Why Bulldozer's spotty performance is good news.

[BarbaricSoul]

Because you don't run both at the same frequency, stock its 9550's 2.83Ghz vs 955's stock 3.2Ghz (for the non-overclocking croud), and then ~4ghz for 9550 and about the same (or a bit less) for the 955 in the overclocked performance. You don't go about electronically limiting your Mustang to 100mph and then claim that your Toyota Prius goes faster because it is not limited to 100mph: you have to take two readings, one with limiters on (whatever it is, in this case corresponding to stock), and then another without limiters (overclocked).

But that's my point, he's trying to compare the BD with OC'ing(no limits) to a I5 with no OC'ing(limited). How's what I said any more wrong than what he said? Also, from your earlier post about price/performance, shouldn't we also add in the increased cost in electricity to BD has to use to even come close to the I5 performance, as it will cost more to run a BD chip compared to a I5 or I7 2nd gen chip.

 

[Fourstaff]

But that's my point, he's trying to compare the BD with OC'ing(no limits) to a I5 with no OC'ing(limited). How's what I said any more wrong than what he said? Also, from your earlier post about price/performance, shouldn't we also add in the increased cost in electricity to BD has to use to even come close to the I5 performance, as it will cost more to run a BD chip compared to a I5 or I7 2nd gen chip.

Yes there are a lot more to add to that, but it remains the fact that you are using a faulty measurement to prove your point, and in my eyes you are no better than him (and you get minus points by calling people fanboy, which is why I was annoyed at you and consequently attacked you).

 

[John Doe]

Or limit the rev to 2000rpm, knowing that the Mustang's optimum rev is much higher than that, while Prius is closer to its optimum rev.

Hmm... Mustang at 2000 RPM or a Prius? Prius should take off after the 3rd gear. Oh well, it is a Ford. It will mess the thang up anyway.

 

[BarbaricSoul]

Well I didn't take it as a attack, this isn't an arguement to me. Just a friendly debate.

I do see what your saying, and yes I will concede that a PII stock clocked at 3.7 ghz does out perform a C2Q at 3ghz. The main point I was trying to make is that no matter what speed's a PII or a BD chip is clocked at, it does not out perform the Intel equivalent chip when compared performance/GHZ or performance/core. And that is the performance that matters to me as a consumer.

 

[Dent1]

And then I read this and the best I remember the Core 2 Quad beat up the Phenom II x4 up pretty good unless your only counting gaming, where most of these quad-core chips perform so similarly that it isn’t worth factoring in the tenths of a frame.

Gaming aside. The orginal Deneb Phenom IIs with C2 stepping (4MB L3) and Core 2 Quads traded blows very evenly. But as time progressed and the Phenom IIs got tweaked to 6MB L3 cache/95W with the C3 stepping and the Phenom IIs pulled ahead a fair bit, although the Core 2 Quad still held its own.

 

[BarbaricSoul]

Gaming aside. The orginal Deneb Phenom IIs (4MB L3/125W) and Core 2 Quads traded blows very evenly. But as time progressed and the Phenom IIs got tweaked to to 6MB L3 cache/95W the Phenom IIs pulled ahead a fair bit

That depends on which C2Q your comparing the PII to. The only way the PII pulled ahead of the high-end C2Qs is higher clock speeds. Intel moved on to the I series, while AMD was stuck with the PII. If Intel wanted to increase the clock speeds of the C2Qs instead of moving on to newer architecture, PII wouldn't have a chance at beating the C2Q. Like I've said all morning long, at EQUAL CLOCK SPEEDS, a PII does NOT out perform a C2Q Q9*50.

 

[John Doe]

Gaming aside. The orginal Deneb Phenom IIs with C2 stepping (4MB L3) and Core 2 Quads traded blows very evenly. But as time progressed and the Phenom IIs got tweaked to 6MB L3 cache/95W with the C3 stepping and the Phenom IIs pulled ahead a fair bit, although the Core 2 still held their own.

