

 

[BallaTheFeared]

AMD did a good job with Thuban. I enjoyed mine even though it became a tiresome bottleneck in some games for me.

 

[nismotigerwvu]

Comparing clock-for-clock, 65nm Core 2 Quad is equal to just a tiny bit faster than Phenom II X4.

In terms of the not so realistic example of a single core operating at an identical clock rate sure, but things get a little more complicated when we take scaling into consideration. Remember the fairly outdated FSB and limited memory bandwidth start presenting issues as the thread count rises. Not knocking C2D or anything (amazing how relevant the chips still are) but it's foolish to just dismiss the strengths that Deneb had. Keep in mind how many of the features introduced into the product space (but by no means invented) by AMD ended up being incorporated into later intel designs. Also, it's a fairly childish to call the AMD engineers "stupid" as you've done quite a few times here. The only thing on display is what an exponentially larger budget can provide. Intel has been operating essentially a full node ahead of AMD since Conroe and has at the very least double the manpower on every project. The fact that AMD is even remotely competitive is a testament to clever, hardworking hardworking engineers they have. I mean look at Bobcat versus Atom, there's about a 40% IPC gap there despite Bobcat cores being smaller (at release, I'm not sure what the area's compare to now). Does that make the Atom team stupid? Of course not.

 

[LOL_Wut_Axel]

Huh? I was responding to your post. You said, and I quote:

Originally Posted by LOL_Wut_Axel
This may caused bruised e-peens, but the reality is that AMD at the very best can only match Intel's CPU architecture from 6 years ago.

AMD's best certainly isn't quad core phenom 2, not while the FX-8150 and 6 core phenom 2 exist. Now, perhaps you were saying one thing while you actually meant something else, perhaps you meant "AMD's best ON THIS PARTICULAR CHART", however even in that case ignoring multithreaded performance is a flawed way to compare the CPU because that chart did indeed include a 6 core phenom 2, and several dual core Intel CPU.

So yeah, I stand by what I said. You pointing out that AMD and Intel both have quad core CPU is irrelevant when the original statement I responded to was not limited to quad core CPU in any way.

Have you been paying attention to any of the discussion in this thread? We're discussing CPU ARCHITECTURE. You evaluate how fast a CPU architecture is in comparison to another by comparing them at the same clock speed and with a single core, to evaluate per-core performance. Simply adding 50% more cores onto the exact same architecture won't magically make it a faster architecture. The 3930K has 50% more cores than the 2600K. That doesn't mean it has a faster architecture, which for both is Sandy Bridge. Simply adding cores doesn't make a CPU architecture faster, easily seen by the fact that the Phenom II X6 and Phenom II X4 are within 1% of each other when comparing architectures in the chart.

Also, this has been known for months now, but the Bulldozer architecture is slower than Stars, which is the architecture powering the Phenom II and Athlon II.

Again, the topic is CPU architecture, and in terms of CPU architecture it's being discussed that Bulldozer<Stars<Conroe<Penryn<Nehalem=Westmere<Sandy Bridge<Ivy Bridge.

 

[LOL_Wut_Axel]

In terms of the not so realistic example of a single core operating at an identical clock rate sure, but things get a little more complicated when we take scaling into consideration. Remember the fairly outdated FSB and limited memory bandwidth start presenting issues as the thread count rises. Not knocking C2D or anything (amazing how relevant the chips still are) but it's foolish to just dismiss the strengths that Deneb had. Keep in mind how many of the features introduced into the product space (but by no means invented) by AMD ended up being incorporated into later intel designs. Also, it's a fairly childish to call the AMD engineers "stupid" as you've done quite a few times here. The only thing on display is what an exponentially larger budget can provide. Intel has been operating essentially a full node ahead of AMD since Conroe and has at the very least double the manpower on every project. The fact that AMD is even remotely competitive is a testament to clever, hardworking hardworking engineers they have. I mean look at Bobcat versus Atom, there's about a 40% IPC gap there despite Bobcat cores being smaller (at release, I'm not sure what the area's compare to now). Does that make the Atom team stupid? Of course not.

