Wednesday, October 12th 2011

Review Consensus: AMD FX Processor 8150 Underwhelming

It's been in the works for over three years now. That's right, the first we heard of "Bulldozer" as a processor architecture under development was shortly after the launch of "Barcelona" K10 architecture. Granted, it wasn't possible to load close to 2 billion transistors on the silicon fab technology AMD had at the time, but AMD had a clear window over the last year to at least paper-launch the AMD FX. Delays and bad marketing may have cost AMD dearly in shaping up the product for the market.

After drawing a consensus from about 25 reviews (links in Today's Reviews on the front page), it emerges that:
  • AMD FX-8150 is missing its performance expectations by a fair margin. Not to mention performance gains in its own presentation, these expectations were built up by how AMD was shaping the product to be a full-fledged enthusiast product with significant performance gains over the previous generation
  • AMD ill-marketed the FX-8150. Hype is a double-edged sword, and should not be used if you're not confident your offering will live up to at least most of the hype. AMD marketed at least the top-tier FX-8000 series eight-core processors as the second coming of Athlon64 FX.


  • FX-8150 launch isn't backed up by launch of other AMD FX processors. This could go on to become a blunder. The presence of other FX series processors such as the FX-8120, six-core and four-core FX processors could have at least made the price performance charts look better, given that all FX processors are unlocked, buyers could see the value in buying them to overclock. TweakTown took a closer look into this.
  • There are no significant clock-for-clock improvements over even AMD's own previous generation. The FX-8150 drags its feet behind the Phenom II X6 1100T in single-threaded math benchmarks such as Super/HyperPi, the picture isn't any better with Cinebench single-threaded, either.
  • Multi-threaded data streaming applications such as data compression (WINRAR, 7-ZIP) reveal the FX-8150 to catch up with competition from even the Core i7-2600K. This trend keeps up with popular video encoding benchmarks such as Handbrake and x264 HD.
  • Load power draw is bad, by today's standards. It's not like AMD is lagging behind in silicon fabrication technologies, or the engineering potential that turned around AMD Radeon power consumption figures over generations.
  • Price could be a major saving grace. In the end, AMD FX 8150 has an acceptable price-performance figure. At just $25 over the Core i5-2500K, the FX-8150 offers a good performance lead.
  • Impressive overclocking potential. We weren't exactly in awe when AMD announced its Guinness Record-breaking overclocking feat, but reviewers across the board have noticed fairly good overclocking potential and performance scaling.
In all, AMD FX-8150 has almost become another example to cite at a marketing class, of how to effectively handle hype. It is sure to underwhelm some. If it's any compensation, Duke Nukem Forever is still the most underwhelming development this year for the gamer-overclocker community.
Add your own comment

450 Comments on Review Consensus: AMD FX Processor 8150 Underwhelming

#1
Super XP
CDdude55 said:
I really don't see what you're saying, those benchmarks are just again proving that if you have a 6 core Phenom II chip you might as well stick with it.

The BF3 benchmark was the only game that really showed BD in a good light considering it's multithreaded, it also only showed that the peak framerate was better but the average framerate was still only about 3.8% better then Sandy Bridge.

Again, i think it's just people trying to make up excuses again to justify the chips existence/performance.
I know Bulldozer is very imature at the moment, my point in that post was websites are starting to re-test Bulldozer and it's performing better than first thought. Is there an issue with the ASUS Crosshair V Formula? Don't know, but what I do know is every single review site got the same package from AMD which included that motherboard, the motherboard I also have pending my Bulldozer purchase.

I suppose in about 2 to 3 weeks we will see more reviews go up with newer updated bioses and different system config's.

Anyhow in regards to those review links, check out the benchmarks, Bulldozer was right up the 2600's alley.
Posted on Reply
#2
entropy13
Super XP said:
Here is a QUOTE from another site.
Something is going on with the Motherboards and the Bulldozer CPU.
This quote seem to make a load of sense. Also no reason to buy a PII when Bulldozer is out.
Following your reasoning then, since there's a slight increase in performance of using an ASUS Maximus IV Extreme-Z over the Z68A-GD65 (which was the Sandy Bridge motherboard in the [H] review) they should also retest everything, but with a different board for the AMD and Intel CPUs?
Posted on Reply
#3
CDdude55
Crazy 4 TPU!!!
Super XP said:
I know Bulldozer is very imature at the moment, my point in that post was websites are starting to re-test Bulldozer and it's performing better than first thought. Is there an issue with the ASUS Crosshair V Formula? Don't know, but what I do know is every single review site got the same package from AMD which included that motherboard, the motherboard I also have pending my Bulldozer purchase.

