AMD FX: Energy Efficiency Compared To Eight Other CPUs

[techelite]

[citation][nom]techpops[/nom]The funny thing about this is it only comes out right after the fail that was Bulldozer. No mention of it in the tech press before then. AMD didn't bring it up as any kind of excuse for BD's lack of performance. In fact, only some AMD fans have got this story, not really into the press as such but into comments around articles talking about Bulldozer.Come on guys, get real, accept it, BD is a flop. Lets hope AMD do better next time but no need to go all conspiracy theory nuts and declare all benchmarks everywhere null and void now. Geez this is getting pathetic.[/citation]

You amuse me with your contradiction. Quote - "The funny thing about this is it only comes out right after the fail that was Bulldozer" and in the next breath "Lets hope AMD do better next time but no need to go all conspiracy theory nuts... ". But on a separate issue, at least you admit that you are "... not really into the press". This case and it's issues were widely reported. Tom's hardware only touched on it without going into the damaging details http://www.tomshardware.com/news/amd-intel-1.25-billion-settlement,9258.html?xtmc=amd_intel_settlement&xtcr=4 . One of the important points raised out of my posts is that it is important that credible tech sites properly evaluate new technologies. If they are unable to do so, they should include caveats regarding the inherent biases in their testing rigs and to reiterate them in the results. Windows OS will always have a performance edge on Intel processors because it is compiled with Intel's compiler. The closest I have seen to a comprehensive test between SandyBridge and Bulldozer architectures was on Phoronix.com, using OpenBenchmarking software http://www.phoronix.com/scan.php?page=news_item&px=MTAxMzY . Believe it or not, on a neutral platform and compiler, Bulldozer was faster in the majority of the tests (not necessarily more power efficient, but that's not tested). Although, I would still like to see a benchmark using Intel's and AMD's own compilers on Linux. That would be an even better representation of their respective engineering.

 

[Draven35]

Unfortunately, a 'neutral platform and compiler' isn't what most users are going to be using these processors with. They are going to be running actual applications, more likely than not under Windows. People cite generic benchmarks as not reflecting real-world usage, and then ask for completely generic benchmarks.

And most of the Phoronix suite's benchmarks that are heavy on FPU operations quite clearly show problems with Bulldozer even when used with a 'neutral compiler'.

 

[techpops]

[citation][nom]techelite[/nom]You amuse me with your contradiction. Quote - "The funny thing about this is it only comes out right after the fail that was Bulldozer" and in the next breath "Lets hope AMD do better next time but no need to go all conspiracy theory nuts... ". But on a separate issue, at least you admit that you are "... not really into the press". This case and it's issues were widely reported. Tom's hardware only touched on it without going into the damaging details http://www.tomshardware.com/news/a [...] ent&xtcr=4 . One of the important points raised out of my posts is that it is important that credible tech sites properly evaluate new technologies. If they are unable to do so, they should include caveats regarding the inherent biases in their testing rigs and to reiterate them in the results. Windows OS will always have a performance edge on Intel processors because it is compiled with Intel's compiler. The closest I have seen to a comprehensive test between SandyBridge and Bulldozer architectures was on Phoronix.com, using OpenBenchmarking software http://www.phoronix.com/scan.php?p [...] px=MTAxMzY . Believe it or not, on a neutral platform and compiler, Bulldozer was faster in the majority of the tests (not necessarily more power efficient, but that's not tested). Although, I would still like to see a benchmark using Intel's and AMD's own compilers on Linux. That would be an even better representation of their respective engineering.[/citation]

I stand by everything I wrote, that you have problems with it just informs me all the more you're deluded. As the next commenter points out, people want to know how their hardware will really perform in the real world. Nobody but some desperate fanboys give two hoots about how Bulldozer performs under Linux running an AMD compiler. Maybe some server guys are interested in that and if so I'd direct them to sites that focus on the enterprise and linux's role in that with code that's often hand optimised anyway for each job.

