TB 1200 AND PC2100

G

Guest

Guest
I have got 128 DDR PC2100.
Can I use these sticks for a ThunderBird 12x100?.
I´ve heard it needs PC1600 and not PC2100.

Regards and thanks un advance.
 

yoda271828

Distinguished
Dec 31, 2007
527
0
18,980
I think you can do it with a DDR mobo, by using a 9.0x133 multiplier and FSB. I don't see any reason that wouldn't work since your not even overclocking the CPU, and the DDR mobo's are designed to run a 133 FSB. You would, of course, have to unlock multiplier settings on the CPU.
 
G

Guest

Guest
first, you need a DDR-ready motherboard like AMD760;

second, if your T-Bird is a current 100MHz FSB version, the AMD760 mobo will automatically use the PC2100(133MHz DDR) DDR-RAM in PC1600(100MHz DDR) mode.

so, you will be able to use your 100MHz-FSB T-Bird and DDR-RAM together, but you will waste more than 20% of the bandwidth that your PC2100 DDR-RAM is able to provide if combined with a 133MHz-FSB T-Bird.


-----------------------------
Some are ignorantly happy,
While some, happily ignorant.
 
G

Guest

Guest
actually, with current stepping, a 100MHz-FSB T-Bird or Duron is not ready to reach 133MHz FSB, even when multiplier's downclocked.

the 100FSB T-Birds and Durons of current stepping that can reach 133FSB are real gems. you may find one out of more than a dozen, whereas stability is another issue.


-----------------------------
Some are ignorantly happy,
While some, happily ignorant.
 

yoda271828

Distinguished
Dec 31, 2007
527
0
18,980
What is different about the 133FSB Athlons? I though that they were just the same CPU, but with a lower multiplier that would run at a higher clock speed (ie. a new 133FSB-1.2Ghz is just an old 100FSB-900Mhz that will run at 1.2 Ghz).

I always thought that FSB was completely independant of CPU core speed. How is it that some PIII's can reach reach over 150Mhz FSB? I though it was the intel chipset, not the CPU. Isn't it the AMD EV6 bus that creates the problem overclocking the FSB on an Athlon. Using that logic, if the Mobo could support a 133Mhz FSB, why can't the CPU? Please enlighten me.
 
G

Guest

Guest
actually it is all about stepping and micron process, and sometimes the material of the interconnects(aluminium or copper) also.

that is why when Intel recalled its PIII 1.13GHz, Intel said that it might need a new stepping. of course, they could also use .13 micron process to make it work, but a new process would be way too much more expensive. and so would the implementation of copper interconnects.

also please note, as Tom says, the original 1.13GHz PIII ran flawlessly at 850MHz. 1.13GHz = 133MHz x 8.5; and 850MHz = 100MHz x 8.5. so you see the influnence of FSB. and Intel motherboards should be very FSB-friendly. FSB overclock simply generates much more heat and puts the CPU under higher stress to its design limits than multiplier overclock does.

just my $0.02.


-----------------------------
Some are ignorantly happy,
While some, happily ignorant.<P ID="edit"><FONT SIZE=-1><EM>Edited by frankel on 12/04/00 08:01 PM.</EM></FONT></P>
 

yoda271828

Distinguished
Dec 31, 2007
527
0
18,980
I think were talking in circles. Your talking about overclocking the FSB with a constant multiplier setting. I'm talking about overclocking the FSB and lowering the multiplier setting.

All processors have a maximum clock speed that they can reach, but it doesn't matter if it's 100x10.0, or 133x7.5, or even 166x6.0. The same 1Ghz processor could support all of these settings. If the Motherboard and all the system components could support it, you could theoretically run it at 1Ghz FSB and a 1.0x multiplier. With the PIII-1.13Ghz the PIII core just can't reach 1.13Ghz (They are limited to 1Ghz), it has nothing to do with the FSB speed.

Athlons(T-birds) can reach 1.2Ghz possibly because <i>it does</i> have copper interconnects. And shrinking the die to a 0.13 micron process would increase the clock speed potential, but it has no effect on the bus speed.

