AMD FX-8150 Bulldozer finally tested

[Kobaltrock]

So will Piledriver be 10% just because of a bug fix, or will it be 10% faster than a Bulldozer without a missfire?

Pile driver's 10% gain will most like be optimizations, but since those pics were more leaks, I have no idea how true that 10% figure is.

What is being talked about in the (3%) patch is the way BD accesses memory. It needs to do a extra fetch when the data isn't aligned correctly, and that can get expensive (time wise) since it must clear the cache, and start over again.

They are not related.

The thing we are all hoping for is for PD to tape out quickly, since from all the leaked BD benchmarks show, a 3% gain is pretty trivial.
We need competition in the high end, not the low end.

[SKYMTL]

I want to draw everyone's attention to the fact that the patch people are suddenly talking about is for Linux ONLY and it WILL NOT address Windows-based performance. In addition, the Linux community has determined the "patch" itself is pretty much a hack job that could negatively impact system stability and in-app performance. So, it should not even be discussed at length at this point in time.

[Hondacity]

thanks Mr Skymtl

how many nailbiting hours before nda lifts? are we following eastern time or when oct12 hits one country?

[nyeah]

thanks Mr Skymtl

how many nailbiting hours before nda lifts? are we following eastern time or when oct12 hits one country?

yea, can't wait till NDA lifted, too many speculation based on wild guessing

[Gilgamesh]

I want to draw everyone's attention to the fact that the patch people are suddenly talking about is for Linux ONLY and it WILL NOT address Windows-based performance. In addition, the Linux community has determined the "patch" itself is pretty much a hack job that could negatively impact system stability and in-app performance. So, it should not even be discussed at length at this point in time.

Just a little question: for Bulldozer you will adopt same metodolgy and programs of your last review here?
hardwarecanucks.com/forum/hardware-canucks-reviews/44339-intel-core-i3-2120-core-i5-2400-lga1155-processors-review-4.html

I like it.

[solofly]

Two more days to Halloween, silver shamrock...

[Kuroimaho]

I want to draw everyone's attention to the fact that the patch people are suddenly talking about is for Linux ONLY and it WILL NOT address Windows-based performance. In addition, the Linux community has determined the "patch" itself is pretty much a hack job that could negatively impact system stability and in-app performance. So, it should not even be discussed at length at this point in time.

You MR are the reason in this madness.

[AzureSky]

I want to draw everyone's attention to the fact that the patch people are suddenly talking about is for Linux ONLY and it WILL NOT address Windows-based performance. In addition, the Linux community has determined the "patch" itself is pretty much a hack job that could negatively impact system stability and in-app performance. So, it should not even be discussed at length at this point in time.

as I understand it, theres a patch for windows either incoming or already in the hands of some testers.

and I for one couldnt give less of a about what the Linux community "determined", I quit with desktop linux years back after I finnely realised that it wasnt going anywhere, they refuse to support a unified audio/sound fraimwork, linux lacks a stable driver abi/api and ties drivers closely with the kernal so a small kernal patch can break your drivers....I could go on and on, but no need, I gave up on linux as a desktop OS quite a while back and have been happier since, as I prefer to use my computer rather then having to recompile the kernal with patches added/removed for hours/days on end to find what broke my video/audio/nic drivers.
oh yeah, and on windows/osx/bsd no dependency hell to deal with, no needing 3 versions of gtk installed because some idiot app dev decided to tie his/her app to a specific gtk version and hasnt bothered to updated to support newer gtk....

bah!!!


I have to say a few things about other comments about marketing and how "normal people" will see 8 cores being slower then 4 cores as fail....

Normal "Joe Sixpack" type users have no bloody idea about how a chip preforms, they buy what appears to be the best value, having worked in PC sales over the years, most of them dont even really care if its intel/amd/via/wtfe they only care that it looks like a good deal, hence the p4 sold so well despite it being so HOT and SLOW.

the fact is, most users are FAR FAR different from the users of this and other tech forums, they got no idea whats really in their computer, they just know what the sales people told them.

example: I recently had a friend of the family ask me to look at his computer, he bought it a few months ago and was told it was "Really fast" I get over there and its low end i3 prebuild with a 1gb gf210 and 2gb of VERY slow ram, he thinks its fast, because thats what he was told, well i cleaned off the preinstalled crapware and tweaked what i could to get it to run better, told him he needed to order ram and a videocard minimum if he wanted to play WoW on the 27" screen they sold him (1080p), He actually got mad at me for telling him this because they guys at the store told him it would game great and was fast...

i got him to come over to my house and try my system, he was quite mad when he found my system to be like lightning to his systems turtles stampeding thru peanut butter in January....

note: this office system im on smokes his, and its just a 9650@2.7 and 8gb ddr2 with a 8800gts512......

so he ordered the ram and gpu (he got a great deal on a 5770, under 80bucks on sale), I installed it for him, he was shocked how much better windows ran, let alone wow, now hes talking about cpu upgrades....on a system thats less then a year old that some smuck at bestbuy sold him to play wow on...ROFL......(think im gonna tell him to just get a board/cpu combo and a psu, the case is nice, just really bad liteon psu and horrible motherboard...)

meh, long story short, most users are idiots you could give a system running windows 8 on arm and as long as they could run ms office and IE they would be happy and not know the dif between that and their x86 system....


as to this dozer situation, I will wait and see how things flush out, i mean its not that big a deal to wait a bit and see if the windows patch shows up and boosts perf and to see how BD actually preforms.

