• Please review our updated Terms and Rules here

Am5x86 reports itself as 100mhz?

Raven

Veteran Member
Joined
Mar 7, 2009
Messages
2,752
Location
DE, USA..
I have several Am5x86-133 chips. In every machine I've run them on recently (didn't measure in times prior to recently) they run at 100mhz according to CHKCPU. It says it's either an Am486 or an Am5x86 running in 3x mode. Am5x86s should be running in 4x mode when receiving a 2x multiplier from the board, no? So why is this? Do they report 100mhz to utilities like this even though they are running at 133, or do I really have a problem cutting into my CPU speed? I just got an Evergreen 586 (Am5x86-133 SMC) and installed it, and it's doing the same thing.

Anybody?

Edit: Here's my benchmarks so far..

BYTEcpu

am5x86-133adw @ 101mhz (in valuepoint)
integer: .538698
float: .336536

i486dx2-66 @ 66mhz (in vectra)
.404714
.233155

Evergreen 586 (Am5x86-133) @ 100mhz (in Presario 425)
integer: .540404
float: .335312

Speed600

POD83 @ 83mhz (in AVA)
int:680.9x
fpu:1398.x

i486DX2 @ 66mhz (in AVA)
int:316.60
fpu:487.22

Am5x86 @ 100mhz (in AVA)
int:474.92
fpu:730.87

Evergreen 586 (Am5x86-133) @ 100mhz (in Presario 425)
int:471.9x
fpu:660.20

Edit: Just about every other 486 chip can do it (for a while), but just so everybody knows, don't put a Cyrix 5x86 in a 5v system - it got wonky and if i left it it may have died.
 
Last edited:
Fixed the Evergreen! It has a jumper which some fool set to 3x instead of 4x. My question still stands, however, for non-upgrade variety Am5x86 chips and why they run at 3x in systems where they should run at 4.

Edit: New benchmark at proper speed:
Evergreen 586 (Am5x86-133) @ 133mhz (in Presario 425)
int:624.00
fpu:795.17

Seems the rumors of the POD being inferior are quite false, as it beats the Am5x86@133 even at integer by a bit.

Have you checked the wait state settings on the system board?

No? Wait state for what - be more specific please, and is it still relevant to me now that I've got it working?

Raven:

No real answer yet from me concerning your problem here, but check this url out when you get a chance - - - http://www.linuxjournal.com/article/2640

Yes I've seen that link, I enjoyed it and it was part of why I bought this CPU.
 
The Premio 486 mobo that I ordered off eBay was delivered today. That's the one where the fellow wanted $75 but took my offer of $40. Much to my suprise, it came with a 486DX-100 already installed (still with warranty tab attached.) It also has a small heat sink with a fan that's never been run. So, I think I'll forgo upgrading the CPU at this time. However, thanks for the offers.

P.S.
When I get the memory installed sometime this week and light it up, I'll post a narrative of the whole project.
 
Mind being more specific about that 100mhz chip? What manufacturer, what branding, is it an OverDrive, etc.?

If it's an Intel OverDrive it might be running the system bus at 25mhz * 4 to get 100mhz and you could SUBSTANTIALLY benefit by going to a 33mhz chip even if it was also at 100mhz (like my Am486-100s).
 
Raven: Here's what what it says . . .

Am486 DX4-100
A80486DX4-100SV88
C 9603DPD
3 VOLT
HEATSINK AND
FAN REQUIRED Designed for
Microsoft
Windows 95

BTW - Its a Socket 3
 
That chip is actually capable of running in a Socket 1 or 2 as well, by the way - I ran it in my Presario 425 until right when I started this thread, actually. Supposedly the chip is designed for 3.3v operation and can run at 5v "until it slowly dies" but I ran it for a month or so of heavy use with no trouble. Doesn't matter now though, I have a proper chip in there. You can get voltage converters for that btw.

Thanks for letting me know, I'm always on the lookout for interesting 486 chips I don't have, but I do already have that one. It's a good chip - 33mhz bus with 3x multiplier, so you'd fall into the "negligible performance boost" when going to a 4x 133mhz. An Am5x86 at 100mhz is exactly the same thing as an Am486-100 though, and my benchmarks (and experience) disagree with that notion and claim a pretty large increase - so if you find stuttering in Duke3d or the like, try a 133.

If your system supports 40mhz bus operation, though (mine does not), then you can benefit substantially by running at 160mhz 40x4 in that machine.
 
.....
An Am5x86 at 100mhz is exactly the same thing as an Am486-100 though, and my benchmarks (and experience) disagree with that notion and claim a pretty large increase - so if you find stuttering in Duke3d or the like, try a 133.
.....
Just for the record, his AMD DX4 has 8kb wb cache while the 5x86 has 16kb. Atleast if his part number is indeed SV8B instead of SV88 ;)
 
Ah, mine is part no. SV88 and I assumed that the B was a misread or typo because I wasn't aware there were two versions - in that case you'd also benefit when using the CPU cache, which is pretty often.
 
Ah, mine is part no. SV88 and I assumed that the B was a misread or typo because I wasn't aware there were two versions - in that case you'd also benefit when using the CPU cache, which is pretty often.
There are atleast 3 different versions (not taking into account 5V and 3.3V versions), V8T, V8B and V16B. There might be a V16T version also. T stands for write Through and B for write Back.
Obviously 8 and 16 mean 8kb and 16kb cache respectively.
iirc the V16B's were basically 5x86's scaled back from 133Mhz to 100Mhz. They may actually have a 4x multi but I never tried that on my V16B AMD DX4. Most I have are the V8T's and V8B's though.

The Intel DX4's also have 2 different versions, most are write through
 
There are atleast 3 different versions (not taking into account 5V and 3.3V versions), V8T, V8B and V16B. There might be a V16T version also. T stands for write Through and B for write Back.
Obviously 8 and 16 mean 8kb and 16kb cache respectively.
iirc the V16B's were basically 5x86's scaled back from 133Mhz to 100Mhz. They may actually have a 4x multi but I never tried that on my V16B AMD DX4. Most I have are the V8T's and V8B's though.

The Intel DX4's also have 2 different versions, most are write through

What's the SV88 that I have, then? Is it a 5v version? That would be nice.
 
What's the SV88 that I have, then? Is it a 5v version? That would be nice.

AMD_486_DX4_100_SV8B_Front.jpg
 
How on earth do these chips survive and operate fine at 5v? I ran my Am486DX4-100 in my Socket 1 box for weeks without the slightest hint of a problem until my Evergreen 586 arrived.
 
How on earth do these chips survive and operate fine at 5v? I ran my Am486DX4-100 in my Socket 1 box for weeks without the slightest hint of a problem until my Evergreen 586 arrived.

Maybe you just had a good part?
Those old chips weren't as critical about voltage compared to modern parts.
btw, did you order your Evergreen at the link I posted earlier?
 
No, I ordered mine from an auction for a pair of CPUs, one Cyrix 100mhz and one Evergreen 586 133mhz (heatsink variety). I wish I had known about these, because I spent $30 on the pair and the Cyrix isn't functional (it turns on, but gibberish and errors after post).
 
:pRaven:
Ran across these neat little files for checking the guts of your CPU's. You may already have this. One is for DOS and the other is for Windows.
 
Back
Top