• 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.

Athlon 1 GHz = Athlon "C" 1 GHz?

J

joohang

Lifer
Based on what I read from somewhere in this forum that the following can be done:

=== Scenario =========

I have an Athlon 1 GHz (100 MHz DDR FSB) on MSI K7T Pro2-A.

I really want that extra FSB badly, so I get a KT133A-based motherboard.

Now, rather than selling/trading my existing Athlon, I can lower the multiplier and run it at 133 MHz DDR FSB. Since it's running at 1 GHz, it should have no problems.

=======================

I personally believe that this assumption is false. What's your input?
 
No.

In that fake scenario, I'm buying a KT133A board.

I'm wondering if my Athlon can run in the same setting as an Athlon "C".
 
Thanks for the compliment in the other thread, Joohang. Unfortunately this will far more brief than my usual posts because I should have been in bed two hours ago.

I can't, of course, address this specific example. But speaking generally, Tthere are limits on the circuits that drive the front side bus (FSB) just as there are anywhere else on the chip. So even if your memory and your motherboard and everything else on the entire system were capable of running at, say, 200MHz (not double-pumped, 200MHz as the real frequency - this is just an example), your CPU almost certainly couldn't. This would be way too far out of spec. Pin-timing tests are just as rigorous as the tests that test the frequency of the rest of the chip.

So essentially if you run a chip that's marked as a 100MHz FSB chip and change the multiplier so that it runs at the same frequency at 133MHz FSB, then you are still overclocking your chip because you are now running the circuits that drive the FSB at a frequency higher than the spec.

I'm not saying that it won't work. I'm just stating that it's not a valid assumption to assume that it will always work to increase the FSB and then change the multiplier so that the overall frequency remains constant.

Ok, well, that was a lot longer than I thought it would be... man, I just can't help typing too much. Ok, now I'm really going to bed. Good night.
 
It should work if your mobo allows use of lower multipliers. Some people using the MSI K7T Turbo are having problems. Also, if you need to use a jumper to go to 133 and above, this may prevent you from being able to boot as the mobo may use your default multiplierx133MHz FSB for a very short period before bios implements the lower multiplier.
 
Thank you for your insight.

I plan to upgrade to a dual Athlon setup with the Tyan AMD760-based motherboard soon. I was hoping to run 2x Athlon "C" 1 GHz + PC2100 SDRAM, so I was wondering if I could save some $$$ and trouble by using my existing Athlon.

I guess I'll just buy myself 2x Palomino instead. 🙂
 
Someone once said, "Those who say it cannot be done, should not interrupt those of us that are doing it."

I have been running 1.2GHz Tbirds at 1.5GHz ~ 10 X 150MHz fsb. That is 300MHz DDR, double pumped. 🙂

I've been doing the same with Durons and many other Tbirds. Not bad for a CPU designed for the 100MHz fsb, or 200 double pumped. I have serious questions about the motives and facts behind the 266 chips. Frankly, I think the core has been fabbed that way all along and they are going to use it for a money tool with the ol relabel trick. :Q
 
compuwiz1,

I get your point.

However, that's overclocking and I refuse to clock any of my CPU's out of the spec. Since I have to overclock the FSB to run my Athlon at the Athlon "C" level, I'm just going to buy 2 new CPU's instead. 🙂
 
joohang,

If you lower the multiplier appropriately, you can get the speed increase of the 133MHz FSB without actually overclocking your processors.
 
dennilfloss,

If I do that, as pm mentioned, the CPU will not be overclocked but the FSB will be. So my CPU will be running at a FSB that is higher than its spec. Then that's still considered overclocking, right?
 
One more question:

If I increase the FSB to 133 MHz, does that only stress the CPU? What other computer components become stressed by overclocking the FSB?
 
Nothing in system is overclocked at 133MHz fsb, since all buses are at spec speed.

