Return of AMD FX: My OC'd AMD FX 8150 / 6990 review - Check in Daily for New Results!

[FlanK3r]

SIMON: of course, it hard work show so many benchmarks to others people. I promised before some months continue practice test (no benchmarks) after my first part of FX, but still I have not free time for this ....Still Im working or benchmarking diferents PC systems...But next week Il have holiday time, so..

[polyzp]

Winrar / cinebench 11.5 revisited. Patch scores up!

[polyzp]

Updated 7zip benchmarks up! with Patch!

[polyzp]

DIRT 3 benchmarks up!

[polyzp]

TrueCrypt benchmarks up!

[Radon]

TrueCrypt benchmarks up!

Linux is king! Makes sense as most servers run Linux and that's where AMD can increase their profit margins.

GCC 4.7 will be released soon and will bring another round of Bulldozer and Sandy/Ivy Bridge improvements, Fedora 17 will be the first to ship it in May.

[polyzp]

Symark SiSoftware 2012 Benchmarks up! tomorrow Alien Vs. Predator Revisited!

[FlanK3r]

n1 man! Could be here some FPS from Starcraft II?

[Lightman]

Really good job polyzp!
Very informative thread with lots of data points

[polyzp]

thanks for the feedback guys! more would be nice ^^ , I am also taking all requests for benchmarks. At this rate I will have significantly more benchmarks to evaluate FX as a whole than most review sites. It seems to me that most review sites leave out benchmarks where FX shines, and include more where it doesnt, all coming to the same incomplete conclusion that it isnt very good. People actually think an 1100t is a better buy than a FX 8120 when i have seen countless times where Phenom II is left in the dust by both sandy bridge and FX. Low single core performance is more of a myth than reality. At 4.8 Ghz FX matches a 4.1 Ghz Phenom Core, and at 5 Ghz, matches a 4.25 Ghz phenom Core. So when you overclock both the difference in single core performance begins to minimize. Unless you really push the 1100t above 4.3 Ghz, single core performance will not be better than FX. You might aswell take the extra scaling (Phenom II X6 - 5.92 scaling , FX 8150 -6.66 scaling). Also, with a good enough cooling solution Bulldozer can reach clocks as high as 5.2 Ghz, pushing its single core performance above what Phenom II can achieve. The more you overclock FX , the better the OC benefits you, there doesnt seem to be a bottleneck wall at max OC, while Phenom II is generally known to suffer from this.

[Olivon]

Don't waste time and bandwidth dude. We all know the Truth.

[polyzp]

Don't waste time and bandwidth dude. We all know the Truth.

Most people would disagree and for that reason alone I will continue to waste all the time and bandwidth i like ^^

[BeepBeep2]

thanks for the feedback guys! more would be nice ^^ , I am also taking all requests for benchmarks. At this rate I will have significantly more benchmarks to evaluate FX as a whole than most review sites. It seems to me that most review sites leave out benchmarks where FX shines, and include more where it doesnt, all coming to the same incomplete conclusion that it isnt very good. People actually think an 1100t is a better buy than a FX 8120 when i have seen countless times where Phenom II is left in the dust by both sandy bridge and FX. Low single core performance is more of a myth than reality. At 4.8 Ghz FX matches a 4.1 Ghz Phenom Core, and at 5 Ghz, matches a 4.25 Ghz phenom Core. So when you overclock both the difference in single core performance begins to minimize. Unless you really push the 1100t above 4.3 Ghz single core performance will not be better than FX. You might aswell take the extra scaling (Phenom II X6 - 5.92 scaling , FX 8150 -6.66 scaling). Also, with a good enough cooling solution Bulldozer can reach clocks as high as 5.2 Ghz, pushing its single core performance above what Phenom II can achieve. The more you overclock FX , the better the OC benefits you, there doesnt seem to be a bottleneck wall at max OC, while Phenom II is generally known to suffer from this.

No, not really.

First of all, FX is barely doing 4.6-4.8 GHz stable when Phenom II can easily do 4-4.3 depending on if you have a good or bad chip. FX doesn't do 5GHz stable, we can only hope stepping and process improvements make that possible. Therefore, by your post, low single core performance is a reality. By the way, AMD was already behind intel considerably in single thread performance, Phenom II being "low single core performance" already.

Then you talk about the scaling, yep...BD scales ~6.66x vs Phenom II's 5.9x. However BD has 8 "cores" and Phenom II has 6. BD scales 83% efficient while Phenom II manages 98%. That's worse scaling.

