|
|
Post by austin86 on May 4, 2022 13:27:22 GMT -5
Back when I used to bench I was never able to hit 180fsb on any 440bx board, most boards seemed to top out around 170-175. From what I recall the PCI/AGP bus dividers were never up to the job. I ran into bad video corruption fallowed by system lock ups.
Looking at a chipset diagram it seems that the ISA bus is ran off the south bridge and not the NB. Would I be right to think a ISA video card would not run into the same problems as a PCI and AGP card would when pushing past 170fsb? |
|
|
|
Post by Mr.Scott on May 4, 2022 15:03:56 GMT -5
Possibly.
Something I've never tried.
|
|
|
|
Post by cbjaust on May 4, 2022 23:38:00 GMT -5
|
|
|
|
Post by Macsbeach98 on May 5, 2022 1:10:40 GMT -5
Any sychronis board will top out once the IDE hard drive gets over around 41Mhz.
133 Mhz to 170Mhz is a jump of 37Mhz divide that by 4 is 9.25Mhz add that to 33Mhz equals 42.5Mhz I would say your hard drive is whats collapsing.
I would try a IDE to Sata adaptor and a PCI video card MX400 are hard to find these days but a S3 Trio should be easy to find for benching.
|
|
|
|
Post by austin86 on May 5, 2022 8:02:21 GMT -5
Any sychronis board will top out once the IDE hard drive gets over around 41Mhz. 133 Mhz to 170Mhz is a jump of 37Mhz divide that by 4 is 9.25Mhz add that to 33Mhz equals 42.5Mhz I would say your hard drive is whats collapsing. I would try a IDE to Sata adaptor and a PCI video card MX400 are hard to find these days but a S3 Trio should be easy to find for benching. The IDE ports are ran of the SB on 440bx and not the PCI bus from what I recall, but the SB is linked to the NB via the PCI bus so who knows. I did not have any problems benching in dos at 175fsb ether, it was not until the GUI loaded that I ran into problems. After I hit 170fsb I switched from a voodoo3 to a PCI s3 virge and was good for about anther 5mhz on the FSB so I thought it was the video cards crapping out. I did not try a sata adapter but I did try a ISA ide card and a few PCI cards and none on those helped. I'll have to give a sata adapter a try. In theory I could go without a video card to and use a scrip to run the bench and screen shot but that would be a lot of work lol. And there are ways to get video out of a serial port too so its not like I don't have options.
|
|
|
|
Post by Macsbeach98 on May 5, 2022 16:08:23 GMT -5
It is commom for older boards to stop working around where you are saying and that was generally the thing when you raise bus the south bridge goes up to another thing that can help is different IDE drive but they mostly top out around 41 to 42 Mhz.
The thing is as you clock it up the southbridge and its output rises too and most IDE drives cant take the increase.
|
|
|
|
Post by austin86 on May 6, 2022 11:44:07 GMT -5
It is commom for older boards to stop working around where you are saying and that was generally the thing when you raise bus the south bridge goes up to another thing that can help is different IDE drive but they mostly top out around 41 to 42 Mhz. The thing is as you clock it up the southbridge and its output rises too and most IDE drives cant take the increase. Wait the board or the hdd can't take it? |
|
|
|
Post by Macsbeach98 on May 6, 2022 16:16:18 GMT -5
The HDD is is usually the first thing to clap out followed by other things but it is the actual HDD.
Not many IDE drives will go much over 41 Mhz
|
|
|
|
Post by Mr.Scott on May 6, 2022 18:26:06 GMT -5
Seagates hold longer, and the small Quantum Fireballs. 43-46 MHZ usually, pending the drive.
|
|
|
|
Post by austin86 on May 7, 2022 19:49:36 GMT -5
Seagates hold longer, and the small Quantum Fireballs. 43-46 MHZ usually, pending the drive. that would explain why I hit a wall at 175sfb every time I benched 440bx I was using a Fireballs. odd that it s tarted with video corruption. I'll try a sata drive when I get the system up. |
|
|
|
Post by antinomy on May 7, 2022 20:43:23 GMT -5
Like everyone said above - get a IDE-SATA adapter and try cards like GF2MX, they can go high on AGP frequency. And don't forget that memory runs in sync, 175 is pretty high, not all memory can stand this.
|
|
|
|
Post by Mr.Scott on May 8, 2022 12:10:19 GMT -5
Highest I've had on mine is 168 so you're doing pretty well at 170+.
|
|
|
|
Post by austin86 on May 9, 2022 12:00:09 GMT -5
Highest I've had on mine is 168 so you're doing pretty well at 170+. I was really pushing it and it was far from stable. But from what I recall 165ish was stable to the point that I could run super pi32m. This was on a modded be6-ii and air. I can't think of anther 440bx board that could do that. I wanted to DICE it but was going to sell the board. That was about the time I got out of benching and I had long stopped using dice then. I plan to bench my bx133-raid a little once I get it up and going soon, no dice just water.
|
|
|
|
Post by Mr.Scott on May 9, 2022 17:13:44 GMT -5
I did 32m at 168 so now I feel better.
|
|
|
|
Post by austin86 on May 11, 2022 10:46:07 GMT -5
I did 32m at 168 so now I feel better. TBH I don't recall 100% how far I got 32m stable. I know it was mid 160s, and I want to say upper 170s was were I maxed out at. I do recall using a 750 or 700mhz coppermine Piii at the time. Those 750Piii were some good clockers. |
|
|
|
Post by Mr.Scott on May 11, 2022 18:28:48 GMT -5
I did 32m at 168 so now I feel better. TBH I don't recall 100% how far I got 32m stable. I know it was mid 160s, and I want to say upper 170s was were I maxed out at. I do recall using a 750 or 700mhz coppermine Piii at the time. Those 750Piii were some good clockers. Indeed. I think my 168 run was on a 733. |
|
|
|
Post by austin86 on May 11, 2022 20:25:35 GMT -5
TBH I don't recall 100% how far I got 32m stable. I know it was mid 160s, and I want to say upper 170s was were I maxed out at. I do recall using a 750 or 700mhz coppermine Piii at the time. Those 750Piii were some good clockers. Indeed. I think my 168 run was on a 733. I been told that the 700-750 Piii Es are higher clocked 1ghz ish chips sold as 100fsb chips. No clue if that’s true, but it makes since, I seen some older versions of cpuz that could not be tell a 750mhz PiiiE apart from a 1ghz 133fsb chip |
|
