Review Index:
Feedback

AMD FX-8150 Processor Review - Can Bulldozer Unearth an AMD Victory?

Author:
Subject: Processors
Manufacturer: AMD

Bulldozer Architecture

Introduction

Bulldozer.  Since its initial unveiling and placement on the roadmap many have called the Bulldozer architecture the savior of AMD, the processor that would finally turn the tide back against Intel and its dominance in the performance desktop market.  After quite literally YEARS of waiting we have finally gotten our hands on the Bulldozer processors, now called the AMD FX series of CPUs, and can report on our performance and benchmarking of the platform.

With all of the leaks surrounding the FX processor launch you might be surprised by quite a bit of our findings - both on the positive and the negative side of things.  With all of the news in the past weeks about Bulldozer, now we can finally give you the REAL information.

Before we dive right into the performance part of our story I think it is important to revisit the Bulldozer architecture and describe what makes it different than the Phenom II architecutre as well as Intel's Sandy Bridge design.  Josh wrote up a great look at the architecture earlier in the year with information that is still 100% pertinent and we recount much of that writing here.  If you are comfortable with the architeture design points, then feel free to skip ahead to the sections you are more interested in - but I recommend highly you give the data below a look first. 

The below text was taken from Bulldozer at ISSCC 2011 - The Future of AMD Processors.

Bulldozer Architecture Revisited

Bulldozer brings very little from the previous generation of CPUs, except perhaps the experience of the engineers working on these designs.  Since the original Athlon, the basic floor plan of the CPU architecture AMD has used is relatively unchanged.  Certainly there were significant changes throughout the years to keep up in performance, but the 10,000 foot view of the actual decode, integer, and floating point units were very similar throughout the years.  TLB’s increasing in size, more instructions in flight, etc. were all tweaked and improved upon.  Aspects such as larger L2 caches, integrated memory controllers, and the addition of a shared L3 cache have all brought improvements to the architecture.  But the overall data flow is very similar to that of the original Athlon introduced 14 years ago.

As covered in our previous article about Bulldozer, it is a modular design which will come in several flavors depending on the market it is addressing.  The basic building block of the Bulldozer core is a 213 million transistor unit which features 2 MB of L2 cache.  This block contains the fetch and decode unit, two integer execution units, a shared 2 x 128 bit floating point/SIMD unit, L1 data and instruction caches, and a large shared L2 unit.  All of this is manufactured on GLOBALFOUNDRIES’ 32nm, 11 metal layer SOI process.  This entire unit, plus 2 MB of L2 cache, is contained in approximately 30.9 mm squared of die space.

View Full Size

Continue reading our review of the AMD FX Processor (codenamed Bulldozer)!!

It is well known that Bulldozer embraces the idea of “CMT”, or chip multi-threading.  While Intel supports SMT on their processors, it is not the most efficient way of doing things.  SMT sends two threads to the same execution unit, in an attempt to maximize the work being done by that unit.  Essentially fewer cycles are wasted waiting for new instructions or resultant data.  AMD instead chose to implement multi-threading in a different way.  For example, a Bulldozer core comprised of four modules will have eight integer execution units, and four shared 2 x 128 bit floating point/SIMD units.  This allows the OS to see the chip as an eight core unit.

CMT maximizes die space and threading performance seemingly much better than SMT (it scales around 1.8x that of a single core, as compared to 1.3x that using SMT), and CMP (chip multi-processor- each core may not be entirely utilized, and the die cost of replicating entire cores is much higher than in CMP).  This balance of performance and die savings is the hallmark of the Bulldozer architecture.  AMD has gone through and determined what structures can be shared, and what structures need to be replicated in each module.  CMT apparently only increases overall die space by around 5% in a four module unit.

View Full Size

A closer look at the units reveals some nice details.  Note the dual MMX (SIMD-Integer) units in the FP/SIMD block.  A lot of work has been done on the front end to adequately feed the three execution units.

