Bulldozer Owners Real World Thread

[eternalone]

This thread is for people actually running the chip. I want to know your personal opinions and real world experiences running this chip. Is it as some people say slower than an Athlon X4 etc etc. I have been watching video reviews on youtube, and different people say different things. So has your computer actually gotten slower after updating to BD have you seen improvements if any?? Is BD really what the benchmarks say it is, a downgrade?>>

 

[LOL_Wut_Axel]

http://forums.anandtech.com/showpost.php?p=32921484&postcount=15

 

[Chiropteran]

http://forums.anandtech.com/showpost.php?p=32921484&postcount=15

Clearly you missed the point of the thread. Benchmarks are not real world, unless you happen to have a benchmark perfectly tailored to your exact usage.

 

[dma0991]

http://www.legitreviews.com/article/1838/1/

 

[Puppies04]

This thread is for people actually running the chip. I want to know your personal opinions and real world experiences running this chip. Is it as some people say slower than an Athlon X4 etc etc.

Links? Quotes?... No? Ok we are playing that game are we.

I think you are taking a group of people who are quite rightly pointing out that in most scenarios there is an existing product already on the market that outperforms bulldozer either in power efficiency, performance or cost which makes it a poor choice for an upgrade and turning that into a scenario where people are claiming that you would be better off putting a cabbage under your cpu cooler.

Let's just put the kibosh on this right now before it gets out of hand. The OP specifically created a thread for BD owners to discuss their findings; let's honor that request.
-Thanks
ViRGE

 

[biostud]

The only problem with bulldozer is that intel is offering a CPU that is faster, overclocks better and uses less power for the same price. If intel didn't exist bulldozer would be a fine CPU.

 

[frostedflakes]

FX-8120 seemed more responsive than my X4 955 for general usage and multitasking, was afraid because of the poor single threaded performance it might feel quite a bit slower, but that doesn't seem to be the case. Just my $0.02, though. And I haven't used a Sandy Bridge system, for all I know it would feel even faster.

Benchmarks pretty much tell the whole story, just look at a wide variety and draw your own conclusions instead of letting people cherry pick benchmarks to try to convince you that BD is either a turd or the greatest thing ever. There are some things BD is pretty good at, such as heavily threaded tasks, and there are things it's pretty poor at, such as single/lightly threaded tasks. Of course it's fine for general PC usage just like any other modern processor, but if you don't have any brand loyalties or other reasons for preferring Bulldozer (such as using it for one of the handful of computing tasks it really excels in), Sandy Bridge is still the better CPU IMO. Better single/lightly threaded performance, tends to keep up with and trade blows with Bulldozer in heavily threaded workloads, and has lower power consumption to boot. I mean they both may "feel" just as fast during general usage, but if they're about the same price and you're making a completely objective buying decision, Sandy Bridge is clearly superior.

 

[IonusX]

dont own one. but amd has a shot at redemption. both a better octo module and a better hexa module are on the way. if they deliever on how a new cpu should be. then it might make those more worthwhile (fx-6200, fx-8170). while i dont expect it to overnight become the messiah id at least expect (if they do it right) a 6200 sitting near the i7-920 and an 8170 sitting smack dab next to a 980X.
these are lofty goals to say the least but ya know. maybe the meth they were on will wear off by then eh?

 

[soccerballtux]

Clearly you missed the point of the thread. Benchmarks are not real world, unless you happen to have a benchmark perfectly tailored to your exact usage.

video encoding benchmarks are real world.
I believe you are mistaking "all benchmarks" with synthetic 3dmark, (especially) Sysmark benchmarks. Real encoding is about as real world as you can get for encoding.
Cinebench is about as real world as you can get for rendering.

 

[AtenRa]

Currently benchmarking an FX8150, I've voted NO

 

[soccerballtux]

http://www.legitreviews.com/article/1838/1/

The joke is: one of those systems was a Phenom2 X4, the other was a bulldozer, and nobody could tell the difference!

 

[SlowSpyder]

I do not own a Bulldozer. But, I did own a Phenom I 9850 that overclocked a grand 200MHz to 2.7GHz. I have a feeling that, for many users, BD will be much like that Phenom I had. It'll be plenty fast, and in many if not most situations a user wouldn't be able to tell much, if any, of a difference between a faster CPU and the Phenom/BD. With that said, if spending a like amount of money, I wouldd certainly go with the faster and more efficient part. And BD will probably age faster than an overclocked i5. But, I'm curious to read some real world reports as well... not that I have any plans of buying one, but to see if it is that bad or generally decent in everyday use.