If talking about Kentsfield, then yeah they do perform similar. However Penryn (Yorkfield) outdoes any Phenom 2 in single or quad threaded performance. It also has better performance per-watt against top end Denebs (95W against 140). Still, things are and have always been cheaper on AMD's end. Also, mobos with nVidia chipsets blow on the Intel side so AMD has a better SLi mobo advantage as well. Bit of a toss up.

 

[Bo$$]

I think we will see some sort of boost in apps like sony vegas, that can really use multiple threads... I am actually liking bulldozer quite a bit. But intel is very strong, for me it is price rather than outright Performance in 1 particular application

 

[Lionheart]

Ok we get it, CQ2 is better then PII :shadedshu, who cares I thought this was a Bulldozer thread.

 

[Super XP]

I couldn't give a rat ass about future , would you still use A64 single core on socket 754 or 939 (future ready back then) because now we can all use "good" software & Windows Vista/7 64bit?

Sorry! but what comes out now is got to be good to use right now , not in 2 years when it will get close to be obsolete or replace by something better...

Beside Interlagos is not doing much better then Opteron base CPU's as far as i heard
Anyhow no point of denied BD is 70% fail , long live PII

I disagree, though what you probably mean is what comes out now should be good now and for the future, it's called balance.

I will say this again, look at benchmarks with retail copies of Bulldozer, they differ from the original benchmarks released on Oct 12, 2011. Why is this? Why is the retail Bulldozer performing better? Something is up here, and with more tweaking and some love, AMD's FX line should do what it was meant to do, Bullodoze Competition

 

[xenocide]

You can claim Bulldozer is a "Server CPU" all you want, but it just sounds like excuses to justify poor per-core performance. Sure, it does great in heavily threaded applications, but those are so few and far between it's hardly worth mentioning when you consider they still marketed this as a Consumer (even Enthusiast) CPU.

 

[Super XP]

You can claim Bulldozer is a "Server CPU" all you want, but it just sounds like excuses to justify poor per-core performance. Sure, it does great in heavily threaded applications, but those are so few and far between it's hardly worth mentioning when you consider they still marketed this as a Consumer (even Enthusiast) CPU.

Bulldozer just needs the right benchmarks to show off it's real performance. That said, AMD needs to release a revision ASAP to ensure it does what it was meant to do, perform like a Bulldozer. Hopefully they can iron out it's issues and get back into the game.

This is almost identical setups for both PC's

 

[de.das.dude]

with all these people buing intels, why do you think the software companies will design softwares depending upon bulldozers architecture?

 

[Dent1]

That depends on which C2Q your comparing the PII to. The only way the PII pulled ahead of the high-end C2Qs is higher clock speeds. Intel moved on to the I series, while AMD was stuck with the PII. If Intel wanted to increase the clock speeds of the C2Qs instead of moving on to newer architecture, PII wouldn't have a chance at beating the C2Q. Like I've said all morning long, at EQUAL CLOCK SPEEDS, a PII does NOT out perform a C2Q Q9*50.

I was talking about the C3 Phenom IIs outperformed C2Q on average as time progressed. Obviously the tests where the Phenom II was triumphant was only slight and negliable and often within margin for error as the two opposing architectures very similar performers.

Yes Intel could of increased the clock speeds but considering that the Phenom II X4 processor offerings was just as fast and sometimes faster, OC'd just as far, cheaper and had AM2/AM2+/AM3 backward compatible AMD would look more attractive to the educated buyer. Also Core 2 Quad with souped up clocks wouldn’t look attractive compared to a cheaper Phenom II X6 Thuban. Intel's move to Core I-series made Intel seem attractive again.

 

[xenocide]

with all these people buing intels, why do you think the software companies will design softwares depending upon bulldozers architecture?

Nobody gave Netburst a break, or catered directly to HyperThreading. Software companies aren't going to develop for obscure changes when 90%+ of the market are still using the same design. What company is really excited to double the amount of threads their software uses when a good number of their customers don't even use that many?

Bulldozer just needs the right benchmarks to show off it's real performance. That said, AMD needs to release a revision ASAP to ensure it does what it was meant to do, perform like a Bulldozer. Hopefully they can iron out it's issues and get back into the game.