I have little to no use for cinbench, its compiled with ICC using commands that make it use the intel cpu dispatcher, meaning AMD chips automatically get slower code paths then intels.

[Pantsu]

how many nailbiting hours before nda lifts? are we following eastern time or when oct12 hits one country?

It's noon EST.

I like how people take it as black or white depending on what one preview test result says and how it's suddenly what justifies the price tag. There's no doubt logic behind AMD's pricing. FX *will* perform in some cases at least between 2500K and 2600K. More than likely it's the 8 thread applications that take full advantage of the architecture. This and the fact that it's the brand new flagship is enough to justify the price at launch at least. FX-8120 might be much more reasonable in terms of price/performance for the avarage gamer/overclocker (compare for example 955BE and 980BE). The question is how AMD's single thread perfomance works (or doesn't) with module architecture and slightly upped clocks. If lab501 results are true, it looks like the extra clocks are nowhere high enough to compensate the loss in IPC. Either they missed the clocks due to power/temp/whanot issues and/or the architecture just takes a much bigger hit than what they expected compared to Stars.

[Dumo]

when we were playing BD a week ago we were looking at each other like "omg this is it?" - 6500+ and the scores were still as ugly as dinos in his summer-suit -.-
Everybody went back to Sandy and GT systems like half an hour later

Lol....Dang, if 6.5Ghz 8C/8T Vantage and '06 runs suck then it will be very sad

[informal]

I don't know what,but imo something just doesn't add up.

[Tha Last Meal]

guys i have the First picture of the amd Bulldozer
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

[radaja]

I don't know what,but imo something just doesn't add up.

what you talking about willis?

[vinta]

guys i have the First picture of the amd Bulldozer
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

[Kuroimaho]

guys i have the First picture of the amd Bulldozer

But it plays smooth jazz so well, just wait and see everybody who will own it will praise it for the smoothness.

[radaja]

is cpu-z just not updated correctly(maybe franck) forgot something?
codename still reads bulldozer,isnt that wrong?shouldnt it be zambezi?

[Oliverda]

the queston was answered, amd said ~3%

It's more than 3%.

[FischOderAal]

It's more than 3%.

AMD themselves said up to 3 percent...
http://permalink.gmane.org/gmane.linux.kernel/1172226

[Oliverda]

AMD themselves said up to 3 percent...
http://permalink.gmane.org/gmane.linux.kernel/1172226

I'm talking about Windows.


mod: Anyway you can try the patch easily becuase Win 8 already contains it.

[saint19]

is cpu-z just not updated correctly(maybe franck) forgot something?
codename still reads bulldozer,isnt that wrong?shouldnt it be zambezi?

Maybe, if you look some current cpuz reports all the 9xx mobos are reported with 850sb

Sent from my Android using TapaTalk

[SKYMTL]

Just a little question: for Bulldozer you will adopt same metodolgy and programs of your last review here?
hardwarecanucks.com/forum/hardware-canucks-reviews/44339-intel-core-i3-2120-core-i5-2400-lga1155-processors-review-4.html

I like it.

It has been substantially expanded actually.

as I understand it, theres a patch for windows either incoming or already in the hands of some testers.

If the "patch" would work without breaking a million other things, you can bet that AMD would be actively pushing it to reviewers. They aren't, which should say a LOT about its viability and its compatibility with Bulldozer rather than Piledriver...

People seem to forget that Boris posted these patches for Thurban as well: https://lkml.org/lkml/2010/3/22/336 Did any of these ever amount to anything substantial on the Windows platform? No and no.

[Oliverda]

...

[Hornet331]

I'm talking about Windows.


mod: Anyway you can try the patch easily becuase Win 8 already contains it.

We know nothing about windows (7) and you are talking about the rewritten task scheduler (which of course includes the patch but is also a lot different from the one in win7), but windows 7 wont get that scheduler.

[Oliverda]

We know nothing about windows (7) and you are talking about the rewritten task scheduler (which of course includes the patch but is also a lot different from the one in win7), but windows 7 wont get that scheduler.

I hope that you didn't include me in your "we".

[Hornet331]

So tell me do you know the performance increase the patch will give you on win7 vs an unpatched version?

And pls no NDA BS since a percentage from a unknown number is still unknown.