You just have nothing but some vague ideas and are trying to hold them up as some reason to question whether Bulldozer has some hidden power no one but you is seeing. It doesn't. It's not my opinion about that, it's that of the programmers and developers who create the software I use everyday. If they were suggesting you should run Intel because it will perform better, you can bet I would be taking note of that but they aren't saying that. Again I quote Maxon, a huge developer of highly complex software written for Windows, OSX and I believe they have special versions for linux too if the customer orders big enough. They say it doesn't matter whether you're running Intel or AMD and if anyone would know, its them. Neither one makes a blind bit of difference. You buy based on your budget, end of story.

There was a time in the past when there a lot of Intel only software and for Intel optimised but those days are gone and arguing about micro amounts from compilers is just sad, grasping at straws and I hope this is the last comment I ever have to make on it.

Yours, a fan of AMD since the K6 chips and wishing I wasn't getting tarnished with the same brush as these loonies talking up how Bulldozer could be great, if only...

 

[techpops]

[citation][nom]Draven35[/nom]Unfortunately, a 'neutral platform and compiler' isn't what most users are going to be using these processors with. They are going to be running actual applications, more likely than not under Windows. People cite generic benchmarks as not reflecting real-world usage, and then ask for completely generic benchmarks.And most of the Phoronix suite's benchmarks that are heavy on FPU operations quite clearly show problems with Bulldozer even when used with a 'neutral compiler'.[/citation]

Exactly right.

On a lighter note, I think we may be seeing the birth of an all new hybrid fanboy. A fanboy so extremist that he or she is both a Linux fanboy and AMD fanboy. Twice the fanboy in the same space. it's scary to think about it really. I'd like to christen this new sub speciesm the lesser spotted Linuxamdopolis, rarely seen in the wild and to be avoided at all costs. Rumour has it that they hunt young lone osx users but will attack Windows users in larger groups, especially if the MS user is weakened with spyware. Be on your guard, lets be safe out there.

 

[alpha754293]

MATLAB is a computationally heavy tool, not an integer heavy tool. It can work on integer/integer-only computations, but it's mostly for intense floating point calculations. Having more ALUs won't help it. In fact, you can see that having FEWER FPUs is hurting the Bulldozer performance, and that it is also clocked at a lower speed than it's predecessors is also hurting it.

 

[de5_Roy]

This is for the benefit of those following this post.

AMD settled in court over Intel compiler optimisations just in late 2010. Purposefully crippling computer code leads to poor performance that leads to power inefficiencies. Relevant and on topic. I'm not sure what the outcome would be using a neutral testing platform (ie Linux) compiled with optimisations geared to each brand of processor. But I'd sure like to know the truth, rather than believing the clearly unreliable results from the biases inherent in the test bench used in this article. Wouldn't you all?

Quote from Wikipedia http://en.wikipedia.org/wiki/Intel_C%2B%2B_Compiler

......
The FTC settlement included a disclosure provision where Intel must:[23]:
" ...publish clearly that its compiler discriminates against non-Intel processors (such as AMD's designs), not fully utilizing their features and producing inferior code."

reiteration: read the stuff you refer.
i read the whole post. was aware of the cpu dispatcher issue.
now the underlined parts stood out to me. not because they're facts, but because you sorta refuted your own complaint. let me explain: you call tom's benchmarks suspect because windows uses intel optimized compilers. then you (and a lot of amd fan(boy?)s) ask for benchmarks that use compilers that have cpu optimizations for each brand - how is that neutral?
like techpops said, a lot of softwares don't care about optimizations and like draven35 said, most users will not run a perfectly non-optimized platform. if cpu optimization is an issue, may be amd fans will blame arm next since win 8.0 will run on arm systems.
even the (possibly non factor in win 7) intel optimizations do not explain the 100+ watts excess power consumption.
i could say a lot more things, but the rest of them have already been said by others.
and lastly, amd and intel settled with each other, why bring this up now? where was this before bulldozer?
just wait till the next fx revision and accept the fact that right now fx is more inefficient than current sandy bridge and some old lynnfield cpus.
edit: or don't accept. everyone is entitled to his/her own opinion. <- my opinion.