Clearly the FSB was't very limited if it allowed the higher-end Core 2 Quad Q9*50 to make short work of the Phenom II X4 in multi-threaded, and that's very well documented.

The R&D argument is an old and tired excuse. AMD's CPU engineers suck, plain and simple, because the GPU engineers have zero problems making excellent products, even though they have a lot less money than NVIDIA.

And yes, the team that made the netbook version of Atom was stupid. The only design goal was to make the processor as cheap to manufacture as possible and to get the lowest power consumption, actual efficiency and performance be damned. The Atom team has switched their focus to making mobile SoCs, and their first attempt has been a very good one. Atom is moving forward when it comes to the mobile (smartphone, tablet) market and not the netbook market.

 

[iCyborg]

Have you been paying attention to any of the discussion in this thread? We're discussing CPU ARCHITECTURE. You evaluate how fast a CPU architecture is in comparison to another by comparing them at the same clock speed and with a single core, to evaluate per-core performance.

So it doesn't matter that 8150 can run at 3.6-4.2 GHZ with more cores (pseudo or not), and Q6600 at 2.4GHz - the latter is a better CPU architecture because it has 10% better IPC? CPU architecture = IPC?
Using IPC exclusively is just as bad as using clocks alone. Not everything is in clocks, but not everything is in IPC either like you seem to assert...

 

[OCGuy]

Isn't Phenom II about equal to Yorkfield clock for clock?

Nope.

 

[LOL_Wut_Axel]

So it doesn't matter that 8150 can run at 3.6-4.2 GHZ with more cores (pseudo or not), and Q6600 at 2.4GHz - the latter is a better CPU architecture because it has 10% better IPC? CPU architecture = IPC?
Using IPC exclusively is just as bad as using clocks alone. Not everything is in clocks, but not everything is in IPC either like you seem to assert...

If you want to compare architectures objectively you need to run the CPUs featuring them at the same clock speed. We're not comparing clock speeds, we're comparing architectures.

Whether an 8150 can run at 4.5GHz or not doesn't change the fact that the architecture, clock for clock, is slower than something that came out 6 years ago from Intel. It's simply a way to point out how bad AMD's CPU engineers are and how far behind they are in comparison to Intel's.

Want to compare CPU to CPU? Okay, fine: compare with varying clock speeds that represents what you'll get out of each. If you want to compare architectures, it needs to be clock-for-clock because we're measuring Instructions Per Cycle and not Instructions Per Second. It'd be completely stupid to say one architecture is faster to the other when you're not putting all of them on equal ground so you can compare their raw speed.

 

[Arachnotronic]

If you want to compare architectures objectively you need to run the CPUs featuring them at the same clock speed. We're not comparing clock speeds, we're comparing architectures.

Whether an 8150 can run at 4.5GHz or not doesn't change the fact that the architecture, clock for clock, is slower than something that came out 6 years ago from Intel. It's simply a way to point out how bad AMD's CPU engineers are and how far behind they are in comparison to Intel's.

Want to compare CPU to CPU? Okay, fine: compare with varying clock speeds that represents what you'll get out of each. If you want to compare architectures, it needs to be clock-for-clock because we're measuring Instructions Per Cycle and not Instructions Per Second. It'd be completely stupid to say one architecture is faster to the other when you're not putting all of them on equal ground so you can compare their raw speed.

Hi,

You forget to note that clock speed is limited a whole lot by microarchitectural decisions. So what you really want to look at is performance per thread per watt.

 

[iCyborg]

If you want to compare architectures objectively you need to run the CPUs featuring them at the same clock speed. We're not comparing clock speeds, we're comparing architectures.

Whether an 8150 can run at 4.5GHz or not doesn't change the fact that the architecture, clock for clock, is slower than something that came out 6 years ago from Intel. It's simply a way to point out how bad AMD's CPU engineers are and how far behind they are in comparison to Intel's.

Want to compare CPU to CPU? Okay, fine: compare with varying clock speeds that represents what you'll get out of each. If you want to compare architectures, it needs to be clock-for-clock because we're measuring Instructions Per Cycle and not Instructions Per Second. It'd be completely stupid to say one architecture is faster to the other when you're not putting all of them on equal ground so you can compare their raw speed.

