While there have been many shiny new chips in the metaphorical sense, x86 (and x87) chips have never been known to be literally shiny. The typical packaging is ceramic or some form of brushed metal, and neither of these surfaces is shiny. But there was one x86 (or rather x87) chip which was actually very shiny, with a gold sheen to it:
33 MHz C&T J38700DX (1992)
It’s a Chips & Technologies SuperMathDX chip aka J38700DX, a 387-compatible math coprocessor. It’s a relative of the similarly obscure C&T Super386 CPU (which is not shiny at all).
Taking photos of this chip is a little challenging. Not only does it have an almost mirror-like polish but the lid is concave. Here’s a photo which captures the surface of the chip:
A very shiny J38700DX
At a different angle, the same chip looks black and flat:
A “black” J38700DX
On the other hand, the bottom side of the chip is utterly boring and bears no markings whatsoever.
Bottom side of J38700DX
For fairly obvious reasons, the FPU market was much easier to break into than the CPU market. The only 386 clones were Cyrix and C&T, and only the former was successful. But 387 FPUs were made by Cyrix, IIT, ULSI, and C&T, with the first three producing significant quantities of chips.
Were FPUs of this era ever heat-sinked?
Not to my knowledge. Up to and including 40 MHz, 386 Cpus had no heatsinks. Some but not all 33 MHz 486s had heatsinks, and starting around 50 MHz heatsinks were required. 387 FPUs were mostly ceramic, some had a metal head spreader (IIT). Intel never put heatsinks on their 386/387 chips.
The clock-doubled/tripled Cyrix and IBM 386/486 hybrids had heatsinks, but not the FPUs.
Nexgen 586 processors were pretty cool looking too! But heat sinks really killed that old style sexy chip packaging. It’s crazy how modern machines have massive sinks and fans, while before we ran bare.
I am fairly certain that my Alaris Blue Lightning board had a heat sink on the 387 variant used for the FPU. Tiny chip, poorly placed, putting out a good amount of heat, it probably needed it.
I have an Alaris Cougar board with a triple-clocked Blue Lightning soldered on, and the CPU does have a heatsink glued on. The FPU does not, but it was also dead when I got the board… so perhaps it should have had a heatsink too 🙂
Hmm, the Nx586 I have needs a heatsink + fan. After the gold-capped Pentium Pro and Cyrix 6×86, it got really boring.
>The only 386 clones were Cyrix and C&T, and only the former was successful.
I take it licenced 386 production (hello, AMD!) doesn’t count?
Right. Whatever the deal between Intel and AMD, as far as I know the Am386 is indistinguishable in software from an Intel i386. It has the exact same quirks, bugs, everything — but then again it’s supposed to use Intel’s microcode, so it should. The Cyrix and C&T 386s were developed from scratch, and are intentionally or unintentionally different.
Do you know if this C&T FPU needs the C&T CPU to work, or can the C&T Super Math work with a Cyrix CPU?
The ‘T9M12B’ code on the front of the chip, along with the ‘JAPAN’ line, leaves me to believe that Toshiba fabbed that particular Chips and Technologies SuperMath chip.
I’m 99% certain you’re right. I have seen pictures of other Toshiba-made chips which looked very similar.