 

[techelite]

[citation][nom]de5_roy[/nom]reiteration: read the stuff you refer.i read the whole post. was aware of the cpu dispatcher issue. now the underlined parts stood out to me. not because they're facts, but because you sorta refuted your own complaint. let me explain: you call tom's benchmarks suspect because windows uses intel optimized compilers. then you (and a lot of amd fan(boy?)s) ask for benchmarks that use compilers that have cpu optimizations for each brand - how is that neutral? like techpops said, a lot of softwares don't care about optimizations and like draven35 said, most users will not run a perfectly non-optimized platform. if cpu optimization is an issue, may be amd fans will blame arm next since win 8.0 will run on arm systems.even the (possibly non factor in win 7) intel optimizations do not explain the 100+ watts excess power consumption.i could say a lot more things, but the rest of them have already been said by others.and lastly, amd and intel settled with each other, why bring this up now? where was this before bulldozer?just wait till the next fx revision and accept the fact that right now fx is more inefficient than current sandy bridge and some old lynnfield cpus.edit: or don't accept. everyone is entitled to his/her own opinion. <- my opinion.[/citation]

de5_roy, have you heard of the term astroturfing? I don't quite understand your logic. It's a little twisted. Firstly to those who are suggesting that I could be an AMD "fanboy" aka a non-thinker you're quite wrong and it is in fact the contrary, if you were to read my posts clearly and with an open mind. I will declare the following out aloud for the sake of adding credibility to my arguments - I have often chosen AMD products on the basis that AMD are the underdogs and keep Intel honest with consumers. Without AMD's competition, Intel would most probably not bother to tick and tock their processors (remember what happened with the Pentium 4 and the lack of competition at the time?). I have been satisfied buying AMD processors in the past that have generally reached 85-90% of the far more costly Intel ones. To continue, I think I have made some important technical points in my posts that were not mentioned and are highly valid and have been rather whitewashed. Let's make it clear. I have never stated that Bulldozer is a superior architecture than SandyBridge. I would not be surprised if, as most results suggest, that it is inferior running Windows 7 and associated platform software. I have suggested partly why this is so - the Intel compiler that cripples the performance of AMD et al. Now listen de5_roy, I never said that the results were "suspect". That's your own word and interpretation of what I wrote. I avoid using vague language like this. What I have criticised is the testing methodology in producing fair and proper results. I do however concede that most people reading Tomshardware are home "prosumers" who probably don't give a stuff about server farms and supercomputers and are run Windows 7. However, a caveat should be placed in the test conclusion regarding the test rig and its inherent bias toward the Intel processor. A fair and proper test, I repeat, a fair and proper test would see either the use of a test bench compiled with Intel's and AMD's own compilers with their respective optimisations or to use a neutral one (ie GCC 4.6). A relatively neutral test bench has been employed in achieving the results at www.Phoronix.com (who is a one guy show who has often chided AMD on several occasions for their poor public relations and in not sending him samples of latest hardware to test, given he is the programmer of openbenchmarking software). Results are there for the world to see. Look at the facts for yourself.

Thanks everyone for the arguments. And stop trying to fry eggs on AMD processor dies! Moore's Law and beyond!

 

[techelite]

[citation][nom]techpops[/nom]Exactly right. On a lighter note, I think we may be seeing the birth of an all new hybrid fanboy. A fanboy so extremist that he or she is both a Linux fanboy and AMD fanboy. Twice the fanboy in the same space. it's scary to think about it really. I'd like to christen this new sub speciesm the lesser spotted Linuxamdopolis, rarely seen in the wild and to be avoided at all costs. Rumour has it that they hunt young lone osx users but will attack Windows users in larger groups, especially if the MS user is weakened with spyware. Be on your guard, lets be safe out there.[/citation]

Where's David Attenborough when you need him? Damn, probably sharing a pot of tea with Richard Dawkins. Where's Richard Dawkin's when you need him? Damn, probably sharing a pot of tea with David Attenborough.

