Repo Man11 wrote on 2021-01-14, 04:14:
Sphere478 wrote on 2021-01-14, 03:50:
Edited above post. […]
Repo Man11 wrote on 2021-01-14, 03:43:
AFAIK, the only way to get higher than 83 MHz FSB on a TX board is to replace the clock crystal. If anyone has actually done that, someone on this forum will likely know about it. With the PCI bus at 41.5 MHz at 83 MHz FSB, you're already right on the edge.
Edited above post.
Which crystal could I try? Won’t that mess up the system clock? Make it count too fast?
I could maybe use a removable crystal socket and an assortment of different ones?
Also, if you have ideas of how to get to 83 that would be great cause I’m stuck at 75 according to notes
The Tom's article Oldie Tuning goes over how it might be done on the P55T2P4 ( https://www.tomshardware.com/reviews/oldie-tuning,216-4.html ), and I'd imagine the same would be possible on your board. As far as getting it to 83, perhaps there is a place on the PCB where there is a missing jumper that could be soldered into place? Maybe someone else reading this thread might know of a way?
fun read I can't seem to be able to get the tables they mention to load though. I wonder if the links are broken? they might provide clues as to how it all works.
I'm thinking that the 21pin? header for my fsb/ratio may contain direct links to the clock gen but it's weird that they don't have a seperate jumper system for the ratio and the fsb so there must be a connection to the cpu there also.
if I can somehow decipher the pins that will trigger the 2x multiplier and the 83 or 100mhz fsb I will be happy. my guess however is that they may not even be in those pins or next to one another. a set of jumper wires would be handy. and a gerber file for this motherboard 🤣 what do you think the chances are of someone having the gerber file laying around 🤣?
I really want to find a k63 powerleap interposer I think it severs the link between the motherboard and the cpu for the multiplier and allows you to set your own also it allows more voltage control. this board, though I have not found the notes for it yet I remember (I think) that you get 2.0, 2.1, or 2.2 by removing all the voltage jumpers which would work but it would be nice to have full control. does anyone around here have a powerleap laying around?
ultimately it may come down to taping the clock gen pins directly and some of the socket pins with little wires and epoxying a little breadboard with jumper pins to a blank space on the board. i would like to avoid that if I can.
I'm having trouble finding a datasheet for the clockgen though and I haven't found any software that can interface with it
btw, in a reply in that other thread there was a now broken link to utilities for the k6 does anyone have those utilities? i need to download them to play with. 🤣
debs3759 wrote on 2021-01-14, 04:00:
Using an NVME drive via adapters sounds like a waste of money to me. I can't see it being faster than a basic SATA 3 SSD when sending and receiving data via PCI. Good luck if you can pull it off though 😀
🤣 what part of this build gave you the impression that cost to performance ratios were important here 😂 joking aside, the pci to pcie adapter can be used wit gfx cards also there is a video of a guy using one. and it wasn't too bad, like 30$ the m.2 was free, the m.2 to pcie adapter was like 5$ so all in all it wasn't very costly but yes, I do believe that a sata2 might be just as fast on this system in practice. you're forgetting cool factor though!
^^^^here is the vid^^^^