• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Does FSB matter anymore?

Sukhoi

Elite Member
In the past (I'm talking i440BX here) FSB speed was a big thing. AGP/PCI/RAM speeds were directly related to it.

With a new motherboard like the EPoX 4BDA2+, does the FSB speed matter at all (forget the higher speeds you can overclock to with higher FSBs)? AGP/PCI is always at 66 MHz/33 MHz. RAM has its own ratio too now. So theoretically shouldn't a 175 MHz FSB with a 1:1 RAM ratio (DDR350) be exactly the same speed as running a 131 MHz FSB with a 4:3 RAM ratio (DDR350)? Am I missing some big thing here? 🙂
 
yes it matters - but not as much as the old i440BX days.

1:1 memory does have a slight advantage over using memory ratios to achive the same speed.
 
Hmm, forgot about the memory ratio thing..I wish there wasn't a latency hit for running memory at a different speed than FSB.

So other than possibly being able to run a 1:1 RAM divider, does a higher FSB have any speed advantage?
 
Well, maybe not for 1 or 2 MHz, but if you can increase your FSB 25+MHz, you'll notice. If you want to test this yourself, find yourself a good benchmark; something that represents what you do most often. Sandra is a good benchmark for comparisons on the same hardware. Sandra might give some pie-in-the-sky numbers when comparing different architectures, but for measuring differences in settings its pretty good.

Grab yourself a motherboard that can alter the multiplier, and an unlocked chip or unlock one yourself. Thunderbird, Duron, AthlonXP, doesn't really matter. Try different settings that yield the same speed like 10x100 and 7.5x133. You'll see a pretty good difference.
 
Doesn't running the CPU at a higher FSB give it a theoretically higher maximum bandwidth that it could utilize?

By running a system at a lower FSB I belive that you are limiting the amount of bandwidth that could possibly be available to the CPU. At some point you will constantly be starving the CPU of data and will limit your performance. With a 175Mhz FSB you've given the CPU almost 1/3 more bandwidth than at a standard 133Mhz.

Also, running a CPU at a higher FSB will equal lower clock multipliers. If a CPU has to wait one clock pulse for data, it's actually wasting clock_multiplier CPU cycles. The lower the multiplier, the less CPU time wasted.

I might be somewhat off about the theoretical bandwidth part but the clock multiplier thing definitely is true. 🙂 If increasing FSB wasn't advantageous we probably would have seen a LOT more work put into developing better asynchronous bus technologies (like RAM/AGP/PCI ratios). Also synching asynchronous busses can be difficult...

Gaidin
 
Back
Top