 

[Guest]

why most people here discard the cpu optimization firsthanded when testing cpu performance(ipc,energy,etc,etc),but the same people cry loud when a crappy console ported to pc game does not perform up to the capability of the pc.
We have seen this over an over,nvidia to (ati)amd,intel to amd.
lets do the same with an application optimize to run on 3dnow,amd64 or wathsoever and you will get the same.

remember when amd launch the athlon 64,and intel the launch a processor with a simulated 64 bit instruction set? or the lack of ssee3 in the amd proccesors,or the intel ht fiasco.

In not saying that the test are not properly done;but we still haven't seen a true test of the BD capabilities(amd code compiled optimization).

Maybe another time.


anyway im not an amd fanboy.

Still using G processor mac ,and a ppc cell ps3 as a compute station.

 

[Draven35]

[citation][nom]opcodeoptimum[/nom]anyway im not an amd fanboy.Still using G processor mac ,and a ppc cell ps3 as a compute station.[/citation]

Hopefully you aren't using it because you think its faster...

 

[Guest]

nope i use it because it still get the work done.
and the ps3 is actually a 6 clustered units.

G3 mac is used to make the reports , graphs and all the paperwork,the cluster is used to calulate,analyze data etc,etc.

Maybe i will migrate the system to a cuda based one.

I know the system sound lacking,but believe me optimization is a big deal.
Not acussing bias or anything,but to show the true potential of a processor optimized code must be used.

 

[Draven35]

I saw some great Cell processor demos at SIGGRAPH a few years ago..

 

[techelite]

[citation][nom]Anonymous[/nom]nope i use it because it still get the work done.and the ps3 is actually a 6 clustered units.G3 mac is used to make the reports , graphs and all the paperwork,the cluster is used to calulate,analyze data etc,etc.Maybe i will migrate the system to a cuda based one.I know the system sound lacking,but believe me optimization is a big deal.Not acussing bias or anything,but to show the true potential of a processor optimized code must be used.[/citation]

Wow, someone else who understands (or at least acknowledges) what I'm talking about. Windows is compiled with the Intel compiler and is optimised to run on Intel processors. All test benches running Windows to test AMD processors are going to be "non-optimised AMD test benches" (actually AMD cippled). This is not some fanboy conspiracy theory. It is in fact a reality. Do we want to compare apples with apples and oranges with oranges? I would suspect not, if you are an Intel share holder.

 

[techpops]

[citation][nom]techelite[/nom]Wow, someone else who understands (or at least acknowledges) what I'm talking about. Windows is compiled with the Intel compiler and is optimised to run on Intel processors. All test benches running Windows to test AMD processors are going to be "non-optimised AMD test benches" (actually AMD cippled). This is not some fanboy conspiracy theory. It is in fact a reality. Do we want to compare apples with apples and oranges with oranges? I would suspect not, if you are an Intel share holder.[/citation]

Here's a little fact that puts a spanner in your argument. If you spend the least amount of money you can today on an AMD based PC, so that would include a cheap AM3 motherboard with everything on it that matters today like USB3 and such, then use a 6 core 1090T and overclock it. There isn't actually anything from Intel that can beat it on price/performance.

The reason I'm so aware of this is I've studied the Cinebench results of all the processors out there and their prices then compared build costs to get the best 3D animation box I could for the money. And while Intel has much faster solutions with the 2600k (which is looking everyday more like the best CPU we've ever had) it still costs significantly more for what is a fairly small percentage speed bump in Cinebench. Or to put it another way, when comparing how much the CPU's can do on all cores going flat out, which is something you want to test for 3D work as that's exactly what will be happening when rendering.

So contrary to your belief that AMD are at a disadvantage and Bulldozer just shows how much of an issue that is. The 1090T build despite its age tells a different story.

I can pull this same trick with the server market and show you how a 2 cpu setup for AMD is faster than anything Intel has on the market today for the same money.