I'm not sure what some people are smoking, but phase change is necessary for 5.2 GHz on Bulldozer for things other than maybe only running Cinebench at dangerous voltages.

[polyzp]

In terms of overclocking ability and ease,

4.6 Ghz is to FX as 4 Ghz is to 1100t
4.8 Ghz is to FX as 4.2 Ghz to 1100t
4.9 Ghz is to FX as 4.3 Ghz is to 1100t
5.0 Ghz is to 4.4 Ghz 1100t

Plus or minus 0.1 Ghz on each for error

[FlanK3r]

In multithread is no doubt about better FX performance stock vs stock. Example rendering, video encoding, adobe working is FX better product. In games is 1100T vs 8150 simillary,still are better quadcores with high clocks (X4 980 BE or FX-4170). In all, FX is about 15% better in practice aplications and simillary in games. If u have no problem with cooling (air/watter), u can use 1.5V with FX for 24/7 ! It is no right, FX must has lower voltage because at 32nm etc (or people forgot for Phenom 65nm vs Phenom II? Phenom I was more sensitive for voltage than Phenom II)

[polyzp]

It is generally accepted that FX 8150 is a better gaming cpu than a phenom II X4 980, and better than an 1100t.

http://www.tomshardware.com/reviews/...ck,3106-5.html

[BeepBeep2]

It is generally accepted that FX 8150 is a better gaming cpu than a phenom II X4 980, and better than an 1100t.

http://www.tomshardware.com/reviews/...ck,3106-5.html

Because Tom's Hardware told you so?
They put FX-8150 and Phenom II X4 975/980 in the same category, and Phenom II X6 1090T/1100T in a lower category. That is a ing joke...
FX-8150 is an equivalent gaming CPU to a 1100T as long as you can overclock it to 4.6-4.8. Actually, it might still be a little behind in a lot of games.

I think that 600 MHz gap you talk about is more like 800 MHz in reality. If I can run my 8150 @ 4.8 24/7, that's fine with me (this one, I can. First one struggled with 4.6 and some others with 8120's can barely do 4.5) I had my 1100T @ 4 to 4.2 before. The reason $250 FX-8150 is able to equal the $190 Phenom II X6 1100T at stock is simple. 1100T's Turbo doesn't work, and FX-8150's does. I've looked at this myself and I have both the CPUs if you'd like me to test that.

It's a fun CPU to play with and overclock...don't get me wrong!...I plan on picking up another some time later on however the performance isn't there.

FlanK3r is spot on this time...~10-20% better multi-threaded performance, slightly decreased single thread performance, and zero gaming benefit over the previous generation which means FX does NOT win against Sandy Bridge or even Nehalem in any type of gaming. (Unless you can measure the invisible 'smoothness factor')

Your gaming (Dirt 3?) and some other tests (Compression, did you use the same files and file formats with equal compressibility?) are laughable because the reviews you inject your results into were on different machines in a different environment. I can select about three different ways to achieve "8x (effective) MSAA" in VISION Control Center. It's one thing to compare these type of results to get a rough idea. It's another to claim your findings in these scenarios as fact or make it seem reputable.

EDIT:
Don't get me wrong, I know you've put a lot of work in. I just wish you'd have written "Return of AMD FX: My OC'd AMD FX-8150 / 6990 observations - Check in Daily for more!" as unfortunately your results are not reputable at all and barely paint a rough sketch of BD performance.

[polyzp]

Yes it is put into the same category, however each catagory is also put in order of performance from top to bottom. Again, when FX can be optimized properly it clearly wins in game tests, and any other tests. Most benchmarks and games FX lacks in are years old.

I strongly believe that if price to performance is not put into consideration, although it almost always is, and you are just looking for the best AMD gaming CPU, that the FX 8120/8150 is the best option period. It is their flagship mainstream CPU, and as much as you try to convince yourself its not, it isnt true.

If you are a gamer these days, you will overclock, and most likely try to achieve the highest overclock you can possibly achieve. Overclocking is FXs thing, thats been made pretty clear :P. Not to mention, some benchmarks really expose K10.5's aging architecture.



http://amdfx.blogspot.com/2012/01/pa...enchmarks.html

Overclockability of Bulldozer = 4.8 Ghz / 3.6 Ghz = 33%
Overclockability of Phenom II X6 = 4.1 Ghz / 3.3 Ghz = 24%

I am not surprised that AMD switched from K10.5 to Bulldozer archetechture, because there was a clearer future. More cores (8 now, 10 later, and eventually more IPC (+10-15% per gen) will eventually dominate. Although, I dont suspect AMD will push into the enthusiast market (500 USD +) just yet.

