2 days ago, pcinpact.com (a French IT news site) release a preliminary benchmarks under Linux using PTS:
It only compare FX8150 vs X6 1100T.
Phoronix: First Linux Benchmarks Of AMD FX-8150 Bulldozer
Here's the first Linux benchmarks of AMD's FX-Series Bulldozer desktop CPUs that launched on Tuesday. Specifically, it's Gentoo Linux performance results for an AMD FX-8150 Bulldozer...
in my point of view the 128bit ram channel driven 8 core bulldozer isn't the reverence for amd.
its the socket g34 based 256bit ram channel driven opteron 6200 with 8 cores.
because in the same 8 core you do have much more speed with less MHZ and less power consuming.
bulldozer appears to be a major failure.
it only wins in tasks that can be heavily multithreaded, and it's not very power effective.
FX processors based on Bulldozer microarchitecture managed to show their strengths only in a small variety of common user tasks. There are very few popular applications, which would generate simple multi-threaded integer load and this is the only case when Bulldozer really performs at its best. As a result, in certain applications the new Bulldozer is not just slower than competitors from Intel, but is even slower than the previous-generation Phenom II X6. And it means that AMD didn’t succeed in launching a revolutionary desktop CPU.
In fact, FX is just another Phenom, which looks pretty good especially compared with the predecessors. Overall, FX processors are faster than Phenom II, they overclock much better and consume slightly less power, so they will be a good replacement for the CPUs on old K10 microarchitecture.
I guess they just care about professional reviews, instead of somebody who just throws out one "news"-entry after another without any content, just to drive add revenue.Evidently they don't care too much about Linux coverage for this less-than-stellar product launch
I am glad you know how to swear in german.Scheiße!
An analysis of why the bulldozer is the best cpu amd could build:
The question is why they have doubled the integer units to 8cores and the floadingpoint and (SSE) SIMD units are only 4core the answer is simple:
Intel always use the "the instruction set war" against AMD (SSE) SIMD the intel's command set always used by intel cpus first and 2-3 years later the competitor can use it at the first time.
the bulldozer is in FMA4 and VXA incompatibility never ever gotten a code snippet from the intel compiler because the amd FMA4 and VXA are not bit compatible and because why? intel force them to be not compatible.
Intel always make it so the last 20-30 years so a 3-4 year old technology for the competitors in real intel only will give max SSE4.2 to amds bulldozer
so why amd is now worth as much to integer units?
is simple because the instruction set of the integer units since the 486s and the 64 bit extension more or less the same without chances.
amd can deliver the integer performance of Intel compilers to the customer against the will of Intel.
Now who says the bulldozer was a bad CPU simply has no plan of the market situation and policies of the Intel instruction set
intel play this game on sse2 vs sse3 at gemcht and sse3 vs SSE4.1 and SSE4.2 vs sse3 this means amd always lose on SIMD.
only in integer performance, they can not
because it is "real" computing power without any fake cheat logic
if you look at the benchmark 7zip on bulldozer has the cpu is clever and genius !
a very good source about this the instruction set war: http://www.agner.org/optimize/blog/read.php?i=25
intel won always on the "the instruction set war" AMD only chance is the "integer" performance to compete with Intel.
Last edited by Qaridarium; 10-14-2011 at 10:09 PM.
I will probably never get how to use the openbenchmarking website.
So I see those test results:
But how can I view those results side-by-side with other CPUs and configurations?
Also, gentoo seems to still be on gcc 4.5. I'd like to see another comparison with gcc 4.6 (with "-Ofast -march=native" on both configurations!) as well as maybe Open64 (and for the sake of comparing: path64).
This patch for bulldozer in the kernel, that linus was upset about in the beginning: is it in 3.0? 3.1?
As for the mean things towards Phoronix I really don't get them. I like the news here and yes the reviews I have seem could have been a bit better but this probably is just one of the few review sites for Linux. The only one that does Linux reviews in English that I know of in fact. So yeah AMD could have sent a Bulldozer. Anyway it's not like I care. I'm still going to support AMD because left alone Intel mean big trouble, already their prices are way off the scale, if they get to dominate even more ... it won't be good.
I don't feel Bulldozer is worth buying right now. One of the old Phenoms yes, they have very good prices and perform decently.