AMD FX-8350 Review: Does Piledriver Fix Bulldozer's Flaws?

[blazorthon]

[citation][nom]AntDX316[/nom]problem is the most overlooked item of cost saving is the electricity usesure it's $100 cheaper than the 3770k but the 125W TDP vs 77W TDP.. in the long run you will save with intel.. and have better performance..[/citation]

That depends greatly on how long you intend to keep your CPU. You seem to forget that many (perhaps most) very high end users don't use the same system for very long and at that point, the power cost may or may not even meat the price difference, let alone recoup it for Intel, depending on how much power is actually being consumes and how much the owner(s) pay for electricity. Furthermore, better performance is extremely situation dependent. Intel doesn't win in everything.

TDP is not equivalent to power consumption, so mentioning it without even stating this fact doesn't help your case. You're generalizing to an extreme and that doesn't work.

 

[agello24]

all i saw was 8 cores vs 4 cores with 4 ghost cores. i want 8 real cores regardless to how it performs.

 

[army_ant7]

all i saw was 8 cores vs 4 cores with 4 ghost cores. i want 8 real cores regardless to how it performs.

I'm not pertaining to any specific model or brand with what I'm about to say. Why would you like 8 physical cores regardless of how it performs (i.e. even if a 4 core can beat it in every single way hypothetically)?

 

[Thomas Creel]

AMD isn't terribly bad like some claim, their color scheme is boss XD

 

[ProSnipor]

I'm planning on grabbing an fx 8350, and a HD 7950. I hope it gives me some big performance jumps over my X6 1035T @3.3ghz and my XFX HD 5770

 

[army_ant7]

I'm planning on grabbing an fx 8350, and a HD 7950. I hope it gives me some big performance jumps over my X6 1035T @3.3ghz and my XFX HD 5770

That HD 7950 would most surely give you an all-around, substantial performance boost, but the FX-8350, when it comes to gaming, it might be best if you refer back to the benchmarks, but when you said "performance" you could've meant productivity programs or both that and gaming.

 

[blazorthon]

[citation][nom]army_ant7[/nom]That HD 7950 would most surely give you an all-around, substantial performance boost, but the FX-8350, when it comes to gaming, it might be best if you refer back to the benchmarks, but when you said "performance" you could've meant productivity programs or both that and gaming.[/citation]

Well, it'd generally beat a Phenom II x6@3.3GHz even in gaming performance, although I agree in that it shouldn't be a "big" gaming CPU performance boost overall.

[citation][nom]agello24[/nom]all i saw was 8 cores vs 4 cores with 4 ghost cores. i want 8 real cores regardless to how it performs.[/citation]

Army covered the flawed performance-related logic in this, but I'll also say this: that is not how Hyper-Threading works.

 

[oxford373]

It's basically an 8150 + 400mhz. There's been no / to little improvement like I predicted. This is a very disappointing offering from AMD. Would not buy.

piledriver cores are 15% faster per hertz than bulldozer in most programs, tom's hardware and many other reviewers said that before.
Quote from this article http://www.tomshardware.com/reviews/a10-5800k-a8-5600k-a6-5400k,3224-2.html

We took the A10-5800K, set it to 3.8 GHz, turned off Turbo Core and any power-saving feature that’d spin the chip down. Then, we took FX-8150, overclocked it to 3.8 GHz, and disabled all of the same features. By running a single-threaded workload like iTunes, we could neutralize the difference in core count (though, if anything, FX could have benefited from its 8 MB L3). Nevertheless, Piledriver clearly completes our workload much faster, yielding a 15% improvement, per clock cycle, over Bulldozer.

Turning off two of FX-8150's Bulldozer modules gives us the opportunity to run a threaded workload like 3ds Max without slanting the result toward Bulldozer. And once again, the Piledriver-based APU wins by roughly 15%.

and here is how to measure per hertz performance using calculator
http://www.tomshardware.com/reviews/fx-8350-vishera-review,3328-2.html
in itunes fx-8150 finished in(1:42)and fx-8350 finished in (1:27)
(1:42)102seconds/(1:27)87seconds=1.1724 so in itunes pildriver is 17% faster per hertz faster than bulldozer.
in 3ds max fx 8150 finished in (2:58)and fx-8350 finished in (2:35)
(2:58)178seconds/(2:35)155=1.14838 so in 3ds max piledriver is 14.8% faster per hertz than bulldozer.

