I don't think so. Socket 4 was the first Pentium socket and socket 5 and 7 were for later Pentiums and AMD and others.

chrismeyer6 wrote on 2020-12-03, 01:51:

I don't think so. Socket 4 was the first Pentium socket and socket 5 and 7 were for later Pentiums and AMD and others.

Well Socket 6 was a legacy 486 Socket right? That's what I had thought when I saw it a few months ago on some list of Sockets, and never was really interested until an hour or so ago.

It seems like they released at the very tail end of the 486 life cycle and I can't find hardly anything on it. I'm now interested in hopefully finding out more about it.

Me too Chrismeyer! Yes, it definitely came out after Socket 5.

I'm coming up basically empty when searching for it.

Presumably it was only used on Intel reference boards.

Apparently they dropped legacy pin for old Overdrive CPUs.

I believe it exists in Japan. I think I saw a picture of one in a PC98 system. That's the only time I've ever seen one.
It's not like they're anything special. They just don't accept 5V CPUs.

Anonymous Coward wrote on 2020-12-03, 03:10:

I believe it exists in Japan. I think I saw a picture of one in a PC98 system. That's the online time I've ever seen one.
It's not like they're anything special. They just don't accept 5V CPUs.

Interesting. Does anyone know if there are any AT motherboards with them?

I think Socket 6 was supposed to move the 486 onto the ATX motherboard standard but since the 486 and AT were both on the way out, manufacturers and consumers both just ignored it.

Edit: I did a little more digging and came across a bit more info in the 1996 edition of "Upgrading and repairing PCs" on archive.org.

The last 486 socket was created especially for the DX4 and the DX4 (Pentium) OverDrive Processor. Socket 6 is a slightly redesigned version of Socket 3, which has an additional two pins plugged for proper chip keying. Socket 6 has 235 pins and will accept only 3.3v 486 or OverDrive processors. Currently, this means that Socket 6 will accept only the DX4 and the DX4 (Pentium) OverDrive Processor. Because this socket provides only 3.3v, and because the only processors that plug into it are designed to operate on 3.3v, no chance exists that potentially damaging problems will occur, like those with the Socket 3 design. Most new 486-class systems that come with DX4 chips initially will use Socket 6. Figure 6.7 shows the Socket 6 pinout.

The DX4 100MHz processor can draw a maximum 1.45 amps of 3.3v (4.785 watts). The DX4 (Pentium) OverDrive Processor that eventually will replace that processor will draw a maximum 3.0 amps at 3.3v (9.9 watts) and 0.2 amp at 5v (q watt) to run the fan, for a total 10.9 watts. Like all the other OverDrive processors that have active heat sink/fan assemblies, the processor has an easy-to-remove- fan that can be replaced should it ever fail.

One strange thing with the socket 6 is that it still has the extra row of pins despite not being 5V, only 3.3V even though pentium overdrives are 5V only.

On the other hand, I heard that pentium overdrive for socket 3 was intended to be named "pentium SX".

Based on these two statements (the second needs a confirmation), my theory is that perhaps intel planned to reproduce the same schema they made with 386SX/386DX ?

Maybe they wanted to make "pentium DX" for socket 4/5/7 with a 64 bit bus and a "pentium SX" for socket 6 just like they made 386SX for kinda the same plaform the 286 was using and meanwhile the real improvements would have been made on the high-end platform just like 386DXs were ?

I think they inially planned to milk the 486 platform for longer, but quicky abandoned it in favour of Socket 5 / Pentium.

Just like later Super Socket 7 which came later without Intel's involvement as they moved to Slot 1.

As far as I know there are no AT S6 implementations. It appears in some proprietary/industrial Asian systems.

mpe wrote on 2020-12-03, 19:29:

I think they inially planned to milk the 486 platform for longer, but quicky abandoned it in favour of Socket 5 / Pentium.

It appears in some proprietary/industrial Asian systems.

Intel (per their 1990 roadmap) had originally planned to be at 200mhz in 1999

They planned for socket 4 to exist longer and span into 80mhz and oddly if AMDs 386 had not of released we would have likely saw the Pentium 33,40 and 50mhz clockings release alongside the dx2

They had planned for 486 to last longer and to produce a DX4 Pentium overdrive which assumably would have bridged beyond 100mhz.

But due to actual competition (and Intels anti competitive cohoes and arrogance) none of that happened they slid up the roadmap releasing some half baked chips earlier than expected like the original Pentium 200
they also abandoned the 486 earlier because they didn’t view it as profitable and their extremely delayed overpriced overdrives were not as popular as hoped (1.5years earlier and they would have sold not only well but at the high prices, it’s likely OEMs might have even came from the factory with the p24t pre-installed earlier on) but Intel was too afraid people would upgrade instead of buy a new system so they fatally delayed overdrives to the point they were way too late)

The odd part is slot 1 would have never occurred if speed, memory and cache pressures didn’t exist, given a more leisurely pace and lower clock speeds they would have gotten everything back on die assumably skipping that generation of cartridges going straight to s370.