Gone is the three pipeline integer unit of the Athlon.  Bulldozer uses a new four pipeline design which further divides the workloads being asked of it.  These include multiply, divide, and two address generation units.  Each integer unit is fed by its own integer scheduler.  The decode unit which feeds the integer units and the float unit has also been significantly beefed up.  And it had to be.  It is now feeding a lot more data to more execution units than ever before.  The original Athlon had a decode unit comprised of 3 complex decoders.  The new design now features a 4 decode unit, but we are unsure so far how the workload is managed.  For example, the Core 2 had a 4 decode unit, three of which were simple decode, and the fourth was a complex.  My gut feeling here is that we are probably looking at three decoders which can handle 80 to 90% of the standard instructions, while the fourth will handle the more complex instructions which would need to be converted to more than one macro-op.  While this sounds familiar to the Core 2 architecture, it does not necessarily mean the same thing.  It all depends on the complexity of the macro-ops being sent to the execution units, and how those are handled.

The floating point unit is also much more robust than it used to be.  The Phenom had a single 128 bit unit per core, and Bulldozer now has it as 2 x 128 bit units.  It can combine those units when running AVX and act as a single 256 bit unit.  There are some performance limitations there as compared to the Intel CPUs which support AVX, and in those cases Intel should be faster.  However, AVX is still very new, and very unsupported.  AMD will have an advantage here over Intel when running SSE based code.  It can perform 2 x 128 bit operations, or up to 4 x 64 bit operations.  Intel on the other hand looks to only support 1 x 128 bit operation and 2 x 64 bit operations.  The unit officially supports SSE3, SSE 4.1, SSE 4.2, AVX, and AES.  It also supports advanced multiply-add/accumulate operations, something that has not been present in previous generations of CPUs.

In terms of overall performance, a Bulldozer based core should be able to outperform a similarly clocked Intel processor featuring the same number of threads when being fully utilized.  Unfortunately for AMD, very few workloads will max out a modern multi-core processor.  Intel should have a slight advantage in single threaded/lightly threaded applications.  AMD does look to offset that advantage by offering higher clocked processors positioned against the slower clocked Intel units.  This could mean that a quad core i7 running at 3.2 GHz would be the price basis for a 4 module Bulldozer running at 3.5 GHz.

View Full Size

Exact specifications have not been released for the individual parts, but we can infer a few things here.  First off is the fact that it appears as though each core will utilize 2 MB of L2 cache.  This is quite a bit of cache, especially considering that the current Phenom II processors feature 512 KB of L2 cache per core.  Something that has allowed this to happen is buried in GLOBALFOUNDRIES 32 nm SOI process.  They were apparently able to get the SRAM cell size down significantly from that of the previous 45 nm process, and allow it to also clock quite a bit higher.  This should allow more headroom for the individual cores.  With the shrink, we should also expect to see at least 8 MB of shared L3 cache, with the ability to potentially clock higher than the 2 GHz we see the current L3 caches running at.

October 12, 2011 | 01:07 AM - Posted by ZackJ (not verified)

Another great article Mr Shrout. I really appreciate the honesty you provide in your assessment of this processor. Hearing different things on the web I was very interested in this processor. Great new design for the future but it just doesnt seem to compete as well as I would like. I currently have a AMD Phenom II X4 965 and this review has me not really wanting to upgrade to it. I plan to get a new mobo and ddr3 ram so looks like that and a new vid card will be my only purchase in the near future. Although given this review would u recommend the FX-8150 or Phenom X6 1090T/1100T?

October 12, 2011 | 10:24 AM - Posted by Ryan Shrout

It is hard to deny the value of the X6 processors now based on their price. If money is kind of tight, I have no qualms recommend the 1100T.

November 8, 2011 | 07:55 AM - Posted by drbaltazar (not verified)

Ryan and his crew please stop doing bogus article on fx 8150.unless amd or someone at ms or linux has some benchmark that test with fx feature in mind all test will be irelevent.fma4 alone isnt supported and you try to compare actual fma proc .intel. vs fx.comparing a cpugpu vs a cpu?core parking activated?threading issue in window?mobo bios issue that wont be fixed in w8.come on just stop benchmarking this cpu give a call to mobo.ms.amd then when they have fixes you can revisit[about6 month worth]some silly test seams to forget one fact very important here in 2 year intel will have fma4.2 year.amd isso far ahead they cant even speak to any one for fear of copying.it isnt they dont want to.the techno is just plain further ahead then all expected.people wonder why fx is selling like hotcake.when was the last time you had a 2 year in the futurtechno avail today?it rarelly happen .this is huge new for corp in the software business be it gaming or anything else.i bet a lot are hard at work opimising .or trying.for fma4 and all other lessb highlighted feature.but like i mention this proc was released 6 month too early

October 12, 2011 | 01:18 AM - Posted by dreamer77dd

I wonder what Trinity will be like. hmmm

October 12, 2011 | 02:18 AM - Posted by player-x

Nice article only the tables are hard to readout :(

I understand the dilemma of sorting by name or rank.

