Skip to main content Start reading the article

AMD FX-8130P tested at 3,2GHz

AMD FX-8130P Cache and memory test results and complete beat-down by Intel platform based on Core i7-920 processor in SuperPi test

While Zambezi debut is postponed by 2-3 months there is no problem finding numerous screenshots of engineering samples of upcoming AMD processors now. User nicknamed sunnykfc went even further and published a short video of AMD FX-8130P CPU running at 3200MHz with couple of tests followed.

Full CPU details are provided by AIDA64 CPUID system diagnosis tool:

AMD FX-8130P specs provided by AIDA64 software suite

The data was known long ago, so we will just refresh what we have by now: 8-core FX-8130P is equipped with 2MB of L2 cache for each pair of cores (overall of 8MB of L2 cache). The instruction set has changed a bit: among specified extensions you don’t see 3DNow! instruction typical for AMD processor architecture, however, several typical for Intel architecture extensions were added — SSE4.1, SSE4.2, 256-bit AVX instructions, AES and CLMUL encryption instructions. In addition, AVX 128-compatible instructions: AMD XOP, FMA4 and CVT16, were ported from AMD SSE5.

AIDA64 cache and memory test results of FX-8130P

Impressed by first results of FX-8130P we decided to compare it with obsolescent Core i7-920:

AIDA64 cache and memory test results of Core i7-920

In spite of FX-8130P having better memory timing numbers, good old Core i7-920 have bitten its rival, although, this could happen due to absence of proper support for new AMD processors in AIDA64 software suite.

Core performance was estimated by SuperPi test, that uses one processor thread:

FX-8130P results in SuperPi 1M

Here are Core i7-920 results for comparison:

Core i7-920 results in SuperPi

In this situation, you can’t write off FX-8130P full core’s (1 thread = 1 core) defeat, compared with virtual half (1 thread = 0.5 core ) of Core i7-920 processor, just for insufficient software support — even significantly higher clocked FX-8130P looses to Intel’s representative with almost double margin.

According to unconfirmed information gathered in Internet, AMD FX family currently includes six 8-core models, three 6-core processors and three quads:

Model NumberCore numberSteppingCacheHTTDPSocketRelease date
L2L3
FX-8170P8C08MB8MB3.2 GHz125WAM3+Q4 2011
FX-8150P8C08MB8MB3.2 GHz125WAM3+Q3 2011
FX-8130P8B28MB8MB3.2 GHz125WAM3+Q3 2011
FX-81208C08MB8MB3.2 GHz95WAM3+Q4 2011
FX-81108B28MB8MB3.2 GHz95WAM3+Q3 2011
FX-81008C08MB8MB3.2 GHz95WAM3+Q3 2011
FX-61206С06MB6MB3.2 GHz95WAM3+Q4 2011
FX-61106B26MB6MB3.2 GHz95WAM3+Q3, 2011
FX-61006C06MB6MB3.2 GHz95WAM3+Q3 2011
FX-41204C04MB4MB3.2 GHz95WAM3+Q4 2011
FX-41104B24MB4MB3.2 GHz95WAM3+Q3, 2011
FX-41004C04MB4MB3.2 GHz95WAM3+Q3 2011

Although FX-8130P is one of the top solutions in that table, tested Zambezi sample is based on B2 stepping, while newer C0 stepping can bring some specification improvements.

2 Comments

# Avatar Benz,

My Phenom II X4 955 @ 3.2 doesn’t need 27 seconds to pass 1MB of SuperPI so this is obviously fake…

I feel so sorry for some people that they need to lie and post fake benches cause they have to compensate for their own short comings… They don’t have a life they sit all day behind their damn computers and create fake benches…

Go find a chick… Get laid for crying out loud… Geez…

# Avatar Taron,

You’re right, man! My superclocked 1090T (4GHz) only needs 16 seconds to perform the SuperPi test loops.

I guess the Bulldozer might be slightly faster. Otherwise AMD would loose much reputation if their new baby isn’t faster than the old Thuban CPU architecture.

I guess, the screenshts have been very well photoshopped or the CPU has been scaled down afterwards to let AMD stand there like a fool against an I7-920…

maybe some Intel-Fanboy who fears the upcoming power of the new AM3+ CPUs

Leave a Reply to Taron

Avatar