I suppose in about 2 to 3 weeks we will see more reviews go up with newer updated bioses and different system config's.

Anyhow in regards to those review links, check out the benchmarks, Bulldozer was right up the 2600's alley.
I think you're trying to find excuses for the chip since you seem like a heavy AMD fan (just basing it off your past posts), you can't cope with the idea that they won't be as good as Intel's current line of chips so you try to find anything to push the illusion that it's still a good buy against the competition.

Again, saw all those benchmarks, the only thing that is wrong is the CPU their using, the boards are fine. Naturally a different motherboards/BIOS will perform differently, but it still doesn't do anything of significance for the chip, you'd STILL be better off with an old Phenom II or if you really want performance, go Intel.

My system is all AMD, so no bias here, i see the reviews and it's crap. Will it improve?, maybe in the future, but it will take a while, it won't be an overnight BIOS update or motherboard replacement, thats a bunch of bullshit.....
Posted on Reply
#4
Damn_Smooth
CDdude55 said:
I think you're trying to find excuses for the chip since you seem like a heavy AMD fan (just basing it off your past posts), you can't cope with the idea that they won't be as good as Intel's current line of chips so you try to find anything to push the illusion that it's still a good buy against the competition.

Again, saw all those benchmarks, the only thing that is wrong is the CPU their using, the boards are fine. Naturally a different motherboards/BIOS will perform differently, but it still doesn't do anything of significance for the chip, you'd STILL be better off with an old Phenom II or if you really want performance, go Intel.

My system is all AMD, so no bias here, i see the reviews and it's crap. Will it improve?, maybe in the future, but it will take a while, it won't be an overnight BIOS update or motherboard replacement, thats a bunch of bullshit.....
I think with lower prices and a power consumption fix, we could have a contender. I can't see that happening any time before the 8170 though. If even then.
Posted on Reply
#5
Goodman
FreedomEclipse said:
any retailer that jacks up prices of older chips to take advantage of the situations like these deserves to be shutdown.
Twice... just to be sure ;)

The computer store around here price of the PIIx6 is just under the FX8120 but it cost more then the FX6100 (except for the 1055T) , why am i not surprise? :laugh: :rolleyes:

Posted on Reply
#6
CDdude55
Crazy 4 TPU!!!
Damn_Smooth said:
I think with lower prices and a power consumption fix, we could have a contender. I can't see that happening any time before the 8170 though. If even then.
I agree.

Anyways keep in mind, i wrote that post that brash mostly cause i'm tired, got home late, so im mad, and then i go on my favorite tech site to find more crap about how there's ''untapped potential'' in Bulldozer. More bullshit about how ''if you make this small tweak, it will change the entire performance of the chip.'' when back in reality it performs like shit and will be shit until they actually focus on the chips themselves, stop blaming the motherboards, stop blaming Windows 7.

Damn i need to sleep, still a bit angry. Be back in the mourning!!
Posted on Reply
#7
Damn_Smooth
CDdude55 said:
I agree.

Anyways keep in mind, i wrote that post that brash mostly cause i'm tired, got home late, so im mad, and then i go on my favorite tech site to find more crap about how there's ''untapped potential'' in Bulldozer. More bullshit about how ''if you make this small tweak, it will change the entire performance of the chip.'' when back in reality it performs like shit and will be shit until they actually focus on the chips themselves, stop blaming the motherboards, stop blaming Windows 7.

Damn i need to sleep, still a bit angry. Be back in the mourning!!
I definitely see where you're coming from man. I expected more from BD too. I still think they have the potential to make something good with the architecture but I would have preferred another delay over what they released.

Have a good night. :toast:
Posted on Reply
#8
Goodman
CDdude55 said:
More bullshit about how ''if you make this small tweak, it will change the entire performance of the chip.'' when back in reality it performs like shit and will be shit until they actually focus on the chips themselves, stop blaming the motherboards, stop blaming Windows 7.

