FSB (Front Side Bus) confusion
I have an AMD XP 1800+ and when running a cpu program, I get the following: CPU AMD Athlon XP/MP/4 1500 MHz FSB 133 MHz Memory 1024 MB Why does FSB say 133 Mhz when the processor is capable of 266 Mhz? Or am I confusing something here?
I have an AMD XP 1800+ and when running a cpu program, I get the following:
CPU AMD Athlon XP/MP/4 1500 MHz
FSB 133 MHz
Memory 1024 MB
Why does FSB say 133 Mhz when the processor is capable of 266 Mhz?
Or am I confusing something here?
CPU AMD Athlon XP/MP/4 1500 MHz
FSB 133 MHz
Memory 1024 MB
Why does FSB say 133 Mhz when the processor is capable of 266 Mhz?
Or am I confusing something here?
Participate on our website and join the conversation
This topic is archived. New comments cannot be posted and votes cannot be cast.
Responses to this topic
It's "double-pumped", just like how the P4s are "quad-pumped" from a real 100MHz FSB to 400MHz (or 133 to 533). It actually isn't called "double-pumped" as AMD has some other stupid name for it. But basically it's another multiplier that is irrespective of the memory speed. So, the CPU can have a 266FSB while the RAM is still at 133. That's why you can see all kinds of crazy dividers and asynchronous operation of RAM vs. CPU.
your memory (if your memory is DDR) and you cpu FSB are running at the same speed. 133 mhz, but that 133 is double data rate, meaning it sends and recieves data on the rise and fall of the signal...
single data rate sent data only on the rise of the signal, DDR effectively doubles the amount of data transmitted.
it is commonly referred to as 266 mhz FSB, but that is actually erroneous, a marketing gimmick to indicate data throughput equal to that of a 266 mhz single data rate bus.
want another fun one?
everybody seems to think that pc800 rdram has an 800 mhz bus... but that is not so. pc800 runs on a quad pumped bus at 100 mhz (100 x 4 = 400) but the chip INTERNALLY moves data equivalant to 800mhz. it has to run at sutch a high clock rate to make up for the inherently high latency in the design.
single data rate sent data only on the rise of the signal, DDR effectively doubles the amount of data transmitted.
it is commonly referred to as 266 mhz FSB, but that is actually erroneous, a marketing gimmick to indicate data throughput equal to that of a 266 mhz single data rate bus.
want another fun one?
everybody seems to think that pc800 rdram has an 800 mhz bus... but that is not so. pc800 runs on a quad pumped bus at 100 mhz (100 x 4 = 400) but the chip INTERNALLY moves data equivalant to 800mhz. it has to run at sutch a high clock rate to make up for the inherently high latency in the design.
Quote:want another fun one?
everybody seems to think that pc800 rdram has an 800 mhz bus... but that is not so. pc800 runs on a quad pumped bus at 100 mhz (100 x 4 = 400) but the chip INTERNALLY moves data equivalant to 800mhz. it has to run at sutch a high clock rate to make up for the inherently high latency in the design.
kinda the reason why intel has such high clock rates and amd can run processors at lower speeds and get similar performance...
everybody seems to think that pc800 rdram has an 800 mhz bus... but that is not so. pc800 runs on a quad pumped bus at 100 mhz (100 x 4 = 400) but the chip INTERNALLY moves data equivalant to 800mhz. it has to run at sutch a high clock rate to make up for the inherently high latency in the design.
kinda the reason why intel has such high clock rates and amd can run processors at lower speeds and get similar performance...
Quote:kinda the reason why intel has such high clock rates and amd can run processors at lower speeds and get similar performance...
umm kinda... because the p4 has a very deep pipeline, if it messes up in its branch prediction it has to throw away everything in the pipline, with a pipeline 20 ish stages deep, that can quickly cost performance.
athlonxp's are faster at a lower clocked speed through effiency and intelligence of design.
they do MORE per clock cycle.
p4 were made to do a little less per clock cycle but run at mutch higher cycles (IE higher "speed")
in other words:
Athlonxp = Made to run programs fast.
p4 = made to run chip fast.
i am NOT saying one is better than the other, they both have their design benefits and flaws.
umm kinda... because the p4 has a very deep pipeline, if it messes up in its branch prediction it has to throw away everything in the pipline, with a pipeline 20 ish stages deep, that can quickly cost performance.
athlonxp's are faster at a lower clocked speed through effiency and intelligence of design.
they do MORE per clock cycle.
p4 were made to do a little less per clock cycle but run at mutch higher cycles (IE higher "speed")
in other words:
Athlonxp = Made to run programs fast.
p4 = made to run chip fast.
i am NOT saying one is better than the other, they both have their design benefits and flaws.