The main problem with overclocking the FSB on Athlon motherboards is that the EV6 bus isn't very good at overclocking. However, the new AMD 760 chipset does supports a 133Mhz FSB setting.

By unlocking the multiplier on the Athlon 1.2Ghz and setting it to 9.0x, instead of 12.0x, and using the 133Mhz FSB seeing on the DDR board, the processor will not run out of spec. Since the processor is still only clocked at 1.2Ghz it <i>will</i> still work just fine.
 
G

Guest

Guest
hoho~~, so "you could <b>theoretically</b> run it at 1Ghz FSB and a 1.0x multiplier". and <b>theoretically</b> you could also run the processing unit at 2GHz FSB and a 0.5x multiplilier, right?

however, it is <b>wrong</b>, even <b>theoretically</b> speaking!

let's look into the architecture of a PIIIeb for example. it features a 64bit system bus interface running at 133MHz, through which it talks to the northbridge. this is a built-in spec. this is the actually limit. it means you can not put the system bus under too much stress. this also proves that why a PIII667eb(5x133) is way less overclockable than a PIII700e(7x100) or even a PIII800e(8x100), both featuring a much higher multiplier with a lower FSB.

theory ends here. and the fact is you can hardly find a 100FSB T-Bird or Duron of current stepping that can run stably or even post at 133FSB even if the multiplier has been downclocked period.


-----------------------------
Some are ignorantly happy,
While some, happily ignorant.
 
G

Guest

Guest
AMD currently only tests Durons & T-birds at a bus speed of 100Mhz. The processor specs only guarantee this level of performance. Certainly a small percentage of parts will work at excellerated speeds up to 133Mhz. AMD will soon be selling parts that will be spec'd @133mhz. These parts are aimed at the DDR market. Fortunately, there will only be a VERY SMALL price premium for the 133Mhz parts.

<font color=green> slice-N-dice </font color=green>
Nothing travels faster than the speed of pain
 

yoda271828

Distinguished
Dec 31, 2007
527
0
18,980
Ok, so the 1Ghz FSB is unrealiztic :) Your last post was a much better explaination. I hadn't thought of the system bus interface. However there are a few things that still didn't make sense to me. In your last post you said:

<i>let's look into the architecture of a PIIIeb for example. it features a 64bit system bus interface running at 133MHz, through which it talks to the northbridge. this is a built-in spec. this is the actually limit. it means you can not put the system bus under too much stress.</i>

If 133Mhz FSB on PIII chips is the limit, then why are so many overclocked past 150Mhz. Some are even clocked past 166Mhz. At this point the memory can't handle any more (It takes better that 6.0ns memory to run over 166Mhz).I don't think using the PIII as an example of the impossibility if FSB overclocking is valid. PIII's <b>can</b> considerably overclock their FSB. With Intel chips having locked multipliers, this is the only way to overclock them.

I realize the difficulties with overclocking the EV6 bus on the AMD chips because it is double pumped. I always thought system bus itself was the problem. Is it infact the system bus interface of the processor? Do the new 133MHZ FSB Athlons have a new stepping? I didn't think so, but they might. In an AnandTech article (<A HREF="http://www.anandtech.com/showdoc.html?i=1344&p=3" target="_new">http://www.anandtech.com/showdoc.html?i=1344&p=3</A>) about the AMD 760, Anand says this about the difference between the 100Mhz FSB and 133Mhz FSB Athlon chips:

<i>The <b>only</b> thing that sets the 1GHz and 1.2GHz CPUs apart from their 100MHz FSB counterparts is their clock multiplier. While the 100MHz FSB Athlon running at 1.2GHz has a 12.0x clock multiplier (12 x 100 = 1200MHz), the 133MHz FSB version only has a 9.0x multiplier (9 x 133 = 1200MHz). Obviously if you have a motherboard that can get around the clock multiplier lock this won’t pose a problem to you, but if not, then you will want to make sure that you only use a 133MHz FSB CPU on an AMD 760 board.</i>

This is what lead me to believe that one could use a 100Mhz chip at 133Mhz in an AMD 760 motherboard if it has an unlocked multiplier.