Early Bulldozer Review

[jmac698]

http://translate.google.com/translate?sl=auto&tl=en&js=n&prev=_t&hl=en&ie=UTF-8&layout=2&eotf=1&u=http%3A%2F%2Fwww.ozeros.com%2F2011%2F10%2Freview-amd-bulldozer-fx-8150%2F&act=url

Unless you speak Italian, that is.
The rumours and previews were true; it's nothing particularly fast but a good option for the price point between 2500k and 2600k.

[Dark Shikari]

Why are they benchmarking a CPU with GPU benchmarks?

[jmac698]

Well we only had to wait a few more hours for some mainstream reviews.
http://hothardware.com/Reviews/AMD-FX8150-8Core-Processor-Review-Bulldozer-Has-Landed/
I'm sure you'll find the rest yourself.
Same conclusion.

[IgorC]

Many reviewers have made the common error when conclude that FX-8150 is slower than i5 2500k for video encoding. They watch at speed for first pass most of all.

It will be correct to calculate the total speed (fps) or time (sec) for two-pass encoding as (1/fps1)+(1/fps2) = (1/fps_total).

This way FX-8150 is on par with 2600K.

[Dark Shikari]

Anandtech seems to have declared a "new version of x264" to be an "AVX-enabled version of x264", assuming that the only difference between the two is AVX support, and commenting based on that. I'm guessing they didn't quite realize that their "HD benchmark" is using a version that's 2 years old, whereas the version AMD gave them was recent.

The results they claim are quite impossible because Bulldozer doesn't benefit at all from AVX: it already has built-in move elimination, making AVX unnecessary.*

*It still helps code cache efficiency by saving code size, but that's a tiny effect.

[hello_hello]

Many reviewers have made the common error when conclude that FX-8150 is slower than i5 2500k for video encoding. They watch at speed for first pass most of all.

It will be correct to calculate the total speed (fps) or time (sec) for two-pass encoding as (1/fps1)+(1/fps2) = (1/fps_total).

This way FX-8150 is on par with 2600K.

What about single pass, CRF encoding? How do they compare then?

[burfadel]

Apparently the OS handling of the threads on the Bulldozer is improper, and that a new update should resolve this later...

[Atak_Snajpera]

Windows 8 does not help at all.
http://www.purepc.pl/procesory/test_amd_fx8150_bulldozer_kontra_intel_sandy_bridge?page=0,12

[RNiK]

http://translate.google.com/translate?sl=auto&tl=en&js=n&prev=_t&hl=en&ie=UTF-8&layout=2&eotf=1&u=http%3A%2F%2Fwww.ozeros.com%2F2011%2F10%2Freview-amd-bulldozer-fx-8150%2F&act=url

Unless you speak Italian, that is.
OT: the site is in Spanish, not Italian. :p

[IgorC]

What about single pass, CRF encoding? How do they compare then?
Then You should watch at the result of second pass only.

[jmac698]

I was starting to think that Italian looked a lot like Spanish :) uffff :)

[Przemek_Sperling]

OT: the site is in Spanish, not Italian. :p

¡Ay, caramba!:eek:;)

[CruNcher]

Many reviewers have made the common error when conclude that FX-8150 is slower than i5 2500k for video encoding. They watch at speed for first pass most of all.

It will be correct to calculate the total speed (fps) or time (sec) for two-pass encoding as (1/fps1)+(1/fps2) = (1/fps_total).

This way FX-8150 is on par with 2600K.

Sorry Igor but with this data @ hand i would say it's awefull and no first pass here comes to the rescue (it's plain inefficient)
Sure its upto the 2600 but ehh did you looked @ the price you pay for that if Annandtechs measuring are correct ????

http://forum.doom9.org/showpost.php?p=1533509&postcount=1733

Though these measurings are somehow hard to belive looking @ a TDP of 125W and we have a overall output of 230W here
The Intel System has a TDP of 95W and a overall output of 160W (70W overall difference 0 Gain ???)

And if you look how Intel is going to improve this even further down in the next years (Tri Gate) (Haswell,Power Management) you can only Wonder where AMD is heading to currently (this was a major architectural change since K8) :(
And im pretyy sure @ that pace Intel will get faster to AMDs GPU Performance and (maybe even @ much better Efficiency) (they already doubling it with Ivy Bridge, starting early next year) then AMD to Intels CPU Performance and Efficiency.
So Fussion needs to evolve fast or Intel will be @ the front door knocking here soon too :(