Bookmark and Share


An overclocker from the Czech Republic has managed to obtain an engineering sample of AMD FX-8130P microprocessor code-named Zambezi that is powered by Bulldozer micro-architecture and overclock it to 4.6GHz.

According to overclockers blog OBR, an engineering sample of eight-core AMD FX-8130P processor could function at 4635.6MHz clock-speed while being cooled down using an air cooling solution. At such high speed, the processor could run SuperPi application and calculate π digit with 1M precision in over ten seconds. The processor worked with 1.5V core voltage.

Exact specifications of the AMD FX-8130P central processing unit (CPU) are unknown, no stepping or revision are known either.  What is know is that it will operate at over 3.0GHz with all eight cores active. Besides, it is interesting to note that the CPU-Z application indicated 186W thermal design power for the chip.

AMD itself delayed release of its highly-anticipated Bulldozer-based FX-series microprocessors for desktops from June to September. Unofficial sources claimed that the company's current chips did not demonstrate competitive performance at their clock-speeds and that AMD needed new revisions to get the required frequencies and performance levels. Apparently, there are Zambezi chips with high overclocking potential that can run all eight cores are 4.60GHz.

AMD did not comment on the news-story.

Tags: AMD, Zambezi, Bulldozer, 32nm


Comments currently: 16
Discussion started: 06/27/11 11:15:21 PM
Latest comment: 06/30/11 04:33:44 AM
Expand all threads | Collapse all threads



AMD itself delayed release of its highly-anticipated Bulldozer-based FX-series microprocessors for desktops from June to September.

Do math -> July 31st - August 31st = 61-91 days from June 1st

Unofficial sources claimed that the company's current chips did not demonstrate competitive performance at their clock-speeds and that AMD needed new revisions to get the required frequencies and performance levels.

B1 suffered something and AMD is fixing that something and they should fix it by B3 atleast

Other Point:
Well of course...since a Bulldozer is a low clock, balanced IPC/FPPC machine

I expect 2.6-3.2GHz

The change of idealogy for me is that I looked REALLY REALLY closely at the Bulldozer Architecture and I moved the paper side ways and saw what I was pos to see

To be more clear I predict we will see something like

FX-57 @ 2.6GHz vs Pentium 4 EE @ 3.73GHz



FX 8C vs i7 16T
0 2 [Posted by: seronx  | Date: 06/27/11 11:15:21 PM]
- collapse thread

you got it the wrong way dude. Bulldozer has been optimized for the knee of the IPC curve, but is almost certainly not possible for BD to equal Sandy Bridge at single threaded performance, at least at the same clock speed.

So this time, tables have turned. Now it is AMD releasing the high-clocking capability on its architecture, and now the question is how much more clock speed BD is gonna need to match Sandy Bridge in single-thread.

Multi thread is another story, i have no doubts whatsoever that top model Zambezi will crush I7 2600
0 0 [Posted by: CarlosTex  | Date: 06/28/11 12:39:19 AM]
I don't see that

I see an 8 Core(4 Modules) vs an 8 Core(with SMT)

Well balanced design vs Under-provisioned design

Low clock vs High Clock

When two execution cores is used you get really fast processing, low clock(2.6GHz(.5GHz TC) At least 4IPC

When one execution core is used you get really slow processing, high clock(2.6GHz(1GHz TC)) At least 2IPC

Module has 4 piplelines
E.Core has 2 pipelines
Floating Point takes 2 pipelines

So you can be running something dual-core have 1 integer and 1 floating point per core or have something use 1 core and switch between 2 floating point and 2 integer per clock
1 0 [Posted by: seronx  | Date: 06/28/11 04:38:33 PM]
Wrong dude. Read the article about Bulldozer by David Kanter. First of all you'll never reach a 4 IPC. Even Sandy Bridge won't reach 3.

Second don't mix pipelines with execution units.
A Bulldozer module has 8 pipelines, 4 per integer core, but they won't work together for additional single thread performance. Forget about that, it has been stated several times by John Fruehe.

Apparently BD has a 17 stage pipeline, with incresed lantencies, but the front end has been massively improved, so BD should be faster per clock than family10h, even though it has theoretically lower IPC potential than the older architecture. BTW even though family10h had 3 AGU's and 3 ALU's, it worked more has a 1.5 ALU and 1.5 AGU, so BD has 33% more execution resources.

