Sign in with
Sign up | Sign in
Your question

Why is FX8150 bad?

Last response: in CPUs
Share
April 25, 2012 4:59:12 PM

Why in some reviews say that the AMD FX-8150 is bad?
What is inferior about it againts like a 2500k or a 2600k or even the 3770k?

Even in the CPU Hierarchy says that it 4 tiers below the intel i7's
http://www.tomshardware.co.uk/gaming-cpu-review-overclo...

Why and how is it bad?

More about : fx8150 bad

a b à CPUs
April 25, 2012 5:05:51 PM

Because it is slow.

Simply put, the equation for CPU performance is as follows:

Performance = Number of Cycles * Instructions Per Cycle

The 'Instructions Per Cycle' is slower in BD then other AMD architectures, so at the same speed, a single BD core is slower then a single PII core. And because BD isn't clocked significantly higher, it doesn't offer significantly better performance over Phenom II. Problem is, Phenom II is only about as good as a 9000 series C2Q, which is obsolete by todays standards.

There are other issues that have been discusses as well: High cache latencies, lack of SW scaling, poor CMT implementation.

Long story short, BD's performance per clock is down, and that hurt performance more then a higher clock and more cores helped it.
a b à CPUs
April 25, 2012 5:49:08 PM

i remember back in the day people would say e8400 4ghz dual core was better for gaming then a 3.4ghz quad core...now people say higher clocked quad is better than a slower clocked eight core

if you want the best fps go for intel
if you are budget oriented (sub 1 grand build) then go for amd.
Intel is faster but you will pay for that, amd is not horrible but it gives you a better value/features if your budget is low
Related resources
a b à CPUs
April 25, 2012 7:20:23 PM

alvine said:
i remember back in the day people would say e8400 4ghz dual core was better for gaming then a 3.4ghz quad core...now people say higher clocked quad is better than a slower clocked eight core

if you want the best fps go for intel
if you are budget oriented (sub 1 grand build) then go for amd.
Intel is faster but you will pay for that, amd is not horrible but it gives you a better value/features if your budget is low


Well, remember IPC wasn't significantly different between C2D and C2Q, so in many cases, the faster clocked duo WAS the better option for gaming. IPC matters, especially since clock speed isn't advancing that fast...
a b à CPUs
April 25, 2012 7:26:13 PM

alvine said:
i remember back in the day people would say e8400 4ghz dual core was better for gaming then a 3.4ghz quad core...now people say higher clocked quad is better than a slower clocked eight core



that is complete crap and hyperbole, care to link any proof of that otherwise you are lying?

i would take an E8400 @ 4ghz over an athlon II x4 at 3.4 ghz any day. even in multi core optimized games the E8400 out performs a slower athlon II x4. would i take the intel dual core over a phenom II x4, no. everything is relative to how well it performs individually. the FX-8 is an over priced CPU and runs slower then similar priced intel CPUs because of bad architecture on AMD's part.


April 25, 2012 8:13:38 PM

dirtyferret said:
that is complete crap and hyperbole, care to link any proof of that otherwise you are lying?

i would take an E8400 @ 4ghz over an athlon II x4 at 3.4 ghz any day. even in multi core optimized games the E8400 out performs a slower athlon II x4. would i take the intel dual core over a phenom II x4, no. everything is relative to how well it performs individually. the FX-8 is an over priced CPU and runs slower then similar priced intel CPUs because of bad architecture on AMD's part.


It depends on context. Back when multi-core was new and becoming mainstream, no games were being developed to take advantage of them, so fewer faster cores often were much better than more slower cores, because any game would still be single threaded.

Today that is very different, and as we go forward a significant chunk of game logic can be offloaded to the GPU via openCL, so in reality game performance cpu wise has plateaued - path finding, collision detection, state awareness, and other frequently done non-visual game actions are easily writable for massively parallel gpus, so you will end up having a vast majority of a game running there.

It will be slow to happen, but as it slowly does, we won't see higher cpu demands for games by much, since anything new can be done gpu side.

In terms of Bulldozer, the 8150 is at a fine price point if you overclock it and don't care about power. It is inefficient in terms of power usage but its price point makes it more economical than a 2500 (soon to be 3570k) and all the new chip brings to the table is even more of the power inefficiency showing itself.

But if you don't pay your electric bill, you really don't need to care. So an 8 core bulldozer chip is "ok" because the i5 2500k / 3570k will only be 10% or so better in "most" games.
a b à CPUs
April 25, 2012 8:44:07 PM

zanny said:
1- It depends on context. Back when multi-core was new and becoming mainstream, no games were being developed to take advantage of them, so fewer faster cores often were much better than more slower cores, because any game would still be single threaded.

