AMD FX-8150 Bulldozer On Ubuntu Linux

[ciplogic]

Good job Phoronix

I do agree that Phoronix did a terrific job to make those benchmarks. People use the new CPU where is strong and where is strong it shine. In fact we don't care as typical users that we will not get 130 FPS when we will get 80, cause the single threaded performance is not that good, but we care when we have to encode a clip and we do hope that anything is scaled right.
Also I think in some way I do think that AMD have to get a good review as Bulldozer is not a bad CPU. It remembers me to K6 times, K6-2 times, when AMD was not a leader but it will trail well, sometimes worked even better if uses architecture specifics (in K6-2 were the 3DNow! instructions).
I do believe that it will be a good competitor for IvyBridge (it is for middle end of performance of Sandy Bridge, which is quite a feat) when will integrate the GPU core. I will consider to buy an AMD CPU at that time (of course if Linux side work nice).

[mer.at]

anyone knows how to enable TurboCore for the new FX-8150?
i've already loaded the powernow-k8 kernel module, and it says that Turbo is enabled, but the cpu will never clock to 4200MHz, only to 3600. (even when i'm running a singlethreaded program)

dmesg|grep power
powernow-k8: Found 1 AMD FX(tm)-8150 Eight-Core Processor            (8 cpu cores) (version 2.20.00)
powernow-k8: Core Performance Boosting: on.
powernow-k8:    0 : pstate 0 (3600 MHz)
powernow-k8:    1 : pstate 1 (3300 MHz)
powernow-k8:    2 : pstate 2 (2700 MHz)
powernow-k8:    3 : pstate 3 (2100 MHz)
powernow-k8:    4 : pstate 4 (1400 MHz)

kernel version is 3.1

i'm asking here because i saw some benchmarkes where turbocore seems to be working.

http://openbenchmarking.org/result/1110208-AR-AMDFX815044

[nepwk]

This comparison was somewhat unfair. Lemme explain. First of all, fx-8150 is set 400 mhz higher turbo. The clock speed of 2500k and fx-8150 should have been the same because they both are almost identical when it comes to overclocking ability. Clock-for-clock, 2500k should be almost the same as fx-8150 in multithreaded applications. Also, it is not fair at all to compare 8 threads vs 4. 2600k would have been a good processor to use in this comparison. though, i understand since it's an expensive processor to buy just for a benchmark. but hyperthreading alone should give sandybridge enough power to obliterate fx-8150 easily.

WTH is wrong with you people? AMD clearly has a more advanced, forward thinking architecture, and Intel fanboys think if that translates into doing well in any benchmark, that it's "unfair".

Unfortunately, the engineering they did doesn't translate into a clear win in every possible category, but this whining is just unbelievable.

I suppose if Llano or Trinity just stomp Ivy Bridge into the ground in OpenCL benchmarks with their vastly superior GPUs, that OpenCL benchmarks will be unfair in general?

[bug77]

... AMD clearly has a more advanced, forward thinking architecture...

You're absolutely right. It's just that me and a handful of other nutjobs tend not to buy "a more advanced, forward thinking architecture", but just go for the best bang for the buck. Now.

[bug77]

the 1090T should be cheap to buy if you buy it used on ebay.

Well then, is it also cheap for you to buy it on ebay and send it to Michael? No?

[deanjo]

That and the FX-8150 reviews I've read show the CPU being capable of running stable at 4.7-4.9Ghz across all cores with top end air coolers/low end liquid coolers, which puts the performance up there with the 6 core i7s.

I'd wager that with a high end liquid setup you could easily break 5Ghz. High end liquid constitutes a high efficiency block, enough thermal mass in the form of liquid to even out the CPU temperatures at a per core basis and enough surface area on the radiators to never go more then 5c over ambient room temp. Think 4x120+ sized radiators with 1/2" tubing and a 20oz+ reservoir.

Even if all of that wasn't the case, intel and AMD only guarantee that a processor will run at x MHz and that is where they should be benched at. Any over clock is a bonus and core speeds are really only comparable when running against the same family (perhaps even stepping) of processor. Otherwise there are way too many differences that also effect overall performance as well.

[schmalzler]

I suppose if Llano or Trinity just stomp Ivy Bridge into the ground in OpenCL benchmarks with their vastly superior GPUs, that OpenCL benchmarks will be unfair in general?

* AMD A8-3850 with OpenCL on the iGPU brings its Performance on the level of a 2500K.
* Ivy Bridge will be faster then Sandy Brdige.
* Ivy Bridge will get a much better iGPU (read the intel documents on ivy bridge), and as it looks it will be capable of running OpenCL-Code on the GPU.

So I don't think that enabling OpenCL will push Llano or Trinity ahead of Ivy Bridge, though I think with Trinity things get more interesting for desktop users on the amd side (if Global Foundries gets their 32nm-problems away (power consumption) and enhanced Bulldozer really gains 10% performance)

[Kivada]

Could be, yes. But you should also mention, that power consumption increases disproportionally.
Look at this bench for example:

http://www.hardocp.com/article/2011/10/11/amd_bulldozer_fx8150_desktop_performance_review/9

numbers under load:
i7 2600K @stock: 181W
i7 2600K @4,8GHz: 275W
FX-8150 @stock: 258W
FX-8150 @4,6GHz: 452W
Impressive, isn't it?

Meh, it's a desktop, not a laptop, so I don't see the point in worrying about power consumption on a high end desktop. Just look at the people running quad SLi GTX480s.

Now if we where talking a laptop then yes, give me all the cores and cache you can and the BIOS options to underclock and undervolt it as far as it will allow for battery life.

[blackiwid]

I waited for that test because I was interested if its really windows 7 problems that did lead to this bad reviews, and we did saw it was that.

That teaches me to not belive benchmarks from some sides especially from the german site: www.computerbase.de

I did maybe missplaced on a Intel thread (as reaction of a joke about how bad the bulldozer would be) link to this test in the hope that they maybe relativate their topic:

Test: AMD ?Bulldozer?
Ein schwarzer Mittwoch

translated:

Test: AMD ?Bulldozer?
A black Wednesday

instead I earned some Site-or-Intel-fanboys-comments from people who could not accept what cannot be. I tryed to argue further and got now blocked for one week.

I am a bit shocked how biased and unproffessional a hardware-test-site can be.

But thanx again that opened my eyes about this site and I will not use this site again for getting information which hardware I should buy.

Sadly there are not much site that have such good layout, most hardware-test sites have horrible site styles. Phoronix is a kind of good exception but here are not that much hardware-tests.

At the moment I try golem.de but search for better alternatives.

I hope Amd learns from this fanboy-review-debackels and sends first samples to phoronix to not have such myths born like phenom1 or bulldozer1 are bad prozessors.

[crazycheese]

Meh, it's a desktop, not a laptop, so I don't see the point in worrying about power consumption on a high end desktop. Just look at the people running quad SLi GTX480s.

Now if we where talking a laptop then yes, give me all the cores and cache you can and the BIOS options to underclock and undervolt it as far as it will allow for battery life.

Why have these people migrated to 560Ti asap it appeared?
Not everyone is american, not everyone can purchase gallons of fuel to tiny prices. Pity, but fact. Powerconsumption affects me as much as performance.