I have a Packard bell like that.
The VGA is a cirrus logic vlb on the motherboard.
Any isa card will be slower.

You can use philcomputerlab's dos benchmark suite for your tests. Mine are done with it. Google it up 😀

Most probably no isa card is ever gonna be as fast as your onbard et4000.

sunaiac wrote:

I have a Packard bell like that. The VGA is a cirrus logic vlb on the motherboard. Any isa card will be slower. […]
Show full quote

I have a Packard bell like that.
The VGA is a cirrus logic vlb on the motherboard.
Any isa card will be slower.

You can use philcomputerlab's dos benchmark suite for your tests. Mine are done with it. Google it up 😀

Most probably no isa card is ever gonna be as fast as your onbard et4000.

Interesting!
So Packard Bell also went that route... Yes, Compaq also switched to Cirrus Logic after going TSENG for built-in SVGAs.
I don't believe mine is a VLB-like motherboard channel, because it is older than the VESA standards.

I'll surely google the philcomputerlab's dos benchmark suite of tests. Do you think these tests would saturate whatever bandwidth it has, even if it is as large as a real VLB bus?
Yeah, I also believe my onboard ET4000 is faster than a ISA one, just I'd like to exactly quantify that difference!

Like doing the math, like:
• For EACH benchmark, measure the %increase of my onboard W32 compared with ISA ones (I hope I'll find public results from some ISA W32s available somewhere)
• Calculate an AVERAGE %increase based on all the benchmarks as above
• Then, knowing the ISA bandwidth, which is 16 MB/sec, assigning this value to the average result of the public benchmarks of the ISA W32s, and then adding the above mentioned AVERAGE %increase of my onboard built-in W32, I would "convert" the results of the benchmarks for my W32 in terms of MB/sec, so to finally compare ISA bus with my mysterious Compaq "local bus".

I could then even compare it with VLB and PCI specifications, since they all go in terms of MB/sec!

Any thoughts on this?

Hey, have a look also at my new "the perfect 486 motherboard" thread if you can, I'd love to hear your opinion about that!

Thanks!

that mobo comes in a version with onboard graphics which are connected via local bus

aries-mu wrote:

Possibly, with the EISA bus forced to run at around 12-16 MHz rather than the crappy standard 8 MHz which, being a 32 bit bus, would be not bad at all.

EISA runs at 8.3MHz, but it isn't exactly slow. It can do up to 33MB/sec. I think 16-bit ISA is more like 5MB/sec.

Anonymous Coward wrote:

aries-mu wrote:

Possibly, with the EISA bus forced to run at around 12-16 MHz rather than the crappy standard 8 MHz which, being a 32 bit bus, would be not bad at all.

EISA runs at 8.3MHz, but it isn't exactly slow. It can do up to 33MB/sec. I think 16-bit ISA is more like 5MB/sec.

yeah it's a decent bus

Hey. Anyone knows, how performs DX4-75 (25x3) 16KB in comparsion to DX2/66 8KB? Will it be faster, slower, or about the same? Does anyone know link to any test?
Problem is, I have only 5x86 133, and I need to simulate as close as I can to DX2-66 , which I don't have. And it knows only 3x multiplier lowest. So I was wondering, how 25x3=75mhz with 16KB cache goes against DX2-66, but cannot find any direct comparsion, because 75Mhz speed is rare, and almost noone use it in tests.

W.x. wrote on 2021-02-27, 06:02:

Hey. Anyone knows, how performs DX4-75 (25x3) 16KB in comparsion to DX2/66 8KB? Will it be faster, slower, or about the same?

Its is sightly faster (about 10-15%) as the cache and clock easily compensates the lower bus speed.