Damn i need to sleep, still a bit angry. Be back in the mourning!!
No! man you're not sleepy you're well awake , that is what i think also...:toast:
Posted on Reply
#9
wolf
Performance Enthusiast
these CPU's should have never been called FX, that name is now forever tainted by this release. the could have done all the dev work and named the chip's just a month or two before release, I have a feeling if it had been done that way they wouldn't be FX CPUs.
Posted on Reply
#11
Super XP
entropy13 said:
Following your reasoning then, since there's a slight increase in performance of using an ASUS Maximus IV Extreme-Z over the Z68A-GD65 (which was the Sandy Bridge motherboard in the [H] review) they should also retest everything, but with a different board for the AMD and Intel CPUs?
No why should they, Sandy Bridge is not Brand New Built from the ground up. Bulldozer is. The company to blame for the inconsistenty in performance and incompatability via mobo's and such is AMD period. My issue is people are not giving them credit for putting out innovation and trying something new.

We all now know the history about why Bulldozer was not as good as we thought. Upper Management made stupid decisions and now AMD is paying for them, hopefully in the short term until Piledriver comes out.

Bulldozer is far from being a wash though, they may not outperform as much as we would have liked, it's still a great CPU. And it still holds it's ground in gaming along with many other apps, it just needs to be priced about $50 less right now.

As for most of the reviews on October 12, 2011, most if not all need to re-Benchmark Bulldozer with proper updates from scratch.
Posted on Reply
#13
CDdude55
Crazy 4 TPU!!!
Super XP said:
No why should they, Sandy Bridge is not Brand New Built from the ground up. Bulldozer is. The company to blame for the inconsistenty in performance and incompatability via mobo's and such is AMD period. My issue is people are not giving them credit for putting out innovation and trying something new.

We all now know the history about why Bulldozer was not as good as we thought. Upper Management made stupid decisions and now AMD is paying for them, hopefully in the short term until Piledriver comes out.

Bulldozer is far from being a wash though, they may not outperform as much as we would have liked, it's still a great CPU. And it still holds it's ground in gaming along with many other apps, it just needs to be priced about $50 less right now.

As for most of the reviews on October 12, 2011, most if not all need to re-Benchmark Bulldozer with proper updates from scratch.
I still disagree, i believe they will need to revise the actual chip design to make a difference in performance, updating the BIOS and swapping out the motherboards isn't the solution for the problem, because the chip is the problem. I commend them on trying a new design, i'm so glad they did cause K10 was ancient.

I'm personally waiting till Piledriver, but if they revise the current generation BD chips, that'd be great too.

I don't see a point in having all reviewers re-benchmark Bulldozer.
Posted on Reply
#14
HalfAHertz
Well the best you could hope for is a CPU driver like for the old Athlon x2's and a new stepping. If each of those gave you 3-5% perf. then at least BD would be in the same ballpark as the PhII x6...
Posted on Reply
#15
CDdude55
Crazy 4 TPU!!!
HalfAHertz said:
Well the best you could hope for is a CPU driver like for the old Athlon x2's and a new stepping. If each of those gave you 3-5% perf. then at least BD would be in the same ballpark as the PhII x6...
True that would help, still wouldn't be a enough to warrant an upgrade from a Phenom II just yet though.

If they can manage to make BD better then their current Phenom II chips at least in multi-threaded applications, only then would that warrant an upgrade for me. If they can keep the price down too along side that, i'd be a winner.:)

The TDP is also fairly high with BD, i'd also be great if they can lower the power consumption and heat output.

I would of bought one a few days ago at launch if they had done such things.
Posted on Reply
#16
cadaveca
My name is Dave
CDdude55 said:
updating the BIOS and swapping out the motherboards isn't the solution for the problem, because the chip is the problem.
Erocker put it best...it IS a bulldozer....massive amounts of power possible, but very slow.

Put ia big pile of work in front of it, it will push right on through, but small pile isn't going to move any faster than big one.


It almost seems to me like perhaps AMD hoped for something like 5 GHz stock clocks, but fabrication issues prevented that goal from being reached.

To me this makes sense considering that the rumour is that a recent firing @ AMD was due to AMD not securing other foundries i ncase a situation like this happened...plain and simple, AMD didn't have a backup plan.


That said, I'm not going to expect as much as you are from revisions. I expect a die shrink will up clocks, and only at that point will AMD have a good chip, as SKT1155 products won't be the top performer for Intel, SKT2011 will.