 

[Ferzerp]

So now we have gone from the benchmarks must be wrong, to "it feels ok"?

It's kind of like that stupid taste test stunt.

Maybe the claims of incorrect benchmarks were the denial stage, but now people are up to bargaining? "Sure, it is demonstrably poor, but I have spent money on it, and it seems ok to me, so I guess it's not so bad".

No one ever claimed it didn't function as a processor. It is an extremely poor buy that is apparently being made because of feelings about companies instead of the product (or perhaps stubbornness after latching on to the fantasy that it would be the best thing ever, despite the historical performance of AMD processors (only at an advantage when intel had made a truly massive misstep)). I guess it stands to reason that now people are attempting to appeal to emotion instead of easily measurable fact to justify it now.


edit: Just look at the anand bench chart with an 8150 vs. an i7-920. The i7-920 came out in November 2008. That's 3 years before the fx-8150. It costed $40 more on release. And was 1.5-2 moores law cycles before the release of bulldozer, meaning bulldozer should have 3-4x the transistor budget. Yes, a processor that came out that far earlier, trades wins with a late 2011 processor.

http://www.anandtech.com/bench/Product/47?vs=434

 

[LOL_Wut_Axel]

So now we have gone from the benchmarks must be wrong, to "it feels ok"?

It's kind of like that stupid taste test stunt.

Maybe the claims of incorrect benchmarks were the denial stage, but now people are up to bargaining? "Sure, it is demonstrably poor, but I have spent money on it, and it seems ok to me, so I guess it's not so bad".

No one ever claimed it didn't function as a processor. It is an extremely poor buy that is apparently being made because of feelings about companies instead of the product (or perhaps stubbornness after latching on to the fantasy that it would be the best thing ever, despite the historical performance of AMD processors (only at an advantage when intel had made a truly massive misstep)). I guess it stands to reason that now people are attempting to appeal to emotion instead of easily measurable fact to justify it now.

This, pretty much. You can try to justify the FX-4100 and 6100 all you want, but the simple truth of the matter is that both the Phenom II X4 925 and X6 1055T and up are faster while being clocked significantly lower.

 

[lOl_lol_lOl]

So now we have gone from the benchmarks must be wrong, to "it feels ok"?

It's kind of like that stupid taste test stunt.

Maybe the claims of incorrect benchmarks were the denial stage, but now people are up to bargaining? "Sure, it is demonstrably poor, but I have spent money on it, and it seems ok to me, so I guess it's not so bad".

Denial of the competitors benefits (i.e., 2500K vs 8120) is not only delusional but its a loss too. AMD and Intel don't lose. Your simply making the wrong purchase by denying yourself the best possible product for the money.

 

[daveybrat]

The only FX chip that seems to compete well for it's price segment currently is the FX-4100 http://www.newegg.com/Product/Product.aspx?Item=N82E16819103996

For $109 can't really go wrong with a 3.6GHz Quad-Core for budget gaming. It's only competition in that range is the i3 processor.

The FX-6100 at $159 is too close to Intel's i5 series IMO.

 

[LOL_Wut_Axel]

The only FX chip that seems to compete well for it's price segment currently is the FX-4100 http://www.newegg.com/Product/Product.aspx?Item=N82E16819103996

For $109 can't really go wrong with a 3.6GHz Quad-Core for budget gaming. It's only competition in that range is the i3 processor.

The FX-6100 at $159 is too close to Intel's i5 series IMO.

Not sure if serious...

Gets beaten by the i3s and the Phenom II X4s in gaming in every scenario, sometimes by a significant amount. Crappy chip for gaming.

 

[TakeNoPrisoners]

Not sure if serious...
Gets beaten by the i3s and the Phenom II X4s in gaming in every scenario, sometimes by a significant amount. Crappy chip for gaming.

So true it isn't even funny.

Even the X4 955 can beat it at stock clocks which is why the prices never dropped on them...

 

[tweakboy]

I rather have more Bulldozer 8 cores. alto Sandy is 8 core with HT on.
Hows the Opteron 12 core server match up against Sandy and bulldozer.
I rather have more cores then speed in each core. For DAW the bulldozer is enough, for video editing its enough. Ok 1 minute slower in a 20 minute render .. whatever.. People switching to bulldozer over Athlon Phenom II quad will not see a performance difference... If you do DAW or Video Editing, I will rather take 8 core 16 threads.