2 - Today that is very different, and as we go forward a significant chunk of game logic can be offloaded to the GPU via openCL, so in reality game performance cpu wise has plateaued - path finding, collision detection, state awareness, and other frequently done non-visual game actions are easily writable for massively parallel gpus, so you will end up having a vast majority of a game running there.

It will be slow to happen, but as it slowly does, we won't see higher cpu demands for games by much, since anything new can be done gpu side.

3 - In terms of Bulldozer, the 8150 is at a fine price point if you overclock it and don't care about power. It is inefficient in terms of power usage but its price point makes it more economical than a 2500 (soon to be 3570k) and all the new chip brings to the table is even more of the power inefficiency showing itself.

But if you don't pay your electric bill, you really don't need to care. So an 8 core bulldozer chip is "ok" because the i5 2500k / 3570k will only be 10% or so better in "most" games.


1- that is what i said so lets not try to turn my statement around and make you look as if you are correcting me. you are not, you are agreeing with me.

2- fail; if game performance has plateaued then why do I get a higher FPS with one CPU over another. that is not a plateau, if you need to learn what the definition of a plateau is (seems like you do) I will gladly post it below

a : a usually extensive land area having a relatively level surface raised sharply above adjacent land on at least one side : tableland b : a similar undersea feature
2
a : a region of little or no change in a graphic representation b : a relatively stable level, period, or condition
3
: a level of attainment or achievement <the 500-point plateau>


the following current multi-core game does not show stable FPS across the board. certain CPUs offer better performance then others, therefore it is not a plateau. while the GPU does take most of the work load during gaming and no one will see the difference between a 60FPS game and a 80FPS in the real world (that was not your argument), in a benchmark world you can clearly see a numbers difference



3- no one stated it was not an OK chip so lets not start another discussion to make yourself look knowledgeable, you are not. the chip is OK but there are better chips out there for the price.

i will excuse your lack of knowledge as a noob attempt at correcting someone by creating imaginative argument. you failed and look the fool.
Anonymous
a b à CPUs
April 25, 2012 9:16:58 PM

the best thing to do is read every OBJECTIVE review of the FX8150 and try to comprehend the flaws that are pointed out and explained how it affects performance. (if it is an objective review there ought to be some good points also)

to simply just ask the question on an open forum is opening the opportunity for a lot of "internet experts" to defend their personal preference and subjectively try to flaunt their limited knowledge of false logic and attack differentiating opinions.

and EVERYONE with a brain knows that if the letters FX are in the name it HAS to SUCK!
:lol: 
April 26, 2012 2:36:02 AM

Quote:
Stupid thread, deserves the rubbish that all the sheep contribute to.

agreed +1..... the bait was laid and the trap worked. So many threads about Intel vs. BD all over the forums that the OP couldn't have missed.
a c 146 à CPUs
April 26, 2012 4:45:46 AM

It's so bad because of the architecture. The IPC is all messed up making equal to an old Core 2 Duo which by now can be considered obsolete.
April 26, 2012 6:50:22 AM

interesting the fx-4170 is one tier above the 8150 and two tiers above hte FX-8120, -6100, -4100. just shows you how inconsistent BD really is for gaming.
a c 185 à CPUs
April 26, 2012 6:57:41 AM

Because it's a quad core that's clocked in @4.2ghz...
a b à CPUs
April 26, 2012 7:01:21 AM

zakattak80 said:
interesting the fx-4170 is one tier above the 8150 and two tiers above hte FX-8120, -6100, -4100. just shows you how inconsistent BD really is for gaming.


It's still beaten by the 3+ year old i7 920 which tells you something. Such a pointless duplicate thread.
April 26, 2012 7:05:31 AM

amuffin said:
Because it's a quad core that's clocked in @4.2ghz...


yeah but still the phenom II chips fall more in line like they should, if i'm not mistaken the x6 series was slower clocked then its x4 cousin, yet the x6 are above most of the x4.
a b à CPUs
April 26, 2012 7:10:55 AM

This was chumming the waters to attract the sharks.
a b à CPUs
April 26, 2012 12:28:17 PM

zanny said:
It depends on context. Back when multi-core was new and becoming mainstream, no games were being developed to take advantage of them, so fewer faster cores often were much better than more slower cores, because any game would still be single threaded.

Today that is very different, and as we go forward a significant chunk of game logic can be offloaded to the GPU via openCL, so in reality game performance cpu wise has plateaued - path finding, collision detection, state awareness, and other frequently done non-visual game actions are easily writable for massively parallel gpus, so you will end up having a vast majority of a game running there.