No, just as comparing clocks is not comparing CPU architectures, comparing IPC *alone* is also not comparing CPU architectures, not in the sense in which you're using it - to point out that engineers suck.

If IBM came out tomorrow with a CPU that beats IB by 2-3x in every known benchmark while working at 30GHz, would you also point out how far behind Intel they are and how much their CPU architecture sucks and hence their engineers must suck too (it's worse than P3!!)? Completely oblivious of real world performance? Sorry, but that would be laughable. CPU engineers develop CPU architectures to deliver performance in real workloads, not to maximize IPC...

 

[Abwx]

You do realize that that makes absolutely no sense, right?

Both Core 2 Quad and Phenom II X4 have four Int and four FP units. Neither has SMT or CMT.

You just dont realize how you , you are nonsensical...

Core2 has only two integer unit per core , so its so called
as good perfs as phenom2 is due only to softs being
generaly optimized for intel s uarch....
Overall , it has less exe ressources.

For the rest , keep on trolling all the way....

 

[ShintaiDK]

note that this is an estimation, piledriver core probably didn't even existed when this slide was made...

now, this slide is new...(but imo, very high score for gpu)

"Up to" is such a forgiving and easy manipulated way. And its even so bad it needs small notes.

Lets just see shall we, AMD havent been known to tell the truth for many years now. Else Phenom today would be able to compete with Sandy/Ivy

 

[Lepton87]

Comparing clock-for-clock, 65nm Core 2 Quad is equal to just a tiny bit faster than Phenom II X4.

At stock that is true but it's because of horribly underclocked uncore. If you overclock uncore to 2.8GHz IPC should easily match 45nm C2D or even exceed it. Overall Phenom II is a better CPU than 45nm C2Q but just barely.

 

[ShintaiDK]

You just dont realize how you , you are nonsensical...

Core2 has only two integer unit per core , so its so called
as good perfs as phenom2 is due only to softs being
generaly optimized for intel s uarch....
Overall , it has less exe ressources.

For the rest , keep on trolling all the way....

You mix up INT cores and issue ports. Core 2 for example got 3 SSE issue ports per core, but can only use 2 at a time. And be careful looking at macroscaled diagrams. The one you link is actually terrible.

 

[Arzachel]

You just dont realize how you , you are nonsensical...

Core2 has only two integer unit per core , so its so called
as good perfs as phenom2 is due only to softs being
generaly optimized for intel s uarch....
Overall , it has less exe ressources.

For the rest , keep on trolling all the way....

What ops can those int pipes do? What's the average utilization? How long are the cache latencies? Iirc, while Phenom has three int pipes per core, only a single one of them is utilized most of the time.

Comparing a single part of the arch out of context is silly and pointless, just like deciding the best CPU by just comparing core count or frequencies.

 

[T_Yamamoto]

You mix up INT cores and issue ports. Core 2 for example got 3 SSE issue ports per core, but can only use 2 at a time. And be careful looking at macroscaled diagrams. The one you link is actually terrible.

+1

 

[AtenRa]

Note, performance/watt.

AMD seems to be dropping any CPU competition with Intel. In short, they gave up.

Im not sure why you making that conclusion from that chart but if im not mistaken Intel's focus is for better performance/watt also.

AMD is also essentially dead in the serverspace.

http://www.investorvillage.com/mbth...mValue=235910&dValue=1&tid=11666535&showall=1

overall 18.9M units 19.1% share +0.3% over Q4
server 286K units 6.8% share +1.1% over Q4
dekstop 9.7M units 22.7% share +0.4% over Q4
portable 8.9M units 17.1% share +0.1% over Q4

It seams that Server is having the highest gains.

 

[Ancalagon44]

You're still arguing about this crap which went down over half a year ago? Here's a good idea: stop looking to incite arguments with members.

It was JF-AMD that was saying up and down that IPC went up. Naturally, everyone believed him because, guess what, he's the director of product marketing for servers at AMD. But hey, it sure is fun to get to attack someone else for his lies, right?

Yeah, we all got mislead.