But personally i really prefer ordering by rank, but that's me.

A great solution would be to have mouse-over change the ordering so everyone can pick whats best for them.

Next to that some color coding would be nice of competing products
Light Blue for i5 2400 blue for 2500K and dark blue for 2600K
next to that dark green for the FX green for the X6 light green for the X4.

To just suggest something.

Like a said, good read, so so tables

October 12, 2011 | 10:26 AM - Posted by Ryan Shrout

Yeah, that's fair. We'll see what we can do to improve that in the future!

October 12, 2011 | 02:27 AM - Posted by Yangorang (not verified)

It's interesting to note that you guys came to a rather different conclusion than Anandtech did with regards to gaming performance with Bulldozer.

http://www.anandtech.com/show/4955/the-bulldozer-review-amd-fx8150-tested/8

I'd definitely like to see some more testing done on this.

October 12, 2011 | 10:30 AM - Posted by Ryan Shrout

Notice they ran their games at resolutions like 1024x768 and at the highest, 1680x1050 while I ran my tests at 1080p. In truth, the higher the resolution the less important the CPU performance tends to be.

To some people, they just want to know the raw gaming power of the CPU so running at low resolutions, sometimes even lower than is likely to be run by the gamer (who plays games at 1024x768 anymore??) will show the biggest differences.

In my case I thought it more pertinent to show the most "real world" cases and 1080p seemed to be the way to go.

October 12, 2011 | 10:07 PM - Posted by Jeremy Hellstrom

You can't argue the Civ V findings, but [H]ard|OCP used similar resolutions and found similar results to Ryan's.

http://www.hardocp.com/article/2011/10/11/amd_bulldozer_fx8150_gameplay_...

October 12, 2011 | 03:36 AM - Posted by Bamstick (not verified)

I would to see more gaming benchmarks. Having only 2 games on there seems lazy to me. Where is Starcraft 2, Bad Company 2, Rage, The Witcher 2, and heck put World of Warcraft on there, you know games that people actually play. I don't anyone who plays Lost Planet 2.

October 12, 2011 | 03:37 AM - Posted by Bamstick (not verified)

Oh and of course Crysis, Crysis Warhead, Crysis 2. Come on!

October 12, 2011 | 08:28 AM - Posted by gwaland (not verified)

Did I miss in the article where you explained why you used a 1090t instead of the top of the line 1100t for most of your benchmarks?

October 12, 2011 | 10:34 AM - Posted by Ryan Shrout

Ah, good point. We used the 1090T results from a previous article (Llano I think) and didn't have time to get in the 1100T to run the full allotment of tests before publication. Instead, with our time we had, I was able to run the 1100T through some our architectural analysis tests (core scaling, etc) and gaming.

October 12, 2011 | 08:55 AM - Posted by Anonymous (not verified)

So... how well does this CPU FOLD????

WOndering if it can handle bigadv folding...

October 12, 2011 | 10:09 PM - Posted by Jeremy Hellstrom

You ain't the only one! I suggested it to Ryan but he hasn't done it since the PS3. Mind you I'm a BOINCer myself.

October 12, 2011 | 09:24 AM - Posted by krankycheese

Based on this review, it's hard to justify upgrading from my Phenom II 955 especially when my PC is used mostly for gaming. I was hoping for better power consumption numbers when compared to what Sandybridge has.

The architecture is intriguing and has potential. It will be interesting to see what AMD comes out with the next iteration.

October 12, 2011 | 09:33 AM - Posted by bjv2370

ill wait for piledriver for improvements...

October 12, 2011 | 09:42 AM - Posted by Anonymous (not verified)

I just finished reading all 3 reviews( Anand, Toms and PCPER) and just like Yangorang said, WTF?!