Had chip companies chilled out and possibly with stronger recycling and import laws it’s possible we would have taken longer to get higher clocks and honesty the world would have been better for it, instead we had a decade of unreliable overheating top of the line today and obsolete and unusable in 6 months, followed by the stagnation and malaise in the 2000’s

Many consumers were exited in those years but many more became bitter leading to the “computer less” generation of hand held devices

rmay635703 wrote on 2020-12-03, 22:56:

Many consumers were exited in those years but many more became bitter leading to the “computer less” generation of hand held devices

This was a very interesting thread. I always wondered what happened to the Socket 6. The explanation that Intel decided to change their plans and burn down the 486 ecosystem once competitors arrived and their replacement was ready makes total sense for Intel's selfish gain. But another way to look at it is that by using the carrot & the stick, intel encouraged consumers to move to an architecture with a more powerful FPU that they needed, but totally paid dividends for the longevity of x86 ecosystem. If PowerPC macs were the only way for consumers to play Quake in 1995, things might have been very different.

Processors certainly evolved ridiculously fast fast in the 1990's, but I wouldn't say that x86 stagnated in the 00's. Started the decade by breaking the 1Ghz barrier. It moved to 64 bit by 2003, produced consumer dual core chips by 2005, incorporated the memory controller, and before the decade ended, we saw the introduction to the Core architecture, which killed off the PowerPC, Itanium, and nearly everything else that didn't run z/OS or need battery life.

If x86 had evolved more slowly in the 00's, all servers might all be using Itanium now, and the world would probably be worse off for that. High prices, proprietary designs, higher power consumption, lower performance. (Itanium fan boys, feel free to tell me I'm wrong)

x86 has always been in that unenviable position where it has to keep advancing quickly while maintaining backwards compatibility. It had such a head start that even after 8 years of stagnation since AMD's K10 & intel's Ivybridge, it still owns the content creation realm. How many people post in this form from a tablet if they have to write more than 2 sentences?

Things certainly look exciting for the 2020's. We'll see what ARM can do, but considering that 2 of the big three console makers use x86, I wouldn't really say that the kiddies are all flocking to ARM on the consumer side. If intel had been willing to license their IP for greater consolidation, paired up with a modern software development team and didn't mind lower profit margins, we might have seen a lot of x86 phones. Probably still could.

produced consumer dual core chips by 2005

Dual-core was indication of stagnation though, which was quickly followed by doubling down with quad core CPUs.

rmay635703 wrote on 2020-12-03, 22:56:

like the original Pentium 200

What issue did the Pentium 200 have? The early socket 5 Pentiums had the floating point issue but that was even before the time of socket 6 which was announced some time in 1994 along with the only chips that would have worked with it.

Not sure about the benefits you're suggesting from a slower rollout of speeds though. I think the past decade was evidence enough that when Intel lacks direct competition, they don't really seek to innovate, so much as find mediocre ways to eek out more performance from their current design while charging a premium price for minor improvements. Netburst was another example of the wrong road taken and maintained by intel only to be deadended years later and overtaken by AMD for a short time once again, but only after thousands of dustbuster space heaters had been made.

The only "issue" I can think of with the P200 was that it ran hot enough that it couldn't use the ceramic package all other classic Pentiums had.

Im not saying competition is bad, Intels reaction to it was

A slower rollout would have resulted in less ewaste and hopefully more reliable systems along with better software support and upgrade possiblities.

There was a LOT of total garbage in the mid to late 90’s into the early 2k like broken modems right out of the box, unstable chipsets out of the box, memory incompatibilities we didn’t see in earlier systems .

A lot of equipment took a major dive in that time due to how competition and dumping practices were handled. In the early 90’s and late 80’s cheap junk usually at least worked, mid late 90’s you had flakey comm ports, software and hardware.

Some companies were forced out of the us due to legally bad quality issues.

I loved every minute of the 1995-2001 era but didn’t like the side affects that period had on everything else after that point.

Anonymous Coward wrote on 2020-12-04, 06:28:

The only "issue" I can think of with the P200 was that it ran hot enough that it couldn't use the ceramic package all other classic Pentiums had.

The original P200 offered minimal gain over the p166 at a very high price with minimal product available

That architecture was cache and bus constrained at that speed point and was released more as a reaction to competitors while checking a box in theory only.

The Pentium MMX 200 was a proper chip the original was made as a matter of pride only and the fact no one had one shows that fact.
Myself Trying to buy one in the day resulted in finding no one had inventory or they were very expensive.

The 1ghz slot 1 is another example of a pride offering that didn’t really exist in the marketplace (there are others as well)