Sign in with
Sign up | Sign in
Your question

Tbred 1.8GHz potential overclocking

Last response: in Overclocking
Share
April 10, 2002 4:34:16 AM

I propose that one of the upcoming Tbred 1.8GHz chips, overclocked and running at 150MHz fsb (2025MHz) with DDR memory running asynchronously, will be faster than a 2.8GHz P4.

Any thoughts? And the beauty is that many AMD motherboard owners will be able to plug one of these bad boys into their existing boards.

The number of paychecks I'm owed by my .dot com employer: 6 (paid twice/month).
April 10, 2002 4:48:34 AM

2.025GHz Tbred faster than a 2.8GHz P4? Depends on the FSB of the P4, but I doubt it.

<font color=blue>If you don't buy Windows, then the terrorists have already won!</font color=blue> - Microsoft
April 10, 2002 11:25:10 AM

Why overclocked to just over 2GHz? If the TBred is as overclockable as is being rumoured, you could probably go higher and get a 166MHz FSB or faster. However, an 800MHz difference is too big. The P4 will probably be faster.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
Related resources
Anonymous
a b à CPUs
a b K Overclocking
April 10, 2002 6:03:59 PM

Check the article here on THG about "The Final Battle: P4/2400 vs. Athlon XP 2100+ (2400+)" And check the benchmarks - you can see that an Athlon XP at 1918 mhz (with 184 FSB) can beat the P4 (400mhz) at alot of stuff - but the 533 fsb p4 (which will be the next batch from what I understand) beat it at almost everything. the fastest one of those is only at 2666. So a 2800 mhz P4 would kill a 2000 mhz Athlon XP. What I am interested in is how high will the t-bred's go? The Athlon XP release was the 1500+, 1600+ and 1700+, right? So we might could expect a 2400+ (at 1933) which we could probably get to 2250 at least. Then maybe it'd be a little closer. But for now it looks like Intel can ramp up the Mhz faster than AMD can. Perhaps not though, I guess we shall have to wait and see...
April 11, 2002 1:18:46 AM

well in 'simple' terms from everything I've gathered from various sites the chips have different needs. AMD has more or less been able to tweak their CPU to run extremely well at current FSB speeds(100, 133, 200, 266) to the point of where Intel as much faster clock speeds can't compete. Now with Intel being able to ramp up the FSB to 133*4(533 Mhz effective) vs the 100*4(400 Mhz FSB) they currently use they're finally able to feed the P4 with enough data to overtake current AMD proc's.

The real issue is will the T-bred have enough software optimization or future FSB speeds to compensate for the raw bus speed intel has on tap. AMD will need to ramp up the core speeds or memory speeds to match and/or beat once again. The 600 Mhz is a big issue with the new faster memory timings.

I think AMD will answer as they have for the last 5 years. In all honesty though, I'd still rather spend $235 for an XP 2100 vs a 2.4 P4 @$560(pricewatch.com current prices). For double the price you're only getting a minimal performance boost. Time will tell who will come out victorius in the 'next round'. All in all it's great for us though, my XP1700(OC'd to 1900+) is now under $100. Prices continue to fall as CPU's get faster and faster. In 6 months I'll give this old rig to one of my daughters and ramp up to 2+Ghz core speed with AMD and all for probably under $200 for the CPU. Things couldn't be better!!!!!!!!

:cool: Save heating costs on your home, overclock your PC!!! :cool:
a b à CPUs
a b K Overclocking
April 11, 2002 4:20:06 AM

Remember that the P4 will be running at 133 instead of 100, so it will probably be around the performance of a Tbred at 2400MHz. And if it gets paired with the upcomming SiS Dual Channel DDR chipset it COULD be around the performance of Tbred at 2600MHz or more.

What's the frequency, Kenneth?
a b à CPUs
a b K Overclocking
April 11, 2002 4:23:13 AM

Remember that the P4 will be running at 133 instead of 100, so it will probably be around the performance of a Tbred at 2400MHz. And if it gets paired with the upcomming SiS Dual Channel DDR chipset it COULD be around the performance of Tbred at 2600MHz or more.

It's really hard to estimate though, so you can add or subtract a couple hundred MHz on those numbers to suit your taste, until the parts are on the market.

What's the frequency, Kenneth?
April 11, 2002 4:49:45 AM

I tend to disagree, as the Tbred will be overclocked and running at a 150MHz or faster fsb, so it too will perform faster than a native 2400 or 2600MHz Tbred on a 133MHz fsb.

The number of paychecks I'm owed by my .dot com employer: 6 (paid twice/month).
April 11, 2002 4:55:30 AM

Why overclock to just over 2GHz? Just being conservative. You're right. A Tbred at 2.4-2.6GHz would likely anhiliate a P4, even on a 133MHz fsb, at 3.2GHz or faster.