The test show some consistencies but there is still a rather big difference in attitude and benches towards the 8150FX.

I think there is a bit of Fanboy-ism being implemeneted by ANAND and TOMS ( you can see by the comments as well) review. Granted it may not be a 2600k but its gets pretty close between a i5 and the i7 so I feel that those 2 reviews excerted much more biased in their writing towards the intel chip, even when the BD came close.

There are some crazy things like the power usage, but really? Most of the people posting dont really care about their lightbill( multi gpu, plethora of fans and 1100watt PSUs) so why are people complaining that much?

I already bought me ASUS CH-V 990fx mobo yesterday, and my AMDHD5970 (2gb) so I think I will just push on through with the BD. My last build was a core 2 duo so I thing I will be good non the less.

October 12, 2011 | 10:37 AM - Posted by Ryan Shrout

I don't think it is a BAD processor necessarily, but I find it hard to recommend the FX-8150 over the Core i5-2500k or even the i7-2600k if you are building a new system from scratch.

You have a 990FX Bulldozer-ready motherboard and want to get rid of that older CPU? Sure, the AMD FX will improve your system somewhat.

As I mentioned in my conclusion page, the primary issue is AMD thinks its processors are worth more money they probably are for MOST work loads.

October 12, 2011 | 12:22 PM - Posted by Anonymous (not verified)

Thanks for the review Ryan. I bought a Core i5-2500K and z68 mobo 2 weeks ago, and I'm not regretting my purchase one bit. We'll have to see if that sentiment persists thru to when Ivy Bridge comes out. :P

July 13, 2014 | 09:06 AM - Posted by yacon leaves (not verified)

Thanks very nice blog!

Alsoo visit mmy webpage: yacon leaves

October 12, 2011 | 10:44 AM - Posted by Rion (not verified)

No wonder AMD is attempting to get pile-driver out as soon as possible, they probably knew bulldozer wasn't going to light the world on fire. Now the question is do you wait for Trinity/pile-driver and FM2. Somehow I think most people will wait, unless they already bought the AM3+ board. Isn't Pile-driver and FM2 by Q2 2012?

October 12, 2011 | 10:47 AM - Posted by Irishgamer01

OH Dear,
Doesn't even look as if its worth updating from an X6.
Hope the 7000 cards are good cause AMD could be in trouble.

October 12, 2011 | 10:57 AM - Posted by Nebulis01

Ryan,

Thanks for including an older intel proc (the Q9650) I have a QX9650 and I've been looking to upgrade and was hoping to head back to AMD with this Bulldozer release. Sadly I see a i7-2600 in my future.

Thank you as always for the great review!

October 12, 2011 | 12:46 PM - Posted by Ryan Shrout

Thanks for reading!

October 12, 2011 | 11:58 AM - Posted by pdjblum

Thorough, comprehensive, objective, and very informative review. Well done.

October 12, 2011 | 12:51 PM - Posted by Steven (not verified)

Well this make me wish I hadn't already bought it, since I have an 1100t... So much anticipation, and I suppose I'm about to be let down.

October 12, 2011 | 01:47 PM - Posted by Bill (not verified)

I couldn't care less where AMD goes from here in their lineup. I'm done waiting for their next "fast" cpu, which is only going to be a pathetic 10-15% improvement anyway. I'll have a 2500k under my hood now, and AMD will unfortunately be in my rear view, broken down on the side of the road overheating.

October 12, 2011 | 02:16 PM - Posted by Oskars (not verified)

What is the deal with the performance?
Doesn't it look strange that a 2 billion transistor chip (fx-1850) is a tad slower than a 0.9 billion chip (i7 2600k) of witch 1/4 is a gpu.
There are a few major improvements, but still.
Is that just an unpopular code or a task sheduler comunication mishap? Some people speak of imprvements in windows 8.
Could that be it? That Bulldozer is a year early, and not late at all?

October 12, 2011 | 03:19 PM - Posted by hechacker1 (not verified)

Indeed. It seems ridiculous that 2 billion transistors nets them a slightly slower chip than even their last generation.

It doesn't make sense. I think AMD needs hyper threading bolted on to extract more performance or something. All those transistors are going wasted, or it's just insanely inefficient.

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

By submitting this form, you accept the Mollom privacy policy.