It will be slow to happen, but as it slowly does, we won't see higher cpu demands for games by much, since anything new can be done gpu side.

In terms of Bulldozer, the 8150 is at a fine price point if you overclock it and don't care about power. It is inefficient in terms of power usage but its price point makes it more economical than a 2500 (soon to be 3570k) and all the new chip brings to the table is even more of the power inefficiency showing itself.

But if you don't pay your electric bill, you really don't need to care. So an 8 core bulldozer chip is "ok" because the i5 2500k / 3570k will only be 10% or so better in "most" games.


Agree with this 100%.


dirtyferret said:
the following current multi-core game does not show stable FPS across the board. certain CPUs offer better performance then others, therefore it is not a plateau. while the GPU does take most of the work load during gaming and no one will see the difference between a 60FPS game and a 80FPS in the real world (that was not your argument), in a benchmark world you can clearly see a numbers difference

http://static.techspot.com/articles-info/507/bench/CPU1.png


Note: DX9.

DX10 onwards moved a LOT of the logic from the CPU to the GPU. Hence why DX10/DX11 games tend to be bottlenecked by the GPU rather then the CPU. As more of the easily parallized tasks get moved to the GPU, the CPU will matter less and less.

We even saw this trend in the BD review a while back: DX11 titles, like Dirt3 and BF3 showed clear CPU bottlenecks, from the 2600k down to the lowly pentium G series. The DX9 titles, by comparision, almost always saw increased performance as the CPU power increased. Thats because most of their processing is still on the CPU, not the GPU.

Hence, using a DX9 benchmark to try and prove a point about CPU/GPU scaling is silly. Run that same benchmark in DX10/DX11, and you get a totally different result.
April 26, 2012 2:05:48 PM

In terms of Bulldozer, the 8150 is at a fine price point if you overclock it and don't care about power. It is inefficient in terms of power usage but its price point makes it more economical than a 2500 (soon to be 3570k) and all the new chip brings to the table is even more of the power inefficiency showing itself.

But if you don't pay your electric bill, you really don't need to care. So an 8 core bulldozer chip is "ok" because the i5 2500k / 3570k will only be 10% or so better in "most" games.[/quotemsg]


NOT!!!!!!!!!!!!!!!!!
a c 185 à CPUs
April 26, 2012 3:50:27 PM

a b à CPUs
April 26, 2012 4:07:20 PM

As the proud owner of a system powered by an 8120 and another system powered by a 2500K, I can assure you that in day-to-day operations, you will not notice any difference between the two machines. In fact, the only time I noticed a real difference in performance was when I put an SSD in the 8120 box, but it inspired me to do the same thing for the 2500K, and we are back to no comparable difference.

Bulldozer is a very good chip. Granted, it's not a knockout when compared to SB, but then, it was developed using a fraction of Intel's R&D budget. Sure, go ahead and badmouth AMD as a failure because they aren't faster than their much larger competitor, and then sometime in the next few years, you will be complaining about how expensive computer processors have become since Intel no longer has any real competition. I'm rooting for the underdog, not because of any sense of loyalty, but because my ulterior motive is to see the power of competition keep prices within range of reason.
a c 283 à CPUs
April 26, 2012 4:22:25 PM

I would like to say that we should just stop replying to threads like this, but there may be 1 out of 1,000 that are't just trolling, so whatever. Reply with caution, I suppose...
April 26, 2012 5:41:45 PM

God all mighty ....I give up......but for me having an 8 core chip performing as a low end 4 core chip and STILL over the price of the 4 (8150)..it is not a good buy. Just because THAT 8 core is labeled as an 8 core chip and is the first "8 core" chip does not make it a great budget price!!! And it is not really an 8 core ...in performance ....If THAT 8 core was performing near the intel's 6' cores like (at best)

http://www.newegg.com/Product/Product.aspx?Item=N82E168...

, and still at that price then wow !! yes that is a good buy !!!

But then again ...we're all are still comparing apples to oranges!!! .. these 2 chips are 2 different pieces of technology that may have the same end ..but are still 2 different pieces of technology ..

can both play games ....YES!!! Can both do it at the same level of efficiency NO ..... games are till GPU bound than CPU mostly anyway ...

comparing the fx chips to previous phenoms II... well that ..still does not look good lol .....
a b à CPUs
April 26, 2012 7:59:52 PM

Short answer: because is more expensive (in most stores that I know of) than the i5 2500k.