 

[technoholic]

AMD tried to fix its architectural flaws and they did not do so bad, especially if you consider that they did this in a relatively short time. This modular architecture is, IMO, a right step in the right direction. Future of software is (and must be) about parallel workloads, multi-core utilization and/or dividing the calculation amongst many cores, so the work can be done in a shorter time. I know, not everything can be "threaded" in software world, that is another story... AMD only needs to cut the power consumption in their future chips. Let's be wise, if PD was a 95w/77w chip, would you be bashing AMD so hard like you do now?

 

[DjEaZy]

[citation][nom]Katsu_rap[/nom]Vishera Summary:Homerun? No. Improvement? Yes. Power efficiency? Getting better Overclockable? Yes to that Pricing? Not bad at all[/citation]
... and i got one... the last FX waz a 8120... but 4 this price... ?! 8350 it's better... eats less, cooler... a price performer...

 

[greevar]

What I don't understand is why Chris didn't adjust the thread count in the skyrim.ini to match the AMD CPU (8 cores, Skyrim defaults to 4 threads). That may have changed the results of those benchmarks somewhat, maybe significantly.

 

[blazorthon]

What I don't understand is why Chris didn't adjust the thread count in the skyrim.ini to match the AMD CPU (8 cores, Skyrim defaults to 4 threads). That may have changed the results of those benchmarks somewhat, maybe significantly.

That's the first that I've heard of trying that, but to be honest, I'm not sure about it making much of an improvement, if any at all. Skyrim doesn't scale well across more than even two or three cores IIRC and letting it access more than four threads means that it might have some threads sharing a module more than it otherwise would. It might help, but it might hurt instead. It seems like something worth testing, but not necessarily worth doing in a review such as this until it has been tested.

Also, I don't think that they were trying to test games with modified .ini files in a review such as this. That seems like a topic for a different article than a review on a then-new CPU.

 

[greevar]

That's the first that I've heard of trying that, but to be honest, I'm not sure about it making much of an improvement, if any at all. Skyrim doesn't scale well across more than even two or three cores IIRC and letting it access more than four threads means that it might have some threads sharing a module more than it otherwise would. It might help, but it might hurt instead. It seems like something worth testing, but not necessarily worth doing in a review such as this until it has been tested.

Also, I don't think that they were trying to test games with modified .ini files in a review such as this. That seems like a topic for a different article than a review on a then-new CPU.

Well, they're reviewing an 8 core CPU, but they don't seem to be using all 8 cores to their full potential and calling it inferior. I'd really like to see how it handles Skyrim in particular with all 8 cores towing the load.

 

[mousseng]

Well, they're reviewing an 8 core CPU, but they don't seem to be using all 8 cores to their full potential and calling it inferior. I'd really like to see how it handles Skyrim in particular with all 8 cores towing the load.

Right, but most software can't leverage all 8 cores, and due to the way AMD's modular architecture works, running Skyrim on all 8 cores may not actually be its "full potential." Knowing how well that 8-core processor performs in cases of under-utilization (especially when its direct competition is a quad-core) is very important when choosing which platform you want to use.

 

[blazorthon]

Well, they're reviewing an 8 core CPU, but they don't seem to be using all 8 cores to their full potential and calling it inferior. I'd really like to see how it handles Skyrim in particular with all 8 cores towing the load.

I don't think that Skyrim can use all eight cores much if at all (I don't think that it even uses four cores very effectively). If I'm right, then doing what you suggest is probably going to hurt performance, not improve it.

 

[Guest]

the 8350 is not bad, however i would rather get the phenom ii x6 1045t for about 95 dollars from microcenter paired with a hyper 212 evo cooler priced at about 30 dollars, altogether just under 130 dollars and overclock it to about 4ghz and should be decent in performance and price.