Ivy bridge leaked performance only shows about a 9-10% increase in performance.. and Intels sticking with 4 cores for mainstream. And the ivy bridge ES is clocked higher. It has a 3.5 Ghz clock / 3.9 Ghz Turbo, with compared to the 2600k's 3.4 clock / 3.8 ghz turbo.

http://www.obr-hardware.com/2012/02/...0k-tested.html

[NEOAethyr]

From what I've seen posted on this forum in th epast, from what i gather anyways, even an overclocked to the max 8150 still underperforms compared to a stock clocked 1090/1100.

However...
I have been noticing that peeps have been able to clock there mem much higher, which could possibly offset the balance a bit in favor of the 8150.
That's ONLY if you can achieve a overclock on both the cpu, nb and the mem.
Compared to the out of the box performance of the 1090/1100, I don't see the BD as being the winner here.
It's more expensive, and the performance is worse.
It clocks higher, yippie do, it's like comparing a 3ghz p4 to a 2.5ghz athlon, not quite a huge difference or anything but just because one is priced higher and clocks higher doesn't mean it's really any better.

It's not just the single threaded performance that is the prob with these BD's.
It's the fact that it's got less cores then the previous gen ^^.
Why in the heck would I downgrade to a quad core when I personally need and use a 6 core ???
I don't care if single threaded performance is double that of what I got, if it's got less cores it's not that useful to me.
Hyper threading means nothing to me, it just screws over single threaded performance even more, by 50%.

This cpu I think is for avg users.
Not power users.
It lets the avgt person play solitar while they watch youtube and whatever they do at the same time.
Great multitasking ability, abiet with a giant catch, performance is lack luster.

I don't speak from personal experience, just from the benchs I've seen in the past from the new cpu's.
If I truly thought it performed better and had more cores (very important to me the core count), then I'de buy one right now.
At the moment I just don't see it happening.

IF it had 8 cores / 16 threads, clocked HIGHER then 5ghz, then I'de buy it.
That could be 2 gens away from now...

So at the moment, do I upgrade my board for the heck of it, just for efi and a potential for a higher mem clock with the same cpu I got already?
Nah, not worth $200 just for a mouse driven interface to play with .

I was very tempted into getting an intel, then an amd server board.
And I realized, intel kinda sucks, to expensive, not enough cores.
Amd is a joke in servers, if I get a quad socket..., I don't get 8x channels, I get the quad chan and that's it, so no point in going amd with more then 2 sockets.
Latest cpu's, that crumby BD's.
Absolutely no point at all in changing over form the 1090/1100 to the 8150 I think.
But that's just my opinion .

I didn't mean this as a rant or anything like that, to long of a post.
I'm just saying, I think these new cpu's have quite a ways to go before they are even worth considering.

Edit:
I gotta get going for now and goto a buddies house to chill for a bit.
Watch some movies and such.
So I can't go through and shorten up the post sorry.
There is one thing I'de like to mention,. you're oc'ed BD cpu's mem performance, the write performance, is lower then my current 2x1gig old mem set.
Copy looks good, but the write looks horrid.
And your clocks are way above mine...

Anyways don't take no offense to my post or anyone else's for that matter, just keep doing what you're doing.
Doesn't hurt to listen to some of those other guys though .
Laters

[polyzp]

If you think the FX 8150 has less cores than the 1100t for any reason you are mistaken. I guess the definition of a core has really changed, but Scaling is better so its pretty much more cores

And i doubt the difference in performance is due to memory alone..

And if you are suggesting that an overclocked FX @ 4.8 Ghz just matches an 1100t's performance out of the box.. thats just plain wrong sorry.

[BeepBeep2]

Yes it is put into the same category, however each catagory is also put in order of performance from top to bottom. Again, when FX can be optimized properly it clearly wins in game tests, and any other tests. Most benchmarks and games FX lacks in are years old.

I strongly believe that if price to performance is not put into consideration, although it almost always is, and you are just looking for the best AMD gaming CPU, that the FX 8120/8150 is the best option period. It is their flagship mainstream CPU, and as much as you try to convince yourself its not, it isnt true.

Your posts make me laugh. You have a few good points but totally offset by the total fanboism...
Did I not state that the FX-8 core CPUs were about equal to Phenom II X6 and Phenom II X4?

"Each catagory is also put in order of performance from top to bottom." ...okay, any competent person knows that Tom's Hardware is NOT a reputable source, and the chart should just be called "Best single thread performance at stock speeds"...Phenom II X6 belongs in the group with FX-8150 and Phenom II X4 to an overclocker...by the way, I hope you noticed that it is Phenom II X4 with 2 more cores...