Cheers!
a c 83 à CPUs
April 26, 2012 10:06:52 PM

alvine said:
i remember back in the day people would say e8400 4ghz dual core was better for gaming then a 3.4ghz quad core...now people say higher clocked quad is better than a slower clocked eight core

if you want the best fps go for intel
if you are budget oriented (sub 1 grand build) then go for amd.
Intel is faster but you will pay for that, amd is not horrible but it gives you a better value/features if your budget is low

problem there is that it isn't a true 8 core. it has 8 integer cores, but only has 4 FPU's and games rely heavily on FPU performance, along with many other applications. Not to mention the cache is slow. They are the 2 main factors that make bulldozer inefficient.
May 3, 2012 1:54:26 AM

It makes me laugh when I see people talk crap about this CPU without knowing anything about CPU architecture. I am a CPU designer (not X86 or X64 but ARM and the same concept applies). While it is true that Intel chips are faster clock for clock than AMD CPUs, this CPU would easily outperform the newest core i7 Intel in many tasks. The problem is that there is barely any software support, if any at all. There are many important components such as scheduling and queuing that must have 100% support in the software. If an operating system and an application know how to fully schedule all the cores and the task at hand can be broken into multiple threads, there is no way that this FX CPU would lag behind as it shows in the tests. No way. Trust me, I do this every day. Because of lack of software support, the FX chips are crippled in your system and their full power cannot be tapped into. Throw in the right support in the OS and applications and the 8 core chip will easily match the latest core i7 and I am willing to bet that it will even outperform in some tasks. That's a guarantee. Real engineers do not pay attention to benchmarks tests because usually if the test is not biased, it is done incorrectly. Of course, in real world tests, Intel wins. Software optimization and OS support is built around Intel chips and schedulers optimized for the Intel CPU architecture. That's why all tests favor Intel.

A group of friends of mine built a rig using an FX-8150, contacted AMD and obtained programming manuals for the FX chips and technical specs. They took three weeks to write their own microkernel based partially on Linux. Very basic stuff, only support for the 8 core chip among other things were thrown in there. In real world test with the custom OS, the chip outperformed an Intel
Core i7-2700K by 6% for a task of generating encryption keys. This was just a proof of concept to show that with the right software support, this CPU would be significantly better. Get educated and start realizing that benchmarks are done incorrectly. You need to benchmark with the right OS and software support, not without it.

Ask yourselves this question: why is the ipad faster than the best Android tablets with quad core chips? Because Apple ARM chips are faster? No way. It is the software and optimization in the software that makes the biggest difference.
Anonymous
a b à CPUs
May 3, 2012 2:10:37 AM

and it makes me laugh when people cry the chip isn't bad, it the software.

Q. what good is hardware that can't run software?

A. none.

deal with it.
a c 185 à CPUs
May 3, 2012 2:24:51 AM

STOP BRINGING THIS THREAD UP! SOME MODERATOR, PLEASE CLOSE THIS!
a c 146 à CPUs
May 3, 2012 2:26:41 AM

Anonymous said:
and it makes me laugh when people cry the chip isn't bad, it the software.

Q. what good is hardware that can't run software?

A. none.

deal with it.


Exactly what I was thinking. Gotta love the people that cry the software is the peroblem. Next they're going to cry the Intel conspiracy.
May 3, 2012 2:45:09 AM

Hey I am not crying anything here. I could care less which CPU is better. This is not about Intel vs AMD war. I am just stating what is a fact in the industry if you are a real engineer. You could throw all the hardware you want but if you don't have the software support, you are dead in the water. The thread can be closed if required. My intention was only to state a fact of CPU engineering. Like I said, I don't care at all for Intel or AMD.
a c 185 à CPUs
May 3, 2012 2:47:27 AM

Just stop replying, your original post could have started one! :ouch: 
a c 146 à CPUs
May 3, 2012 2:48:45 AM

madscientist55 said:
Hey I am not crying anything here. I could care less which CPU is better. This is not about Intel vs AMD war. I am just stating what is a fact in the industry if you are a real engineer. You could throw all the hardware you want but if you don't have the software support, you are dead in the water. The thread can be closed if required. My intention was only to state a fact of CPU engineering. Like I said, I don't care at all for Intel or AMD.