I should also add this is a specialist solution I'm talking about. That being a computer that will use all cores 100% of the time. Gaming PC's don't need that kind of focus so you'll find better value building an Intel box for gaming as you don't need the 2600k to get great frame rates

I can lay out the prices and the percentage differences in performance if you don't believe this. But really once you accept it, that the efficiency of AMD when it comes to price/performance is still stellar, its hard to make an argument that the world is Intel biased and AMD can't compete unless the world moves over to processor agnostic compilers.

 

[techelite]

That's a different angle on it. As I declared in my previous posts "I have been satisfied buying AMD processors in the past that have generally reached 85-90% (of the performance) of the far more costly Intel ones". But the thrust of my posts has been more on the scientific side of things: how CPU architectures/ engineering ought to be compared more accurately/fairly/ purely/ squarely, on an unbiased test bench. It's like the importance of testing the ability and speed of racehorses from different stables in the dry (Linux) and the wet (Windows), with the jockeys (compiler) atop both beating the horses with equally vigor and determination. Perhaps we will find that it's horses for courses. But we'll never know if we just do a single test using Windows with Intel's compiler refusing to give the AMD processor a good and proper thrashing. However, It does seem to be that the general consensus from other reviews on the net is that the Bulldozer architecture is good at multi-threaded applications - for example transcoding. But because of the slower clock speed it is not looking as good as the Intel Sandy Bridge in single threaded application performance, namely games and office apps. I think what AMD has done here is to bank on the prediction that developers will move their applications to the multi-core paradigm sooner rather than later. The rise of quad-core ARM processors (and Windows 8) will hasten this, surely. Also, the Bulldozer architecture is designed to share intra processor resources in a way that allows it to scale down for more efficiency and therefore higher clocks/ higher thermal envelope. I think that this is a reasonable strategy. But it won't pay big dividends immediately. We'll have to see, most probably in the third generation based on the 22 nm FD-SOI Steamroller core in late 2012/ early 2013.

 

[rick_h]

Folks,

Before I go on a rant, be advised that I have only owned two Intel products in the last 15 years. I have built over 17 systems in that time. I currently own a FX-8120 system (I pre-purchased before product launch) which I regard as the worst technical investment I have made. I am utterly amazed that AMD actually released this abortion. Unlike the problem riddled Phenom that was released in late 2008, the architecture of this CPU is so badly flawed that they should have converted these CPUs into paperweights and taken the loss.
Allow me to point out that without competition, Intel would be selling their CPUs for over $1000 and they would cease to innovate, just as they did back in the early 90's. We need AMD, but if all they can do is build space heaters, than we all lose.
In closing, I don't know what my next system will have inside, but I guaranty it will not be AMD.

 

[techelite]

[citation][nom]rick_h[/nom]Folks,Before I go on a rant, be advised that I have only owned two Intel products in the last 15 years. I have built over 17 systems in that time. I currently own a FX-8120 system (I pre-purchased before product launch) which I regard as the worst technical investment I have made. I am utterly amazed that AMD actually released this abortion. Unlike the problem riddled Phenom that was released in late 2008, the architecture of this CPU is so badly flawed that they should have converted these CPUs into paperweights and taken the loss.Allow me to point out that without competition, Intel would be selling their CPUs for over $1000 and they would cease to innovate, just as they did back in the early 90's. We need AMD, but if all they can do is build space heaters, than we all lose. In closing, I don't know what my next system will have inside, but I guaranty it will not be AMD.[/citation]

rick_h - Everything that you've written sounds legit, although a little vague on detail and coloured with extreme and histrionic language, except for your closing statement that whiffs of an astroturfer. If you were as committed a supporter of AMD the underdog company (and you make out that you are a pretty committed one at that), then I would be totally surprised to find you writing such a through away line. The product is not "flawed". It just hasn't met your expectation of a super jump in performance over your previous processor (tried overclocking it to 4.42 GHz?). AMD makes it clear there will be 4 generations including a die shrink - Bulldozer, Piledriver, Steamroller and Excavator. It doesn't sound like a flawed micro-architecture to me.

