It's always nice to see benchmarks. I recently got SXL2-50 but haven't yet found the time to do a test drive - don't even know if it works at this point. Would like to run it at 1x40MHz and test with different motherboards. Later I hope to find a board that works fine on 50MHz and run it 1x50MHz.

Woah, thank you very much for this comparison! It is really well done!
I always wanted to build a nice, huge 386-based vintage system and was curious how both of these compared! 😀
So the only thing left I want to know is how well a 486SXL2-50 compares to a 486DLC-40 and how the difference in chipset speeds (25MHz vs 40MHz)
affects overall performance. If it is close, I really could build a nice system with these huge, golden 80s parts (except for the Cyrix).. 😁

Golden 80s? As in 1980s?

I thought more of the shiny, golden computer parts from then. But golden '80s (as in 1980s) is also sweet..
That decade hat some cool stuff - music, fashion, films and people.. So it is valid in either way. 😁

Are any 386 boards from that era stable with a 50 MHz FSB? How about 45 MHz? Were there even 90 MHz oscillators that fit?

I tried 50MHz on an OPTi 495SX board. It wasn't stable even with cache disabled and memory timings lowered. 45MHz seemed to work most of the time.
I think I tried my Peak/DM board too, and 50MHz was no go.

I always felt the extra cache didn't make the chip that much faster, but 5% is still something.

What's the deal with the situation where the SXL is actually slower than a DLC?

Anonymous Coward wrote:

What's the deal with the situation where the SXL is actually slower than a DLC?

I think the precision of measurement is lacking. I should probably have run those 3x each.

feipoa wrote:

Are any 386 boards from that era stable with a 50 MHz FSB? How about 45 MHz? Were there even 90 MHz oscillators that fit?

I have my doubts, about 25-33MHz is the limit. Note that this doesn't apply to all mainboards from that era or their chipsets in general, but only to the huge models.
Their physical layouts might be to large and their traces to long to run reliably on higher frequencies. Some of them did not even use chipsets, but discrete parts.
That's why I was interested in SXL2 chips, as they are running at twice the bus speed. So an old 25MHz board would run them safely at 50MHz.

Anonymous Coward: Does the SXL also have the ability to clock double?

Jo22 wrote:

I have my doubts, about 25-33MHz is the limit. Note that this doesn't apply to all mainboards from that era or their chipsets in general, but only to the huge models.
Their physical layouts might be to large and their traces to long to run reliably on higher frequencies.

Don't nearly all 386 boards from 1990-1993 support a 40 MHz front-side bus.

feipoa wrote:

Jo22 wrote:

I have my doubts, about 25-33MHz is the limit. Note that this doesn't apply to all mainboards from that era or their chipsets in general, but only to the huge models.
Their physical layouts might be to large and their traces to long to run reliably on higher frequencies.

Don't nearly all 386 boards from 1990-1993 support a 40 MHz front-side bus.

I think so, but the boards I'm curious about are the early ones, from '86-89, running at a frequency of about 16-25MHz.
And its not only about the size, but also the older BIOS flavors that I think are interesting. 😀
Wait, here's a sample: http://www.ebay.com/itm/386-BIG-American-Mega … s-/181366747975

I am more drawn to the later 386 boards from the early 90's, although boards which have both 386 and 486 sockets do not interest me.

Yes, the 40MHz SXL does support clock doubling. As far as I can tell, they work at 50MHz with no problems.

I'm interested in boards from late 1991 to late 1992. Generally these boards are full length baby AT and still use a true 386 chipset (not hybrid one). Sometimes they even support DLC chips. The problem with the earlier ones is that they're picky about memory/cpus and have a lot of quirks. The later 386 boards often suffer from faulty DMA controllers, use 486 chipsets and sometimes even have a 486 socket. However, the later ones are pretty darn solid.

SXL works at 50 MHz? Maybe I'll put the 100 MHz crystal into my Baby Screamer and see if it can handle it. I only have 70 ns RAM though. Do you think that will pose a problem?

Just increase the DRAM & SRAM latency and reduce ISA clock. I have the same project in plan (already have all the components) but no time to do it 🙁

I wouldn't do 50MHz in the screamer. I would only try 50MHz on boards using 486 chipsets. It was a major problem for manufacturers to produce boards/chipsets that could handle the 486DX-50, and the VLSI topcat probably isn't up the to job. The Screamer was originally only designed for up to 33MHz CPUs, with 40MHz support being added later.

kixs, this motherboard doesn't have DRAM, SRAM, and ISA clock adjustability in the BIOS, which means the ISA bus will be running at 50/4, or 12.5 MHz. That seems pretty insane.

Anonymous Coward, I am going to try it anyway. If/when it becomes unstable, I will try the UMC 481-based board.

I have this 386 board (says "486 ISA BIOS" on the AMIBIOS chip), is this any good? http://www.uncreativelabs.de/th99/m/U-Z/32980.htm

I've owned two of those exact boards, however mine did not have the local bus slot soldered (just a regular ISA slot), though the pads are present.

Both of my boards had some kind of defect with the 82C206 system chip, and DMA did not seem to be working properly. This problem affected my ability to use soundcards and SCSI controllers, especially when used together with a 486DLC CPU. Definitely not recommended.