"Overclockability of Bulldozer = 4.8 Ghz / 3.6 Ghz = 33%
Overclockability of Phenom II X6 = 4.1 Ghz / 3.3 Ghz = 24%"

Again I'll not that turbo does NOT work on Phenom II X6 95% of the time, and it DOES on Bulldozer with Turbo Core 2.0. Make that comparison 4.8 / 3.9, not 3.6. Trust me, that is one reason FX does as well in gaming (because it's bouncing between 3.9 and 4.2 depending on the game) as the 3.3 GHz 1100T. ...and it more so explains why 1100T scores 5.9x multi-core efficiency in Cinebench.

Also, 4.8 GHz stable is with a good chip and a hefty amount of voltage, plus a water loop with at least a 4x120mm radiator and recent waterblock. I can not cool my FX CPU enough to retain stability in Prime95 at 4.8 GHz. Realistically a good OC is 4.6. If you want to compare maximum to maximum, then make that 4.8 vs 4.2 or 4.3...a few guys on the forum could bench 4.5 GHz in Cinebench with their "better" X6 CPUs too if you want to compare a "5.1 GHz max benchable" to 4.5.

If you think the FX 8150 has less cores than the 1100t for any reason you are mistaken. I guess the definition of a core has really changed, but Scaling is better so its pretty much more cores

And i doubt the difference in performance is due to memory alone..

And if you are suggesting that an overclocked FX @ 4.8 Ghz just matches an 1100t's performance out of the box.. thats just plain wrong sorry.

No, scaling is worse. 6.66 / 8 is a lower number than 5.9 / 6.

As far as memory, had Phenom II X6 had an IMC like BD does, it would have had better IPC for sure. BD's IMC at a measly 2400-2600 MHz is more similar to Phenom II X6's IMC at 3600-3800+ and at a handicap in MHz too.

NEOAethyr is extremely wrong as far as his claims about performance...he must not have been around for a really really long while or something.

[polyzp]

since when is Tomshardware not a reputable source? Theres no point in arguing over OC numbers, my original statement still stands and is true for ease of overclockability

4.6 Ghz is to FX as 4 Ghz is to 1100t
4.8 Ghz is to FX as 4.2 Ghz to 1100t
4.9 Ghz is to FX as 4.3 Ghz is to 1100t
5.0 Ghz is to 4.4 Ghz 1100t

with +- 0.1 ghz error

[BeepBeep2]

since when is Tomshardware not a reputable source? Theres no point in arguing over OC numbers, my original statement still stands and is true for ease of overclockability

4.6 Ghz is to FX as 4 Ghz is to 1100t
4.8 Ghz is to FX as 4.2 Ghz to 1100t
4.9 Ghz is to FX as 4.3 Ghz is to 1100t
5.0 Ghz is to 4.4 Ghz 1100t

with +- 0.1 ghz error

Because they post bull. Their overclocking has always sucked because reviewers didn't know how to operate the systems they were testing...
They boast the name "The Authority on Tech" however the website is oriented for those who are new, and can't learn themselves because that's who THEY are.

As far as your opinion, not "statement" standing, I still disagree.

FX is at 4.5 GHz on the AMD "water cooler" which is essentially an H70...very reasonable for the cooling they had.
1100T was at 4.15 GHz, AND on a 990FX board in which it loses a considerable amount of efficiency.




3DMark11, is extremely multi-threaded aware, (just look at 2600K vs 2500K and the real world...)


CPU bound scenarios, difference is much smaller in GPU bound high resolution scenarios, until you add multiple cards in crossfire...so I feel a little bad for FX in the following:






FX is definately not for those with true 120hz screens, if any of those exist. Neither is Phenom II, but you'd have thought AMD could make at least a small increase in single/up to 4-6 thread performance at all over 1 and a half years.

EDIT:
Turbo-


Power consumption-

[Dumo]

A comparo thread between BD and any other chip (AMD or Intel) most likely will ended up with bad shouting match. For some reason a lot of peep still can't forget how BD inflated expectation burst to the ground when retail chip became available.

It is what it is..

@ polyzp....Thanks for your hard work and very informative thread

[polyzp]

A comparo thread between BD and any other chip (AMD or Intel) most likely will ended up with bad shouting match. For some reason a lot of peep still can't forget how BD inflated expectation burst to the ground when retail chip became available.

It is what it is..

@ polyzp....Thanks for your hard work and very informative thread

haha, I didnt want it to come to that either. Thanks for the feedback! more results tmw!