 

[JumpingJack]

The only FX chip that seems to compete well for it's price segment currently is the FX-4100 http://www.newegg.com/Product/Product.aspx?Item=N82E16819103996

For $109 can't really go wrong with a 3.6GHz Quad-Core for budget gaming. It's only competition in that range is the i3 processor.

The FX-6100 at $159 is too close to Intel's i5 series IMO.

The low thread count performance is very poor on bulldozer, which makes it impractical for a gaming CPU, even at that price. The 4100 will also be in a shared resource scenario for thread count > 2, which pushes the aggregate per thread performance even lower. You are better off from both a cost and power perspectives with a similarly priced i3 or a quad core Phenom II 3 Ghz or higher.

 

[Chiropteran]

Not sure if serious...
*bunch of graphs*

Gets beaten by the i3s and the Phenom II X4s in gaming in every scenario, sometimes by a significant amount. Crappy chip for gaming.

Translating your graphs into reality-

farcry- all 60+ fps, all CPUs are playable.

crysis 2- all cpu under 60 fps, no CPU playable

metro 2033- all CPU under 60 fps, no CPU playable

dirt 3- all cpu 60+ fps, all cpu playable

starcraft 2- useless engine, all cpu under 60 fps, no cpu playable

IMO the playable games are perfectly playable on bulldozer. The badly programmed games lag on i7, i5, and bulldozer. I see some flawed games.

 

[Chiropteran]

video encoding benchmarks are real world.
I believe you are mistaking "all benchmarks" with synthetic 3dmark, (especially) Sysmark benchmarks. Real encoding is about as real world as you can get for encoding.
Cinebench is about as real world as you can get for rendering.

Sure, if you are a professional video encoder. If you don't actually use the program being bench marked the results aren't even remotely real world.

 

[hans030390]

I'm satisfied with my FX 6100 for many reasons, some of which do not apply to the average buyer:

1) I got it for $100 and an Asus M5A97 motherboard for $80 on Black Friday. I don't think you'll get much better CPU for that price (especially not a 6-core). This alone easily makes up for its shortcomings.

2) It's extremely easy to overclock. In a short amount of time, I had it running stable at 4.3GHz (up from 3.3GHz stock). It has tons of OCing headroom.

3) It provides a very noticeable performance upgrade from my Phenom II X3 720 at 3.2GHz.

4) It is not a particularly hot running chip. I had no issues cooling it with my Hyper TX3. It does amazingly well with my Antec Kuhler 620 in a push/pull configuration.

All in all, I've been very satisfied with my purchase. Sure, it's no 2500K, and it doesn't feel quite as fast as the 2500k in real-world situations. When you get to gaming, the difference is even more noticeable. For example, Skyrim ALWAYS runs beautifully on my roommate's stock 2500k + 6950 on ultra. On high settings, my FX 6100 + 6870 offers very smooth performance, but it noticeably drops below 60fps more often than my roommate's rig. But, hey, it was less than half the price!

For some downsides, I have to feed my FX 6100 1.43-1.45V to obtain 4.3GHz. In turn, it draws quite a bit of power at full load. Thankfully, real-world scenarios rarely push the CPU like stress testing and benchmarking will. Idle power consumption is great, and the power consumption under most loads is perfectly acceptable. It's just the stress testing that kills power consumption.

Is it an issue for me? Not really. I might need to upgrade my PSU if I want to OC it further. And given what I've said above, I'm not really seeing any increases in power costs.

 

[Chiropteran]

I bought an FX-8120 yesterday using the Microcenter deal ($120 off a motherboard and FX-8120 for $199). Honestly haven't owned it enough to really talk about it's general performance, but installing windows at least was a breeze I am keeping my prior machine (1090T), it's just a second computer that I wanted to setup in a different room. A huge plus is the low idle power consumption, as I am running that computer 24 hours a day mining bitcoins, but the mining works the GPU only and CPU generally remains idle.

 

[Dravic]

FX 8150 @ 4.67 with H100 using 1.43v

Nice and Snappy, faster than the 1055T @ 3.8ghz I migrated to my after my linux workstation (dual socket w/ opteron 265's from old 2005 amd tech tour) was pulled from service after it killed its second PSU.

Small upgrade, but pretty much a side grade. The only game I play currently is BF3 and it performs well there. While the 1055t @ 3.8 and CF 4890's weren't slow, this 8150 with the new 7970 is down right beastly at 19x12.