 

[greevar]

[citation][nom]jihidi[/nom]the 8350 is not bad, however i would rather get the phenom ii x6 1045t for about 95 dollars from microcenter paired with a hyper 212 evo cooler priced at about 30 dollars, altogether just under 130 dollars and overclock it to about 4ghz and should be decent in performance and price.[/citation]

Well, I'd like to see some empirical evidence rather than speculate.

 

[blazorthon]

[citation][nom]jihidi[/nom]the 8350 is not bad, however i would rather get the phenom ii x6 1045t for about 95 dollars from microcenter paired with a hyper 212 evo cooler priced at about 30 dollars, altogether just under 130 dollars and overclock it to about 4ghz and should be decent in performance and price.[/citation]

FX-8320 and FX-8350 should overclock better than Phenom II x6 CPUs. Whether or not a 1045T@4GHz is decent isn't the point. The point is that an 83xx model @ almost 5GHz (maybe even at 5GHz) would be better.

[citation][nom]greevar[/nom]Well, I'd like to see some empirical evidence rather than speculate.[/citation]

I don't see any speculation in that post except for assuming that the Phenom II x6 1045T unit received will be able to reach about 4GHz. If you misquoted with the intention of quoting me, I'd say that current evidence is quite convincing in that. I agree that it is something worth testing, but an article such as this is not the place for that.

This is the place to ask for such tests, just keep in mind that you should ask for them done in a future article, not state your disregard for this article due to it not having tested something you want to see tested.

As we see today, Skyrim seems to be mostly single through triple threaded since even AMD's top quad-core models don't keep up well at all with the i3s compared to how they usually get pretty close and AMD's high core count models don't do much better. That is why I say that I don't think that your suggestion will help.

Letting Skyrim see more than four cores increases the chances of a module having multiple high-performance threads working on its two cores instead of spreading them one per module. This would potentially lead to a decrease in performance, hence why I think that it'll not only not help, but be counter-productive and hurt performance.

 

[Guest]

i hope amd will create a processor capable of superior single threaded performance, because piledriver architecture is capable of multi-threaded performance, very low power consumption..and of course fast!!

 

[Guest]

i still would rather take the thuban, maybe 4ghz is too much, although i seen a few people on some sites saying they were stable at those speeds and a little higher as well, 3.7ghz with a nice aftermarket cooler should be fine. why not take a phenom ii x6 at 3.7ghz with a nice aftermarket fan for only about 130 dollars? thats a really good deal even if it wont beat the fx 8350 at stock speeds, its still a lot cheaper, 70 dollars cheaper.

 

[synd]

Please people, for the love of god. Stop using the "the review didn't use all 8 cores" excuse.
BD/Vishera aren't a 8 REAL, PHYSICAL, CORES CPUs!

 

[blazorthon]

[citation][nom]synd[/nom]Please people, for the love of god. Stop using the "the review didn't use all 8 cores" excuse. BD/Vishera aren't a 8 REAL, PHYSICAL, CORES CPUs![/citation]

Yes, they are eight real cores. There are only four FPUs and some resources are shared within each module, but there are eight real cores nonetheless.

 

[roza95]

You need to take into account that most games wont use the extra cores amd can offer so while it may not be better than an i7 in years to come it will be as when games require more cores those hyperthreaded ones in the intel aren't going to allow it to compete with full sized amd one's, also the pricing over here in the uk is phenomenal i can get this 8350 for £150 which is about $225 maybe cheaper if i looked around but for an i7 3770k your looking at £240 or roughly $340 its very competitive pricing and its future proof because of those extra core's

 

[hytecgowthaman]

Toms give only lightlyThreaded application result.Toms hardware give result of multithreded application ,so we can find which processor is 1st in multithreaded application.

 

[army_ant7]

@hytecgowthaman
Um... I believe the benchmarks where the FX-8350 beat the Core i5's were highly-threaded.