![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Difference between uping FSB and multiplier
- Thread starter Hedger72
- Start date
D
Deleted member 79192
Guest
Wouldn't upping the FSB fry the PCI and/or AGP slots? And on newer mobos with locked PCI and AGP bus speeds, wouldn't it still overclock the memory when upping the FSB? Since the memory speed is based on the FSB:Memory ratio? 
I'm feeling nerdy today.
I'm feeling nerdy today.
Originally posted by Matt Welke
Wouldn't upping the FSB fry the PCI and/or AGP slots? And on newer mobos with locked PCI and AGP bus speeds, wouldn't it still overclock the memory when upping the FSB? Since the memory speed is based on the FSB:Memory ratio?
I'm feeling nerdy today.
the FSB is a bus
PCI is a bus
the PCI bus is connected to the FSB
not the individual slots
Terminus
[H]ard|Gawd
- Joined
- Jul 30, 2001
- Messages
- 1,838
the FSB on the AMD64 line is on the cpu die so it WILL get hot in that case
It is better to achive the max in FSB with a low multiplyer, than lower the FSB and Raise the multiplyer untill the over all speed evans out. That will help you achive the best overall OC. The FSB is used for how fast the RAM is communating with the rest of the compter and it is inportant to have that high so that alot of information can go thought the bus to get to the CPU, the more info that can get to the CPU the more work can get done over the long hall.
There is no FBS on the K8 line of CPUs....rather a HyperTransport link. Read the FAQ about it.
Apparently we need to address the subject of this thread in a FAQ or something, because many of you are wrong or miss informed. A good job goes out to cbakey for his description in simple terms.
mwarps: I think you need to do that semiconductor physics class soon, nothing too hard, they are not EEs.
O and you cant post anymore, as you are 1337
Spiritual Machine
Weaksauce
- Joined
- Mar 21, 2002
- Messages
- 79
nevermind
That is a different discussion.
In the case of the K8, the HT speed is run on a divider of the CPU clock speed. Assuming you leave the CPU clock the same and lower the divider for a higher HT speed, the CPU temp would be about the same. Overclocking the CPU is the only way this would increase the temps.
In general systems, northbridge and CPU being separate, overclocking the FSB *may* increase the CPU temp a bit as the CPU is getting more data to work with, this is the benefit of a higher FSB. If the CPU can recieve more data to work with using a higher FSB it follows that more work is done, more transistors going, less stalls in the pipe waiting for data. The CPU temp can go up a bit with a higher fsb, but it is still more a function of the CPU doing more work rather than the FSB directly causing a temp increase.
D
Deleted member 79192
Guest
So, in the end, a FSB increase is more efficient than a multiplier increase. Just don't bring the FSB too high, or you'll go too high for the memory to handle.
Have ya tried this out, can ya get back to me with the results? I'd love to know how we're violating the laws of physics this week.
The portion of the chip that deals with the actual clock interface is *so small* compared to the rest of the chip that the heat difference is entirely negligible, milliwatts at worst.
Originally posted by CentronMe
There is no FBS on the K8 line of CPUs....rather a HyperTransport link. Read the FAQ about it.
Apparently we need to address the subject of this thread in a FAQ or something, because many of you are wrong or miss informed. A good job goes out to cbakey for his description in simple terms.
mwarps: I think you need to do that semiconductor physics class soon, nothing too hard, they are not EEs.
O and you cant post anymore, as you are 1337
But but but. Aww come on!
I'll start writing one up for a FAQ/sticky. Maybe relic will be kind enough to plop it up there.
We'll see.
The weekend awaits!