So, Intel needs to be SKT1155 only, for the desktop space. Recent news says that Intel is lowering the TDP of 1155 with ivybridge, which could possibly mean that the clocks will not go up much more on 1155, for fear of outshadowing SKT2011.
Posted on Reply
#17
Super XP
CDdude55 said:
I still disagree, i believe they will need to revise the actual chip design to make a difference in performance, updating the BIOS and swapping out the motherboards isn't the solution for the problem, because the chip is the problem. I commend them on trying a new design, i'm so glad they did cause K10 was ancient.

I'm personally waiting till Piledriver, but if they revise the current generation BD chips, that'd be great too.

I don't see a point in having all reviewers re-benchmark Bulldozer.
Well yes I can see them taking Bulldozer's Design and manipulating it with massive tweaking for Piledriver.
Posted on Reply
#18
Covert_Death
all i gotta say is i really wanna see piledriver more focused on desktop enthusiasts and released Q1 of next year for AM3+

i get that BD was more server focused and thats totally fine but they need to give us desktop enthusiasts something to upgrade too and i really hope it is a much better upgrade when PD comes rolling in to town
Posted on Reply
#19
ensabrenoir
5....4...3,,,,2...1

let the over hyping of pile driver begin!:rockout: What harm could it do?:rolleyes: :banghead:
yep.... intels going down... hard!!!!( going down a road made of cold hard cash cause when you make a Ferrari u can charge Ferrari prices):pimp:
Posted on Reply
#20
eidairaman1
The Exiled Airman
ensabrenoir said:
let the over hyping of pile driver begin!:rockout: What harm could it do?:rolleyes: :banghead:
yep.... intels going down... hard!!!!( going down a road made of cold hard cash cause when you make a Ferrari u can charge Ferrari prices):pimp:
Problem is this is not a ferrari :roll:
Posted on Reply
#21
Covert_Death
lol im not saying PD is gonna smoke intel... honestly i don't care, but i have an AM3+ board and would like to actually upgrade to something "competitive" if PD offers that on AM3+ then i will get it, if BD gets revised and looks better then i'll upgrade to that, i really don't care how it preforms to intel but how it performs with what i currently have
Posted on Reply
#22
Neuromancer
Well AMD wont be going anywhere, they still have a strong GPU division.

As for CPUs, they wont be going anywhere either, in both VLPU and Server segments they have a strong enough base.

Having said that, I hope that they take the criticism, both from review sites, readers, owners and past employee testimony to redouble their efforts and make the next architecture more worthy. In the mean time they can try and improve on the design issues of the current gen.

The design of BD could be a success if for low threaded say up to for, it actually did use the 256bit FP, and saw a marked increase in performance as a result.

Perhaps it is up to using the right applications to take advantage of the design, as well as fixing the threading issues, I do not know enough. For now at least, until more useful information is out about it, I am holding off and sticking with my Thuban.

What would be interesting is if they released a good 4 core version not based on 2 modules but on 4 modules with 4 of the integer units designed out. (Or is that Llano? I will know soon enough anyway, FM1 board is here and CPU should be soon :) )
Posted on Reply
#23
Super XP
Posted by seronx
New things in Piledriver that may be in some or all Piledriver CPUs:
Input/Output Memory Management Unit Version 2
Fused Multiply-Add 3
Converged Bit Manipulation Instructions
Trailing Bit Manipulation Instructions
Increased L1 DTLB size 32 entries to 64 entries
IPC & Power Management Improvements
Turbo Core Version 3

Piledriver CPUs:
Trinity
Viperfish(Sepang/Terramar/Vishera?)

What to expect:
10-15% increase in performance ( I think they can squeeze out a lot more with additional Tweaking of the original Bulldozer with the above updates. I say approx: 20% to 25%)
or
5-7.5% increase in performance and 5-7.5% decrease in power usage
or
10-15% decrease in power usage
:D
Posted on Reply
#24
Athlonite
Why's everyone comparing the power consumption of a 4core to an neutered 8core that's like comparing apples to bananas ofcourse it's going to use more power it's got more actual cores DUH

evan if it does perform worse watt for watt (that's just bad design)
Posted on Reply
#25
Horrux
Athlonite said:
Why's everyone comparing the power consumption of a 4core to an neutered 8core that's like comparing apples to bananas ofcourse it's going to use more power it's got more actual cores DUH

evan if it does perform worse watt for watt (that's just bad design)
And it also has more than double the total transistors, at a larger process node.
Posted on Reply
Add your own comment