Found another one: Ukraine store accidentally ships FX-8120 and it gets tested!

[bronxzv]

I wouldn't know because AIDA 1.85 is the one that supports Bulldozer!

well, it looks like you were right for the x87 path:

http://www.aida64.com/product/aida64-extreme-edition/benchmarks

"FPU SinJulia Benchmark" is 80-bit x87 only
"FPU Julia Benchmark" is 32-bit FP multiple paths (x87, SSE, AVX,...)

for a proper review it will be far better to use "FPU Julia Benchmark" and the latest version of AIDA with an FMA4 path

 

[NostaSeronx]

well, it looks like you were right for the x87 path:

http://www.aida64.com/product/aida64-extreme-edition/benchmarks

"FPU SinJulia Benchmark" is 80-bit x87 only
"FPU Julia Benchmark" is 32-bit FP multiple paths (x87, SSE, AVX,...)

for a proper review it will be far better to use "FPU Julia Benchmark" and the latest version of AIDA with an FMA4 path

Oh cool

XOP should technically replace x87 as it can do trignometry and exponential in 128bit and 256bit(I think might be just AVX)

 

[bronxzv]

Oh cool

XOP should technically replace x87 as it can do trignometry and exponential in 128bit and 256bit(I think)

for practical purpose I suppose nobody (for commercial projects) is still using x87 transcendentals anyway, they are incredibly slow (sluggish microcode) on any target, software polynomial approximations are way faster and *vectorizable*

FYI neither AVX nor XOP provide native transcendental functions

 

[NostaSeronx]

FYI neither AVX nor XOP provide native transcendental functions

Most be a different ISA then

 

[bronxzv]

Most be a different ISA then

LRBNI has transcendentals though they are emulated on the 1st MIC products

 

[Skurge]

When is this thing releasing?

I'm sitting with a z68 and a 990FX infront of me. Want to know which path I should go.

 

[Tsavo]

When is this thing releasing?

I'm sitting with a z68 and a 990FX infront of me. Want to know which path I should go.

12th of this month.

Go 2500k if you want a CPU that can do everything well.

 

[3DVagabond]

When is this thing releasing?

I'm sitting with a z68 and a 990FX infront of me. Want to know which path I should go.

A bird in the hand is worth 2 in the bush. SB is a known commodity and is an excellent platform. BD might be as well. If not though you won't be getting a whole lot for that 990 mobo after it releases.

Do you play it safe? Or, roll the dice?

 

[Skurge]

12th of this month.

Go 2500k if you want a CPU that can do everything well.

Yeah, The prices of the hardware here has gone up cause of the exchange rate. Since it's coming down thought I'd wait.

I do prefer the 990fx though. Will see in a few days.

 

[SlowSpyder]

One thing I don't understand, let's say BD is as terrible as many of you are saying it is. Why would AMD release it? Wouldn't it make more sense to shrink Thuban to 32nm, maybe tweak the cache, NB, even the cores a bit and go forward with that?

 

[TemjinGold]

Why did AMD release Phenom 1?

Skurge: The odds of BD blowing SB out of the water are pretty low. You are basically either getting known solid performance with SB or hoping for the chance that BD will more or less match it. I can understand a risk when there is a decent chance at a decent reward, but this particular risk makes no sense...

 

[Skurge]

Why did AMD release Phenom 1?

Skurge: The odds of BD blowing SB out of the water are pretty low. You are basically either getting known solid performance with SB or hoping for the chance that BD will more or less match it. I can understand a risk when there is a decent chance at a decent reward, but this particular risk makes no sense...

It's not a risk. I haven't paid for either board. I have them with me unopened. If BD sucks, I'll just give the 990FX back and pay for the Z68. If it matches SB, I'll keep the 990FX as I like it more.

But It isn't looking good for AMD though.

 

[toolbag]

12th of this month.

Go 2500k if you want a CPU that can do everything well.

Go AMD if you wanna be cool.

 

[Ferzerp]

AMDzone is leaking again, I see.

 

[JWade]

my head hurts from trying to shift through all the posts and comparisions of BD and SB. I understand wanting to compare a 4ghz BD to a 4ghz SB, but to me a ghz to ghz comparision is not what I am interested in. I am more interested in how does the BD compare to a similiar priced intel cpu? it may be posted already but with my eyeballs jumping out of my head i cant seem to see it/them

 

[SithSolo1]

I woke up this morning and a thought occurred to me.

Ok, so synthetic benchmarks are designed to max out processors with extremely heavy loads. The kind of which you normally don't see in everyday work, gaming, media play back, etc. BD uses a shared resource per "core pair" correct?

What if AMD designed the chip with the assumption that under daily use "core pairs" would never be at 100% load. What if BD core pairs reach peak efficiency under 80-90% load, but above 80-90% load efficiency drops like a rocket propelled rock since the shared resource can't keep pace. To my thinking that would make it trash for synthetic bench marking but under real world conditions it might actually perform admirably.

Again I am making a lot of assumptions without having the technical know-how to back them up.

 

[BlueBlazer]

The reivewer speaks......

Latest BIOS, latest drivers. Unfortunately this is how it performs... And that part in me that owned Athlon64 and was looking up at the FX series all my life is now crying

Also here.....

Why does Lab501 get to break NDA?!

In order to break an NDA you have to sign one first...