It doesn't matter, nowadays IPC is not a good measure of processor performance.
1 0 [Posted by: CarlosTex  | Date: 06/29/11 01:45:43 AM]
I know that is server based

but understand that a
Server K10h @ 2.5GHz = Server K15h @ 1.8GHz
while being 33% faster with 33% less cores
0 0 [Posted by: seronx  | Date: 06/29/11 05:38:14 PM]
??? The design is the same, for both Desktop and Server, and it is too early to say which family15h chip will equal a family10h chip regarding performance. It is a fact that BD has 33% more execution resources, but that number won't translate in performance difference.
0 0 [Posted by: CarlosTex  | Date: 06/30/11 12:14:34 AM]
K10h @ 2.5GHz = Server K15h @ 1.8GHz


that is a 72% increase in performance not 33%
0 0 [Posted by: seronx  | Date: 06/30/11 04:33:44 AM]
I sure hope this is fake, otherwise, the IPC rating of bullzoder is down in the dumps. At these clockspeeds, the sandy bridges are somewhere in the low 8 seconds, and this is surely above 10s, maybe even 19s.
0 0 [Posted by: psycho_mccrazy  | Date: 06/28/11 06:40:09 AM]
SuperPI results are irrelevant. You don't rely on x87 code nowadays.
1 0 [Posted by: CarlosTex  | Date: 06/28/11 10:09:33 AM]
Correct, x87 is now completely emulated

Netburust and Core i7 from Intel, emulates x87

K15 from AMD, emulates x87
0 0 [Posted by: seronx  | Date: 06/28/11 07:02:49 PM]

What stepping is this? Assuming that Bulldozer clocks at 3.8GHz out of the box, that's a modest overclock. I think some people may have been hoping for more.
0 1 [Posted by: GavinT  | Date: 06/28/11 02:47:45 AM]
- collapse thread

B0, and expect low GHz

8130P has a 3.8GHz TC for half the cores being utilized(8130P = 2.8GHz(Stock)/3.3GHz(TC 8 Cores)/3.8GHz(TC 4 Cores)
0 0 [Posted by: seronx  | Date: 06/28/11 04:36:23 PM]

That image has a "I'm so photoshopped" aura to it, that I find hard to believe this article XD

0 0 [Posted by: YukaKun  | Date: 06/28/11 08:37:49 AM]

Word on other sites is this is fake, especially because the checksum isn't shown for SuperPI and the time has been blanked:

"The time is for pi to 1M and the time taken is 1X.Y29 seconds (where 0<X<9 and 0<Y<9), meaning it is between 10.029 and 19.929 seconds... which is a massive range, and unless it's closer to 10 isn't brilliant for 4.6GHz..."
0 0 [Posted by: AnonymousGuy  | Date: 06/28/11 02:23:56 PM]
- collapse thread

So? SuperPI is x87 code. It has been deprecated. SuperPi is not a reliable tool for measuring processor performance. Whatever performance BD gets in SuperPI it is irrelevant.
1 0 [Posted by: CarlosTex  | Date: 06/29/11 01:48:24 AM]
I am agree with you,better tool for measure CPU compute power is Linpack 64 (use fortran routine).
1 2 [Posted by: Blackcode  | Date: 06/29/11 04:32:36 AM]


Add your Comment

Related news

Latest News

Wednesday, November 5, 2014

10:48 pm | LG’s Unique Ultra-Wide Curved 34” Display Finally Hits the Market. LG 34UC97 Available in the U.S. and the U.K.

Wednesday, October 8, 2014

12:52 pm | Lisa Su Appointed as New CEO of Advanced Micro Devices. Rory Read Steps Down, Lisa Su Becomes New CEO of AMD

Thursday, August 28, 2014

4:22 am | AMD Has No Plans to Reconsider Recommended Prices of Radeon R9 Graphics Cards. AMD Will Not Lower Recommended Prices of Radeon R9 Graphics Solutions

Wednesday, August 27, 2014

1:09 pm | Samsung Begins to Produce 2.13GHz 64GB DDR4 Memory Modules. Samsung Uses TSV DRAMs for 64GB DDR4 RDIMMs

Tuesday, August 26, 2014

10:41 am | AMD Quietly Reveals Third Iteration of GCN Architecture with Tonga GPU. AMD Unleashes Radeon R9 285 Graphics Cards, Tonga GPU, GCN 1.2 Architecture