Now, my point is, are the 100MHz "spec'd" cpus really running out of spec at 133MHz fsb? I do not believe so. Do you really believe there is a fundamental difference in the core and cache on the P3 700, vs. the P3 733? I'll bet not.
Where the factory will lock the multiplier is the only difference I see.
My bet is the core is the same.

BTW, I am more than a hobbyist or casual observer. 😉
 
As pm mentioned, nobody is saying that it can't be done. However, he is very correct when stating that running the cpu at a higher fsb is still running some of it's circuits out of spec. Therefore, you are overclocking it.

Don't get me wrong, I'm not implying that there's anything wrong with you overclocking your cpu if you so desire. It's your cpu, you can do whatever you want with it.
 
For those of you who are saying that I'm wrong, think of it this way:

if you had a magic motherboard that let you increase the FSB in increments of 1MHz and this magic motherboard could somehow keep everything running at spec. (the PCI bus and the memory bus), and you could somehow dial an infinite variety of multipliers such that the total frequency of the CPU was always constant - say 1GHz - how fast do you think you could increase the FSB before the CPU would fail? At what point when you are turning this dial would you say that you are overclocking the FSB? 150MHz? 200MHz? If the manufacturer specifies that a CPU should be run with a 100MHz, why is it not running it out of spec to run it at 133MHz, any more than it would be to run it at 200MHz?

You guys seem to be saying that the circuitry that drives the pins of a CPU can simply run at whatever speed - that they have no response time to inputs and can always have the data setup for outputs. But I'm saying this is incorrect, the circuits on the FSB are governed by the same rules as the rest of the chip. If some chips that come out of fab can run at 1GHz and others no faster than 700MHz, then why doesn't it make sense that some can run at 133MHz FSB while others can only run at 100MHz? The circuits that drive the FSB are subject to the same PVT problems that all of the other circuits on the chip are.

And I didn't say, "this won't work ever" - I said, "it's not valid to assume it will always work with every CPU". I can assure you that some, maybe an insignificant minority, but still some, will fail.
 
I think that the point that compuwiz was trying to make though was that the parts aren't really running out of spec at 133/266, that they've been capable of that all along.
 
And I'm saying, no, they aren't all capable of it. Maybe most of them are, but not all - which was what the original question is.

If you ask is "a = b" and the answer is most of the time, but not always, then "a != b".

So the question is, am I right, are there a few processors that can run at 100MHz FSB, but not at 133MHz regardless of the internal CPU speed, and I am certain that there are. It is that it is a lot harder to get 133MHz frequencies on a motherboard than you might think. A 133MHz FSB may not sound like much when you are talking 1GHz CPU's, but CMOS devices don't source a lot of current and so it is very hard to drive the huge capacitances of off-chip buses. The timing numbers for 133MHz are not a slam-dunk. I would guess that most of the CPU's work at 133MHz, but I'm sure that not all of them do.


 
I'm pretty convinced that my CPU will work at 133 MHz, but I'm very reluctant towards overclocking for the following reasons:

- Stability
- Stress on the CPU for running at out-of-spec FSB
(unless the Athlon "C" is simply the Athlon with lower multiplier and 133 MHz FSB, as manufactured by AMD, which I doubt it is.)
- Heat issues: I'm using a Molex quiet fan. My Athlon is already running at 50+°C
- Risk vs. Performance increase: the performance increase is quite small

If you agree/disagree to any of my points, please fire away your opinions. 🙂
 


<< For those of you who are saying that I'm wrong, think of it this way >>



pm, I am definitely not speaking about Intel. 😉

I do not think you are wrong. I used Intel as a reference, but my opinion is mostly being expressed about AMD.
 
From what i have read, almost all tbird/durons will run with the 133 mhz bus. And the C athelons are just ones that AMD tested to work at 133 FSB and stamped a C instead of a B on the core. A 1ghz C and a 1ghz B with the mult lowerd running at 133 should produse the same amount of heat. Also, the proformance increase from the higher FSB is said to be fairly significant.
-Az
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top