Your power consumption graph on page 14 has the colors reversed showing the 8150 using 93.12watts and the 2600 using 129.72, although you indicated the correct values in the paragraph bellow.

Correct, we just fixed that, thank you for noticing, and also thank M.Beier for that!

I suppose there is also a small chance that all the final BIOS updates for the BD won't be released until the day of the launch

Actually no, there is no chance, because you need reviewers to have time to test this stuff before publishing a review. Nobody makes reviews in one day...

I suppose there is also a small chance that all the final BIOS updates for the BD won't be released until the day of the launch

I know!!! Trust me, we retested many times, because we found this hard to believe even if the results unveiled under our eyes. We checked everything countless of times, and still get the same results. Unfortunately, this really is how Bulldozer is performing, and that makes me as sad as any other enthusiast in this industry!

Now the saddest thing is not only the performance, but also the way in which AMD tries to manage this. I heard from a little birdie that some folks at AMD will start calling press tomorrow morning to ask them how reviews are going and try to do some damage control (this reminds me of Nvidia calling press before GTX480 launch). Actually many of the press guys I talked to are a little bit puzzled and don't really know how to approach this situation. From my point of view it is pretty clear, the truth (no matter how much it hurts) is the only way.

 

[RussianSensation]

The GTX 470 was only 5% faster than the HD 5850 and costed significantly more as well

Actually, GTX470 was about 10% faster on average, but only 8% slower than the 5870 (1920x1080 4AA/16AF). And GTX470s cost less than HD5870s. Coupled with more games, overclocking headroom to GTX480 (i.e., ~ HD5870 overclocked), it actually presented a great value vs. 5870s. 5850s were still untouchable in terms of value of course.

The point is Fermi consumed more power but was undeniably faster in stock or overclocked vs. overclocked states. Game bundles weren't a bad deal either . However, BD is shaping up to be slower, hotter, not much cheaper than 2500k and it doesn't look like it will overclock better either. That's a bigger disappointment than Fermi was, esp. when you consider it was 5 years in development and is at least 9 months late since SB launched. Ivy is also expected to launch in less than 6 months from now (unfortunately given lack of competition, Intel could easily delay it). More shockingly, a $225-250 8 core BD should never lose to Phenom II under any circumstances, not when X6 1090T overclocks to 4.0ghz and can be had for $160 or so.

A true next generation chip should be at least 15% faster simply based on improvements per core. Once you add more cores and higher clocks, it should be at least 30% faster if not more. Let's hope this design is going to scale much much higher in frequency clocks than Phenom II could. If BD scales to 5.5+ in the next 15 months, it might be able to compete with IVB after all.

 

[RussianSensation]

It's not a risk. I haven't paid for either board. I have them with me unopened. If BD sucks, I'll just give the 990FX back and pay for the Z68. If it matches SB, I'll keep the 990FX as I like it more.

Ya, with 3 days until launch, you might as well wait.

Curious as to why you like the 990FX platform more. Its SATA 3 performance with an SSD is a lot slower.

 

[Black96ws6]

What's really scary is this:

Worse IPC than Phenom, and BD is clocked HIGHER???

And Lab501 is a credible site.

Well I think that does it for me, TWO credible sites (Lab501 and the Dutch PC Magazine) with the same review: Bulldozer is a failure!

Again, by some miracle, perhaps things will change. I was planning on waiting until Black Friday to upgrade anyway to get the best deals.

But if these results are correct?

Wow. Just wow.

And how will the 4 and 6 cores perform??? That 4 core BD is going to get demolished in multi-threaded, if the 8 core version is the same as today's existing 1100T!

 

[StrangerGuy]

What's really scary is this:

Worse IPC than Phenom, and BD is clocked HIGHER???

And Lab501 is a credible site.

Well I think that does it for me, TWO credible sites (Lab501 and the Dutch PC Magazine) with the same review: Bulldozer is a failure!

Again, by some miracle, perhaps things will change. I was planning on waiting until Black Friday to upgrade anyway to get the best deals.

But if these results are correct?

Wow. Just wow.

And how will the 4 and 6 cores perform??? That 4 core BD is going to get demolished in multi-threaded, if the 8 core version is the same as today's existing 1100T!

You ever saw a BIOS update that improves stock CPU performance in the past 10 years? Me sure didn't.

 

[RussianSensation]

What's really scary is this:

Worse IPC than Phenom, and BD is clocked HIGHER???

SuperPi shows about a 48% advantage in favour of the 2600k. That lines up fairly accurately with 4C being able to keep up with an 8C CPU in the benchmarks presented by Lab501. Also SB's IPC was about 40% faster over Phenom II. So once again consistent. It looks like SuperPi will continue to be a good predictor of per core/IPC performance after all these years.

It's unfortunate that many people waited possibly 9 months only to be disappointed. I really think AMD is being too optimistic if they price FX-8150 at $250 or so. $225 or below seems more reasonable.

 

[mosox]

SuperPi shows about a 48% advantage in favour of the 2600k.

That's 93% not 48%.

 

[Imouto]

The maths skills again...

 

[Medu]

One thing I don't understand, let's say BD is as terrible as many of you are saying it is. Why would AMD release it? Wouldn't it make more sense to shrink Thuban to 32nm, maybe tweak the cache, NB, even the cores a bit and go forward with that?

They already have a 32nm Thuban core- it's called Llano and we all know how good that CPU part works...

OT:
Not hard to see why Meyer got fired...