"Better" is always subjective. You need to define what they're supposed to be better at for a good answer.

In any event, the Athlon 1000B (given you're saying KT133, I'm assuming the 1000C with 133MHz FSB is out, as that needs a KT133A) is only a little bit faster than the Duron 1000 in most tasks, although how much depends on the application and how it fits into cache, and if something actually uses SSE (period correct: very rare) the Duron would be faster. A Duron 1300 will almost always be faster than an Athlon 1000B, but again, the margin depends on the application.

One thing to look out for: quite a few KT133 boards are BIOS limited to Thunderbird/Spitfire and don't support Palomino/Morgan cores. Check your motherboard in any event.

I think also Duron Morgan was based on the Palomino architecture and was a lot cooler than Thunderbird!?

The Morgan Durons have less cache than the palomino but still defaults to 1.75 volts so they do run decently hot. You can probably undervolt it a bit, my 1000Mhz Morgan runs fine on 1.65 volts.

And as dionb wrote, a 1300Mhz Duron will beat a 1000Mhz Athlon B almost all of the time You will also get some small improvements with SSE enabled games, like extra added effects. The Athlon has way more cool factor though 😁

compiling linux under duron 800 back in the day took _ages_; this is a cpu only for old games imho

That tiny L2 cache has to be a major caveat in some scenarios.

Is there any real advantage to the Morgan core having SSE? Considering Steam no longer runs on WinXP that is not a consideration anymore.. Maybe certain Linux distributions will run on a Duron 1300 but not on an Athlon 1000? Do any games from 2000-2001 have SSE support?

appiah4 wrote:

Is there any real advantage to the Morgan core having SSE? Considering Steam no longer runs on WinXP that is not a consideration anymore.. Maybe certain Linux distributions will run on a Duron 1300 but not on an Athlon 1000? Do any games from 2000-2001 have SSE support?

Offhand I know that Battlezone 2 and Unreal Engine use KNI/SSE. Battlezone 2 has some slight visual enhancements with it enabled. I remember comparing enabled/disabled with the game in a paused state to see what it actually did. 😉

But no I don't recall ever coming across anything in a game that made SSE super exciting to have. Nothing requiring it either.

Yep, it's a bit like MMX a few generations earlier. Intel's marketing hyped the new instructions up massively, so that people associated the better performance of the P55C to them. Nope, it was just the doubling of the L1 cache that did it. Unless you had an AMR, no application you would have sensibly run with such a machine would require them.

dionb wrote:

Unless you had an AMR, no application you would have sensibly run with such a machine would require them.

On a side note, I know about exactly 3 DOS programs that needed MMX and kind of runs on a Pentium MMX.
The most most important of the trio are the Impulse Tracker MMX drivers. They do resonant filtering that the original non-MMX drivers won't. But there aren't too many .ITs that employ these filters.

The other two are scene demos, Event Horizon and Heaven 7. They are okay-ish on an overclocked P-MMX (and good tools to test P-MMX overclock stability) but run better on faster CPUs anyway.

And I know no DOS programs that would need SSE. Except for an experimental Mandelbrot generator I wrote in assembly when I got my Athlon XP.

L1 cache is huge on both CPUs and L2 cache on K7 Athlons is comparably slow. Due to that, size difference between Duron/Athlon doesn't hurt as much.

But no I don't recall ever coming across anything in a game that made SSE super exciting to have

SSE is also supported by drivers. Although the performance impact is probably mild.

Considering Steam no longer runs on WinXP that is not a consideration anymore

It's not in consideration for quite a while now, because Steam requires SSE2.

Descent3 mentions Pentium III "enhancements" in the patch notes but i'm not sure what that entails. Other Interplay games like Messiah were also supposed to support SSE but i'm not sure they did.

I guess video drivers did support SSE but didn't seem to gain much benefit from it.

well

appiah4 wrote:

Is there any real advantage to the Morgan core having SSE? Considering Steam no longer runs on WinXP that is not a consideration anymore.. Maybe certain Linux distributions will run on a Duron 1300 but not on an Athlon 1000? Do any games from 2000-2001 have SSE support?

Wow, I somehow never knew that the early Athlons and Durons didn't have SSE. That has to really hurt usability for more modern tasks. With a Pentium III or early P4 it can be interesting to see what newer software they can handle, but I imagine that anything without SSE is going to be severely limited. The lack of SSE2 already hurts the P3 and Athlon XP in newer tasks.

SPBHM wrote:

well […]
Show full quote

well

That's a little....suspicious. 🤣

swaaye wrote:

SPBHM wrote:

well

[...]

That's a little....suspicious. 🤣

No, that's an exceedingly rare example of something actually using SSE to the full.

Here are more regular benchmarks:
https://hexus.net/tech/reviews/cpu/211-amd-du … hz-morgan-core/

1GHz Morgan about 0.5%-2% slower than 1GHz Thunderbird (assumedly B given closeness of results) depending on bench.

https://m.hardocp.com/article/2001/09/21/1ghz … re_duron_review

1GHz Morgan ~12% slower than 1GHz Thunderbird C in Q3A

Yeah the Rage Dispatch results say to me that the non-SSE version needed some work. Doubling is a bit too good to be true for SSE outside of synthetics I think.

SPBHM wrote:

well […]
Show full quote

well

so I found the source https://www.anandtech.com/show/254/10, but I still have no clue what exactly "Rage's Dispatch demo" is.
I was right there in the thick of it in 1999, both gaming and working in the industry, and I cant recall any Rage's Dispatch game. Is this it https://www.mobygames.com/game/rage ?

SPBHM wrote:

well […]
Show full quote

well

So much performance increase with SSE? 😲

NostalgicAslinger wrote:

[...]

So much performance increase with SSE? 😲

SSE allows you to manipulate multiple floating-point data with a single instruction. If you have an application that is bottlenecked on doing exactly that and supports these instructions, you will save multiple instructions so multiple cycles and get a performance increase in the order of magnitude you see here.

Question is: how realistic is this workload.
Answer: not, at least not for the vast majority of applications you might throw at one of these CPUs.