I'm not sure at all, but won't the coprocessor just run at the same frequency as the cpu ? How could 33MHz be generated with only a 25MHz quartz ?

Deksor wrote:

I'm not sure at all, but won't the coprocessor just run at the same frequency as the cpu ? How could 33MHz be generated with only a 25MHz quartz ?

Indeed, that's exactly what I'm thinking but the Ebay seller insists it should work no problem 😒

You got a 33MHz part. I should work with any 386DX slower than 40MHz. The seller did you a favour.

Anonymous Coward wrote:

You got a 33MHz part. I should work with any 386DX slower than 40MHz. The seller did you a favour.

So wait a minute, I was into thinking the Math coprocessor had to match the CPU clock speed in order to work in this case with the 386SX 25Mhz and what you are saying is that I got a free upgrade?

Thanks for the help!

Megadisk wrote:

Anonymous Coward wrote:

You got a 33MHz part. I should work with any 386DX slower than 40MHz. The seller did you a favour.

So wait a minute, I was into thinking the Math coprocessor had to match the CPU clock speed in order to work in this case with the 386SX 25Mhz and what you are saying is that I got a free upgrade?

Thanks for the help!

Free upgrade? No. You got a better part, but it will still run @ 25MHz in your motherboard.

A 33MHz coprocessor will run underclocked at 25MHz with no problem.

The frequency at which your computer and it's components in determined by a separate crystal, there is not one integrated into the CPU or FPU, and unlike modern CPUs there is no hardware governing the operating speed.

Fair enough guys. I'll install it and go from there thanks much for the help!

Is this the correct utility for the chip? : http://www.driverguide.com/driver/detail.php? … 93751a59e2ef27b

Megadisk wrote:

Fair enough guys. I'll install it and go from there thanks much for the help!

Is this the correct utility for the chip? : http://www.driverguide.com/driver/detail.php? … 93751a59e2ef27b

I wasn't even aware math coprocessors needed drivers. My integrated DX co-processor on my 486 doesn't, it's just picked up like memory or an ISA card.

It's not a driver. It's probably just a diagnostic to determine if it's functioning properly.

BIOS should report the FPU in some manner. You can try a few benching tools from Phil's dos benching kit, I think the one that uses AT 286/287 scores should work for that, forgot the name tho.

A quick update. So I also got hold of an Intel 387sx 25Mhz coprocessor and the testing begun.

The i387sx worked great with the intel advanced diagnostic utility to test the chip. Then I installed the Cyrix CX83S87-33-JP with its utility and that's when the problem started. It kept giving me a "Transcendental Test" error. What the heck does that mean?

Then I ran intel's utility with the CX83S87-33-JP since it should compatible and still gave me the same exact error, what da?..

1) try the faulty Cyrix FPU in another motherboard
2) try a different Cyrix FPU in the affected motherboard

The CX-83D87 was introduced in 1989. It is the fastest 387-compatible coprocessor and provides up to 50% more performance than the Intel 387DX. The 83D87 also offers the most accurate transcendental functions of all coprocessors. It is the 387 clone with the highest degree of compatibility to the Intel 387DX. Unlike the Intel 387DX, the 83D87 (and all other 387-compatible chips as well) does not support asynchronous operation of CPU and coprocessor. To reduce power consumption the 83D87 features advanced power saving features. Those portions of the coprocessor that are not needed are automatically shut down. If no coprocessor instructions are being executed, all parts except the bus interface unit are shut down.

So, highest degree doesn't mean it runs 100% like the intel. Maybe some incompatibility? Speed, motherboard?

It was noted that the newer Intel 387 copro chips (units with "16-33" written on the face) are almost as fast as the Cyrix ones. I will verify this soon. I am working on getting a stable 386-50 running with a 50 MHz FSB, but need an async FPU at 40 MHz.

sprcorreia wrote:

The CX-83D87 was introduced in 1989. It is the fastest 387-compatible coprocessor and provides up to 50% more performance than the Intel 387DX. The 83D87 also offers the most accurate transcendental functions of all coprocessors. It is the 387 clone with the highest degree of compatibility to the Intel 387DX. Unlike the Intel 387DX, the 83D87 (and all other 387-compatible chips as well) does not support asynchronous operation of CPU and coprocessor. To reduce power consumption the 83D87 features advanced power saving features. Those portions of the coprocessor that are not needed are automatically shut down. If no coprocessor instructions are being executed, all parts except the bus interface unit are shut down.

So, highest degree doesn't mean it runs 100% like the intel. Maybe some incompatibility? Speed, motherboard?

Regardless of compatibility, the Cyrix FPU should test good with the Cyrix test suite which was made specifically to test Cyrix coprocessors. I agree that the chip should be retested in another motherboard though.

I don't see what the problem could be other than a bad CX83S87-33-JP chip because the Intel N80387SX(16-25Mhz) works just fine on it.

So much of that so called "better" Cyrix compatibility I guess unless the 33Mhz against the board 25Mhz does makes a differences after all which is unlikely since accoding to a post above, it runs on a different clock signal 😵

Ensure that your Cyrix FPU is fully functional in another motherboard before assuming it has compatibility problems. Some aspect of the FPU may be defective.

I apologize I am resurrecting this very old thread, but where that "Cyrix Math Coprocessor Demonstration" shown here:

Re: Math Coprocessor speed question....

can be downloaded? I mean the Intel one I found in seconds, but the Cyrix one - I cannot find. thanks!

[EDIT] ok, it took a lot of searching, but i finally found it - attaching it and so we have it archived here.