Like loon said what good is hardware if it can't run software?
May 3, 2012 2:56:12 AM

I agree. I am not saying I do not agree. I agree. Like I said, I am just stating what is a fact. Benchmarkers are benchmarking a crippled chip. There is no true way currently to know whether the chip is actually good. Whether AMD built a good chip there or not is beyond the scope of this discussion and an entirely different topic. The point I was trying to make is that benchmarking this particular chip family the way it's being done is not an accurate test.
May 3, 2012 3:06:47 AM

I am not going to say my opinion here because I know this is a website full of fanbois for both sides who are willing to jump at anything against their belief system. The FX family is a new design and it currently lacks the software support for accurate testing and comparison. That is a fact. OS and software vendors build their optimization using Intel chips. AMD chips get little testing by most vendors. Intel chips are faster and yield better performance clock for clock. These are all facts.
Anonymous
a b à CPUs
May 3, 2012 4:05:11 AM

i don't know if this site is "full of fanbois for both sides who are willing to jump at anything against their belief system". there are SOME but its not the norm. however i digress.

what YOU are doing is coming to a section of a site full of gamers and system builders that exchange experience, tips and knowledge about hardware affecting their main interest and posting about how said software/benchmarking isn't "fair" to a particular piece of hardware.

that makes about as much sense as ME going over to an adobe forum and explaining how intel HD graphics are getting a bad rap because premiere pro is not developed to take full advantage of quick sync.

reverend, you got a good sermon but your preaching to the wrong choir.
a c 283 à CPUs
May 3, 2012 4:11:42 AM

Quote:
reverend, you got a good sermon but your preaching to the wrong choir.


Excellent line, sir.
May 3, 2012 5:24:52 AM

I dont know what these guys are talking about, unless your moving microns over on pluto only benchmarks pull those numbers out of the machines, the codes get updated all the time and as the architecture improves in all of the surrounding hard and software, the true abilities of the advanced architecture in the cpu will be used accordingly, these book and forum heavy guys are not including realtime reactions of every variable at each individual moment. Wether they know it or not a lot of the answers are feeling based and number buffed. Not realtime results at everybodys desk.
May 3, 2012 12:29:24 PM

I apologize if I offended anybody. It was not my intention. Also, I don't mean to preach anything. While you do have a good point that my argument can be brought up about practically anything hardware/software, what I stated is a fact. Benchmarking something right is a lot more than comparing two systems running Windows 7 and some games and some software. To do it right, you have to take lots of things into account. The software optimization required to run the FX chip to the fullest are just not there. That was my point.

A couple of years back I worked on a project for a chip that was theoretically twice as fast as the competing chip. However, when it came time to implement it in the real world, we found the chip performed 40% slower than the competition. The problem was the firmware that was running on the device. It did not have the optimization required to take advantage of the new features in the chip. So we had to abandon that particular design and go with something that would work with that particular firmware.

So you guys know that Intel chips are faster for gaming? So go with Intel. I would gladly pay an extra $100 - $150 for a better chip with better performance and not be stuck with something that doesn't work quite right even if it is selling for less money. I agree with that. I am not denying that at all. The whole purpose of my original post was to point out that the Bulldozer architecture is not getting a fair chance in benchmarks because not even the operating system knows how to correctly schedule the chip not to mention how to take advantage of the new features. I only know of a few compilers that even have these options. This is all fact. I have read the programming guides and technical manuals for the BD architeture and it is a new design. Although I did find several shortcomings in the design that I noticed, with the right software support the chip would perform 20 - 30% better in many applications.
a b à CPUs
May 3, 2012 12:56:28 PM

so what you are saying is that its microsofts fault bulldozer sucks? microsoft sprung windows 7 at the last minute on bulldozer in order to make it look bad compared to intel? the two years plus that win7 was out in the market just wasn't enough time for AMD to run a single test and say "hay is the best performance we can do?" should all bulldozer owners run windows XP instead? was that release in time for bulldozer? or maybe no current software works right with bulldozer and owners should wait for windows 9. buy a FX computer now and wait four years for the right OS to come out?

maybe, just maybe, if AMD wasn't trying to be so cheap and focused on streamlining to increase revenue they could create a chip architecture that serves just the desktop market rather then one that attempts to serve both desktop and server markets while doing neither very well.
May 3, 2012 1:12:17 PM

I don't know whose fault it is or was. The fact is that whoever has one of those in their system whether they run Windows or Linux, they have a crippled chip. I would not buy one, that's for sure. Again, all I was trying to point out is that the benchmarks do not take that into account. Build the right kernel with the correct support and the chip would just run way better.
May 3, 2012 3:38:39 PM

You keep saying it's a "crippled chip", as if it was made lame by anyone but AMD. As dirtyferret just pointed out, Win7 had been out for a while, giving AMD designers and engineers plenty of time to test their "revolutionary" design with the current OS's, and make adjustments where needed. It is AMD, and AMD alone that is responsible for putting an architecture out on the market that is handicapped that much with respect to operating systems. Bad judgment, wrong design, overly hyped marketing.
May 3, 2012 4:06:36 PM

Yeah, ok. However, the design is great, ingenious, just bad implementation.
!