Afaik that's not true.
I had an Am486DX2-66 back in the day. It died, and I replaced it with an Intel 486DX2-66.
My benchmark figures were 100% the same.
As far as I can tell, the Am486 is a carbon-copy of the Intel chip. There's one minimal flaw however, that I cannot explain. For some reason, I could install OS/2 on the Am486, but after installation, it bluescreened when it tried to boot. It could only boot to safe mode.
Once I had it replaced with the Intel, OS/2 ran fine on that system. So it has to be something in that Am486 that's not entirely compatible. But OS/2 is the only software I've ever had problems with on that chip.

I think any comparison where an AMD part beats Intel is only if it's apples and oranges, e.g. 16 KB vs. 8 KB cache, Write Back vs. Write Through. Intel usually wins clock for clock. WB versions of AMD stuff are easier to source, though.

AMD's good reputation in performance comes from higher clocked parts, like the 386DX-40 and 486DX4-120 (and DX5-133) that Intel never matched.

Here's a 2014 post by gerwin where he tests Intel and AMD DX4/100 cpus in the same system. He also downclocks both to 66Mhz. According to his results, the Intel is hair faster at both 66 and 100Mhz.
486 VLB UMC-Chipset, what is it?

But DX4s come in various flavours... 8K cache or 16K cache, and can be either WT or WB cache.
Afaik the DX2 chips are all the same, and AMD and Intel perform exactly the same (others, such as IBM/Cyrix are different though).

Scali wrote:

But OS/2 is the only software I've ever had problems with on that chip.

another one is netware i think

Scali wrote:

But DX4s come in various flavours... 8K cache or 16K cache, and can be either WT or WB cache.
Afaik the DX2 chips are all the same, and AMD and Intel perform exactly the same (others, such as IBM/Cyrix are different though).

there are WB and WT dx2-66's as well

well, it seems it's as I thought. he must of got some bad information somewhere or misspoke.

and part of that issue is that it's LGR - a very popular channel. any misinformation he''ll mention will be echoed among the crowds of easily impressible retro millenials out there and youtube doesn't allow video editing to fix mistakes (except for annotations, where he could put a big block of text. youtube sage status carries great responsibility

Also Moraff had obnoxiously shilled AMD 286-486 cpus in the day as being faster so that might also have something to do with it 😜

The 486 LGR built in that video had no motherboard cache, likely why it ran Duke Nukem 3d so badly.
My Intel dx2-66 runs it perfectly smoothly at 320x200. With a AMD 5x86@160mhz it's smooth at 640x480.

leileilol wrote:

Also Moraff had obnoxiously shilled AMD 286-486 cpus in the day as being faster so that might also have something to do with it 😜

The whole thing is a bit weird anyway, since AMD was always a generation behind.
By the time fast 286s came around, Intel had already abandoned them, and moved to 386. By the way, Harris made the fastest 286, at 25 MHz.
Likewise, the 486 was already out when AMD's 386 arrived. And the Pentium was out by the time the Am486 was launched.
So yes, these CPUs may have been faster than offerings from the same family, but that's only because Intel had already moved to a newer architecture, and delivered faster CPUs there.
People tend to make it sound like AMD had the performance crown back then, but they didn't. The only time that happened, was with the Athlon.

Scali wrote:

The whole thing is a bit weird anyway, since AMD was always a generation behind. By the time fast 286s came around, Intel had al […]
Show full quote

The whole thing is a bit weird anyway, since AMD was always a generation behind.
By the time fast 286s came around, Intel had already abandoned them, and moved to 386. By the way, Harris made the fastest 286, at 25 MHz.
Likewise, the 486 was already out when AMD's 386 arrived. And the Pentium was out by the time the Am486 was launched.
So yes, these CPUs may have been faster than offerings from the same family, but that's only because Intel had already moved to a newer architecture, and delivered faster CPUs there.
People tend to make it sound like AMD had the performance crown back then, but they didn't. The only time that happened, was with the Athlon.

true, i m totally with you on this.