If you make the test bias of course they will win lol. The problem is that overall they don't outperform similarly priced offerings from Intel in a majority of tasks.

 

[cadaveca]

with all these people buing intels, why do you think the software companies will design softwares depending upon bulldozers architecture?

Of course they will. AMD and Intel BOTH are going to be bringing CPUs with like 20 cores to the desktop space in the next couple of years.


I kinda gotta agree with the sentiment offered by the article writer in the OP...devs need good multicore chips with true cores to write and test software on, becuase, as mentioned, both Intel and AMD are increasing core counts. Bulldozer provides an affordable option for devs NOW, so that their software in the FUTURE can be written NOW.


All i can think of is a BD-based design with a good GPU in it, for consoles, would provide a similar core platform like the PS3 does, but in x86, and with truly powerful cores. This makes me get a bit excited for the next generation of gaming consoles.

 

[xenocide]

All i can think of is a BD-based design with a good GPU in it, for consoles, would provide a similar core platform like the PS3 does, but in x86, and with truly powerful cores. This makes me get a bit excited for the next generation of gaming consoles.

Any developer would hate that. The PS3 had a more powerful CPU on paper, but it was such a nightmare to program for that most developers still have issues with it. I guarantee most next-gen consoles will still use IBM CPU's, probably based on a more consumer-friendly version of the platform Watson's hardware used.

 

[FordGT90Concept]

What types of work loads do servers mostly deal with?

FLOPs! The FPU is always the bottleneck, not the ALU.

 

[cadaveca]

Any developer would hate that. The PS3 had a more powerful CPU on paper, but it was such a nightmare to program for that most developers still have issues with it. I guarantee most next-gen consoles will still use IBM CPU's, probably based on a more consumer-friendly version of the platform Watson's hardware used.

I understand, xenocide, but for me, something providing a challenge just makes me rise to the occasion, not bitch about it.

I think the nightmare of the Cell Processor was that it wasn't running x86 code, where most devs are far more comfortable.


You're probably right about the console hardware, but i think it'd be a mistake for them to NOT be running either x86 or x64 code for compatibility reasons. It doesn't make sense for OEMs that are designing not just consoles...but lifestyles...to be using many devices all running different code, when they do not have to.

 

[Fourstaff]

I do see what your saying, and yes I will concede that a PII stock clocked at 3.7 ghz does out perform a C2Q at 3ghz. The main point I was trying to make is that no matter what speed's a PII or a BD chip is clocked at, it does not out perform the Intel equivalent chip when compared performance/GHZ or performance/core. And that is the performance that matters to me as a consumer.

I can see your point too, but you were lucky that the PII and C2Q have more or less the same roof. If PII had a roof of about 8Ghz, compared to 4Ghz of C2Q then, with the current work done per clock cycle per core the PII is going to absolutely demolish C2Q and some more.

 

[techtard]

Please take the fanboy slapfight to an Phenom vs C2Q thread. This is for (yet another) Bulldozer thread.

 

[Bo$$]

even if this isnt as good as expected, the price of the overall platform is cheaper. i think it will be pretty good in a few years as games and apps get more threaded

 

[Disparia]

This is easy then... stop thinking like a bunch of single task consumers

I can't be the only one here who puts a rip to transcode, opens up a game on monitor 1 while a show/movie plays on monitor 2. ...and that's just my instance on my multi-instance home mainframe for my wife and two little users

The first part is true, the last part is where I'm heading. GPU virtualization is here and will continue to get better. That was really the last part needed to make it happen as CPU power, RAM density, PCIe lanes and storage I/O are plentiful/powerful.

 

[Fatal1ty39]

found this video AMD Fx-8 Vs intel core i5 2500K
http://www.engadget.com/2011/10/12/amd-fx-processor-brings-eight-cores-to-battle-we-go-eyes-on-vi/

 

[cadaveca]

I love to see Macci. Thanks!

Pretty funny to hear Sacha say that Bulldozer will be a catalyst for multithreaded programming for devs. I just said that myself!