 

[de5_Roy]

rick_h - Everything that you've written sounds legit, although a little vague on detail and coloured with extreme and histrionic language, except for your closing statement that whiffs of an astroturfer. If you were as committed a supporter of AMD the underdog company (and you make out that you are a pretty committed one at that), then I would be totally surprised to find you writing such a through away line. The product is not "flawed". It just hasn't met your expectation of a super jump in performance over your previous processor (tried overclocking it to 4.42 GHz?). AMD makes it clear there will be 4 generations including a die shrink - Bulldozer, Piledriver, Steamroller and Excavator. It doesn't sound like a flawed micro-architecture to me.

you know he's talking about the system he owns and is unsatisfied with (imo rightfully so) - an fx 8120 system. not bulldozer architecture.
iirc phenom was bad at launch too. but amd managed to revise and revise and bring out ph ii - which is a much better cpu.

(tried overclocking it to 4.42 GHz?)

overclocking it at that speed would make it use a lot more power even though it already uses excessive power on stock (compared to a sb i5). not to mention it'll be slower than an i5 - fx 8120 is a 4 module 8 core processor and its single threaded performance is worse than i5's despite it's high clock speed (4 module fxes do excel at truecrypt benchmark, though).
so far any 'astroturfing' i've seen is from amd fanboys attempting to mislead people by saying fx is a all-round better cpu than intel's core i5/i7. although fanboys don't seem to realize that both intel and amd are big corporations and should be fine without anyone campaigning for them.

 

[techelite]

rick_h - Everything that you've written sounds legit, although a little vague on detail and coloured with extreme and histrionic language, except for your closing statement that whiffs of an astroturfer. If you were as committed a supporter of AMD the underdog company (and you make out that you are a pretty committed one at that), then I would be totally surprised to find you writing such a through away line. The product is not "flawed". It just hasn't met your expectation of a super jump in performance over your previous processor (tried overclocking it to 4.42 GHz?). AMD makes it clear there will be 4 generations including a die shrink - Bulldozer, Piledriver, Steamroller and Excavator. It doesn't sound like a flawed micro-architecture to me.

you know he's talking about the system he owns and is unsatisfied with (imo rightfully so) - an fx 8120 system. not bulldozer architecture.
iirc phenom was bad at launch too. but amd managed to revise and revise and bring out ph ii - which is a much better cpu.

(tried overclocking it to 4.42 GHz?)

overclocking it at that speed would make it use a lot more power even though it already uses excessive power on stock (compared to a sb i5). not to mention it'll be slower than an i5 - fx 8120 is a 4 module 8 core processor and its single threaded performance is worse than i5's despite it's high clock speed (4 module fxes do excel at truecrypt benchmark, though).
so far any 'astroturfing' i've seen is from amd fanboys attempting to mislead people by saying fx is a all-round better cpu than intel's core i5/i7. although fanboys don't seem to realize that both intel and amd are big corporations and should be fine without anyone campaigning for them.

Heh? FX-8120 is a bulldozer architecture. and Yep 4.42GHz eats power. 100 Watts more over idle. It's called overclocking. Agree Intel single core eats AMD single core due to the AMD module core design, that is aimed more at multi-threading applications. Not quite sure about idle power. When comparing CPU power between chips, it is necessary to compare total system power because AMD incorporates much of the Northbridge functionality into their CPUs, unlike Intel. I agree that there are both Intel and AMD fanboys writing rubbish. I respect your knowledge de5_roy, but you must admit that this guy, rick_h, is spinning a croc-o-5hit story. If you're not prepared to admit it, I imagine that you'd loose some credibility with those who are bothering to follow this post into its 8th page now, and would see you as an Intel champion above all else.