Well when I played around with some 486 chips, taking AMDs latest and greatest Am5x86-P75, the IntelDX4 was a tiny bit faster clock for clock.

soviet conscript wrote:

There are WB and WT dx2-66's as well

Yea, apparently there was one very late Intel 486DX2WB model, according to Wikipedia: https://en.wikipedia.org/wiki/Intel_80486
Seems to be a spin-off of the DX4, which had WB as well.
Same with AMD: https://en.wikipedia.org/wiki/Am486
That one is even later, probably a spinoff of the 5x86?
I had a regular WT one though, in both cases.

back in the day, I had a AMD DX2-80, and it was rock solid. great chip that was.

Scali wrote:

The whole thing is a bit weird anyway, since AMD was always a generation behind. By the time fast 286s came around, Intel had al […]
Show full quote

leileilol wrote:

Also Moraff had obnoxiously shilled AMD 286-486 cpus in the day as being faster so that might also have something to do with it 😜

The whole thing is a bit weird anyway, since AMD was always a generation behind.
By the time fast 286s came around, Intel had already abandoned them, and moved to 386. By the way, Harris made the fastest 286, at 25 MHz.
Likewise, the 486 was already out when AMD's 386 arrived. And the Pentium was out by the time the Am486 was launched.
So yes, these CPUs may have been faster than offerings from the same family, but that's only because Intel had already moved to a newer architecture, and delivered faster CPUs there.
People tend to make it sound like AMD had the performance crown back then, but they didn't. The only time that happened, was with the Athlon.

It's more like Intel always being a generation ahead of the consumer in those times. 😉
The 486 came out in 1989 and Pentium in 1993, but they were as good as paper launches to the public until years later.

exactly, also pricewise:
noone cares what a rocket the p2 was on release day as no sane person spends that kind of money;
in the sane pricerange amd was offering more power for the same money and therefore was ahead;

firage wrote:

It's more like Intel always being a generation ahead of the consumer in those times. 😉
The 486 came out in 1989 and Pentium in 1993, but they were as good as paper launches to the public until years later.

Back then you didn't have a Xeon range for servers/workstations.
So new architectures started out in the 'professional' market, and over time, prices dropped, and they became mainstream.

I usually had Intel systems though, because AMD was the budget mainstream stuff, where Intel had interesting offerings for the higher mainstream segment where my dad and myself shopped.
Eg, we bought a 486DX2-66 VLB system by the time the Am386DX-40 became popular. And it seems that many people like us did the same, because in the demoscene, a lot of demos were targeting the 486 already, around 1993-1994.
Likewise we got a Pentium 120 at around the time the AMD 5x86 came out.

The 386DX-40 was out in 1992 and the first gaming applications for a 486 came out in 1993, same with the AMD DX2's up to 5x86 releasing in 1994-1995 and Pentium games coming out in 1996. Yeah, you weren't future-proofing with AMD.

Love them or hate them Intel have always been the market leader.

Even when AMD products are technically better its still the case of "we have to make our CPU's as compatible with Intel as possible"
Its why I don't blame AMD for inventing the PR scheme which many said was a marketing scam, no our chips are just more efficient

So back then while Intel was pushing their new tech which they have spent loads of money on R&D for, AMD, Cyrix, etc where happy to spend a lot less on R&D for what even they knew was soon to be obsolete but still has a few years left in the budget hardware category, and use some of the profit into working out their answer to Intel's new CPU.

chinny22 wrote:

Its why I don't blame AMD for inventing the PR scheme which many said was a marketing scam, no our chips are just more efficient

Cyrix used the PR-rating before AMD did, on their 6x86 series.
The rating in itself is not necessarily a scam, although trying to capture the performance of a CPU architecture in a single figure is doomed to fail, which is why things like MIPS and FLOPS were abandoned as performance figures long before PR was introduced, which is basically the same flawed concept.
However, AMD did not adjust their ratings, despite newer, faster generations of Pentium 4 and more software being optimized for Pentium 4, so the numbers became more and more inflated, which could be seen as a scam.

Also, Intel also used their own 'PR' system, known as "iCOMP index": http://www.computerhope.com/jargon/i/icomp.htm
Intel basically faced the same problem: How to explain that a Pentium 66 is faster than a 486DX2-66, or that a PIII at 300 MHz is faster than a PII at 300 MHz.