Not sure what you're trying to say. With AMD, 2x multiplier is used / interpreted as 4x.

Except it was detected as 5x, and the processor was then detected as a 5x86 by CHKCPU
You should run a few benchmarks to see if this is for real or a misdetection.

even the BIOS recognized it as X5 P75 and Linux Kernel 2.6.22 too.
NSSI CPU Benchmark is between P75 and P100.

Can you recommend CPU only Benchmark ?

i wanted to run run 2x50Mhz on DX4, instead i got this.

SpeedSys in DOS?

Some AMD 486 DX4-100 CPUs supports 2x and 3X multipliers, some supports 3x and 4x multipliers.

If you have 16KB L1 cache then your CPU is the same as a 5x86p75, if you have 8KB then its not the same even if its WB cache.

got 50.09 in Speedsys for CPU.

My 5x86 behaved in the same manner.
When I chose a multiplier of 3x, it recognised it as an AMD 486DX4-100 WB.
With a multiplier of 2x (or 4x), it recognised it as an AMD 5x86 P75.
Generally speaking, the 5x86 is nothing more than a 486 DX4-100 WB with an internal multiplier of 4x.
The 5x86 can also be configured for write through (WT) mode.

NSSI reports 8Kb of L1 Cache, Sysspeed reports 16Kb.

Who is right ? 😀

Now running DX4 @ 160Mhz 😳

EDIT:
cachechk reports 16Kb Cache.

Is there a Method to check real cache size ?

The test in CTCM is quite reliable.
However the graph in Speedsys should show it nicely as well.

matze79 wrote:

Can you recommend CPU only Benchmark ?

Speed Test
run as: speedtst >res.txt

also you should check system's stability - CPU and video. for example Quake demos running in SVGA for a couple of hours will be not bad. linx comes to mind too

Your DX4 runs as 5x86, because it actually IS a 5x86 chip inside. Compare the package number on your chip with the information given here: http://www.cpu-world.com/forum/viewtopic.php? … er=asc&start=15

Yes, 5x86 and 'P75' are just marketing, in response to Intel's Pentium. Very similar to the PR rating that AMD put on their Athlon CPUs in the Pentium 4 era.
The CPUs were actually just 486-derivatives (which is why they perform so poorly in a game like Quake, which expects a true Pentium with a properly pipelined FPU instead of the outdated 487).
See also http://www.cpu-collection.de/?l0=co&l1=AMD&l2=5x86:

Introduced in November 1995, the AMD 5x86 is a standard 486 processor with an internally-set multiplier of 4, allowing it to run at 133 MHz on systems without official support for clock-multiplied DX2 or DX4 486 processors.

So software can't really tell the difference between a 486 and a 5x86, because they're the same chip. The software probably 'guesstimates' that it's a 5x86 because of the clockspeed and multiplier used.

tried some Tests, looks like the Core is really the same like the 5x86, but amd disabled half of 16kb.
but makes no big difference at all.

Core works at 3x50 and 4x40Mhz well, no issues, crashes etc.
i did some mpeg encoding on the CPU under Linux and it was running ~ 30-40 hours for a videocd.

matze79 wrote:

Core works at 3x50 and 4x40Mhz well, no issues, crashes etc.

try linx. 2 hours

matze79 wrote:

tried some Tests, looks like the Core is really the same like the 5x86, but amd disabled half of 16kb. but makes no big differen […]
Show full quote

tried some Tests, looks like the Core is really the same like the 5x86, but amd disabled half of 16kb.
but makes no big difference at all.

Core works at 3x50 and 4x40Mhz well, no issues, crashes etc.
i did some mpeg encoding on the CPU under Linux and it was running ~ 30-40 hours for a videocd.

Video encoding on a 486 😲

Man you have patience.

16 years ago, our 486DX4 100Mhz had hard time to process DivX Files to VCD Format.
Using TMPEGEnc on Windows 95 and Nero Burning ROM.
It never crashed, it had a 2x CD-Burner and a 56k Modem to download movies.

http://www.p-wholesale.com/upimg/2/192a1/vcd- … d-b2000-408.jpg

In Germany VCD wasnt very common so you had to burn them yourself 😀