 

[de5_Roy]

Heh? FX-8120 is a bulldozer architecture. and Yep 4.42GHz eats power. 100 Watts more over idle. It's called overclocking. Agree Intel single core eats AMD single core due to the AMD module core design, that is aimed more at multi-threading applications. Not quite sure about idle power. When comparing CPU power between chips, it is necessary to compare total system power because AMD incorporates much of the Northbridge functionality into their CPUs, unlike Intel. I agree that there are both Intel and AMD fanboys writing rubbish. I respect your knowledge de5_roy, but you must admit that this guy, rick_h, is spinning a croc-o-5hit story. If you're not prepared to admit it, I imagine that you'd loose some credibility with those who are bothering to follow this post into its 8th page now, and would see you as an Intel champion above all else.

i am guessing that you got that 4.42 ghz number from phoronix's fx oc article. even in that article fx seems to consume a lot more power for the tasks it does. those power consumption figures are similar to other power consumption figures i've seen elsewhere. smart of phoronix to hook up a wattmeter since amd's driver is useless (proved in a previous article). tomshardware measured total system power consumption. their power-efficiency measurements are (much) more in-depth than other websites that i've seen.
not to mention this chip is almost incapable of clocking higher than 200-500 mhz (fx 8150 has a top turbo freq. of 4.2 ghz) on air cooling on average. that's not much of a speed gain and it certainly does not translate into performance-efficiency gain. iirc core i5 and i7 reaches fx's power numbers when they use gtx 580s and are oc'ed to 4.6 ghz.
i find it slightly more credible hearing from someone who has had an actual personal experience with something than someone who hasn't. that's one of the reasons i like toms' reviews more than other websites'. i know i cannot verify what everyone says about everything. but usually people are honest about what they feel. i have read a few comments of different fx owners. some of them bought fx to play with the cpu - oc it as much as they want. some of them wanted to upgrade their system. some of them wanted to build a gaming system and so on. a lot of people agree that fx is a joy to oc. i think that's true - if you want similar capability from intel you have to buy pricey x58 or x79 systems. a lot of them state that their systems play all games smoothly at high res. then there are people who are unsatisfied with what they've bought. they also pretty much say how disappointed they are and how fx doesn't offer as much performance and/or efficiency as they paid. just because someone doesn't go into gory details (won't doing that make your favorite product seem even worse?) doesn't make them any less true (or false). so far i've seen a LOT of amd faithfuls (not fanboys) being disappointed with fx.
if i cared about what people think of me i'd never join in any conversation or done anything in my life. people will always make up their own interpretations of what someone says, no one can control that.

 

[rick_h]

First, thanks for your comments and snide remarks ;-). I confess that I did overclock my CPU to be on par with its FX 8150 sibbling. Although it runs stable without a voltage increase, it quickly reaches it's maximum core temperature even with an above average water cooling system. I have retrofitted another fan on the radiator to increase air pressure which now seems to reduce this anomaly. Only in extreme case do I get an alarm. Winter is closing in and I have no heat in my office, so to warm the room up, I simply run Battlefield 2 or rip a blue ray to MP4. My office is now nice and toasty. One last observation, none of you actually own an FX 81xx, therefore spare me your biased dribble, as it is pure conjecture. My grievance is about the terrible power management on this CPU. I knew this when I built the system. However, what I didn't know is that it was a step backwards from the AMD Phenon X6 1190i system.

Rickey M. Horwitz

http://www.hellbentcycles.com

 

[techelite]

I have made up a two FX 8150 rigs for my customers. I benched one against my own AMD Phenom II 1090T rig with Linux OS on board. It outscores it by around 14-18% in video transcoding. Not a huge jump right? 8 cores vs 6 cores. However it still comes under the same thermal envelope of 125 Watts TDP. So there is a real increase in performance per watt. Just not as much as one would have hoped. With 8 cores I would have like to have seen more performance. And I will wait for the second or third generations of these bulldozer processors before I do my upgrade. When you add extra threading to video transcoding, it impacts a little on the picture quality. It's probably more technical than perceptual. But If I'm going for 8 cores, it has to be a flyer. For those who are still stuck on 2 or 4 core AMD platform, it is still worth the upgrade, if you can't wait. Bulldozer is not bull5hit.

 

[Guest]

what about for networking using virtual space, like internet cafe, isn't more cores always better? to mimic more computers...?