The number of paychecks I'm owed by my .dot com employer: 6 (paid twice/month).
April 11, 2002 4:57:43 AM

Well that would be something to see, but based on current P4 performance, a Tbred at 2.6GHz would anhiliate even a 3.2GHz P4. Shoot, even with the 133MHz fsb.

The number of paychecks I'm owed by my .dot com employer: 6 (paid twice/month).
April 11, 2002 5:01:36 AM

Yeah, that is an interesting article, but too bad they're comparing a heavily overclocked P4 with a marginally overclocked XP.

To be fair, if we compared a Tbred overclocked to a 150 or 166MHz fsb (300/333), I think it would be a fairer comparison to the upcoming 133 fsb Northwoods.

God, could you imagine a Tbred running at 2.4GHz on a 166MHz overclocked fsb (333)??? It would probably beat a 3.5GHz P4.

Check the article here on THG about "The Final Battle: P4/2400 vs. Athlon XP 2100+ (2400+)" And check the benchmarks - you can see that an Athlon XP at 1918 mhz (with 184 FSB) can beat the P4 (400mhz) at alot of stuff - but the 533 fsb p4 (which will be the next batch from what I understand) beat it at almost everything. the fastest one of those is only at 2666. So a 2800 mhz P4 would kill a 2000 mhz Athlon XP. What I am interested in is how high will the t-bred's go? The Athlon XP release was the 1500+, 1600+ and 1700+, right? So we might could expect a 2400+ (at 1933) which we could probably get to 2250 at least. Then maybe it'd be a little closer. But for now it looks like Intel can ramp up the Mhz faster than AMD can. Perhaps not though, I guess we shall have to wait and see...


The number of paychecks I'm owed by my .dot com employer: 6 (paid twice/month).
a b à CPUs
a b K Overclocking
April 11, 2002 5:48:09 AM

It's all speculation. Right now it takes an XP about 2000MHz to beet a P4 2400 which has a 100MHz FSB. We know that going to a 133MHz FSB is worth a good increase on the P4, probably on the order of 10%, I'm sure someone will point to the benchmarks. The P4 2.8 will have that advantage, as everything from the 2.5 up is supposed to. We also know that going from dual channel RDRAM to dual channel DDR SDRAM will improve performance by reducing latency. I'm guessing another 10%, but that's only an educated guess. So we're talking about an estimated 20% increase in performance.

OK, so now compare those new advantages to currently available speeds: Imagine the P4 2.4 had these features, that would mean the Athlon would need to be 20% faster than 2000MHz, which means it would have to be 2400MHz. That's PAR!

But let's say those estimates are too generous, and that each enhancement will only add 5% performance to the P4, for a total 10% increase. That's fairly conservative, don't you think? That would mean the Athlon would have to run 10% faster to match the P4 2.4 with both enhancements, as in 2200MHz.

Now fast forward to the real deal, the P4 at 2.8GHz would be more than 15% faster than the P4 2.4 just by clock speed. Plus it will be available with Dual Channel DDR and at 133MHz FSB, adding at least 10%, probably 20%. Compare that to the Athlon 2000MHz. Make it 15% faster to match the clock increase, now it's 2300MHz. It takes 2300MHz on the Athlon to match the P4 2.8 with a 100MHz bus and RDRAM. Ouch. Now add the 10%-20% performance gain I mentioned for using the 133MHz bus speed and dual channel DDR. 2300+10%~2533. 2300+20%~2666.

So it will take a stock Athlon of at least 2533MHz to 2666MHz to match a P4 2800 with Dual Channel DDR SDRAM. Funny thing about the XP+ rating, those processors would be rated at what, 3400+? LMAO@XP ratings!

OK, I'll give you 10% back for overclocking the AMD, giving it faster bus speed. That's still ~2300MHz to match an Atlon with an overclocked bus to a P4 with a stock bus. Nice.

Look I like AMD a lot. I had my board all picked out (the Leadtek Winfast 7350KDA) and everything for my new system. But AMD didn't release the Thoroughbred on time, so it's their fault. Intel's biggest mistakes on releasing the P4 were it's reliance on Dual Channel RDRAM instead of Dual Channel DDR SDRAM, and releasing the Willy in a crippled state. But those mistakes are now being addressed. Sometimes you have to step into the dark to see the light!

Before you go blabbing garbage, think about just HOW LEAN that 10% performance improvement on the Northwood "B" mounted to a Dual Channel DDR board really is. Were talking a VERY conservative estimate. I fully expect the gains to be on the order of 20% or more.

Now I can only fault Intel for jerking customers around with the Willy and Rambus, and for using an abusive pricing scheme. AMD will probably not overtake them again until 64-bit programs for the Hammer are popular.

What's the frequency, Kenneth?
April 11, 2002 2:52:34 PM

You made some fair and reasonable points, first and foremost of which is that this is all speculation, as there aren't either 133MHz fsb P4s (unoverclocked, that is) or mobos that support dual channel DDR and P4s at the moment.

But when you said, "Before you go blabbing garbage..." I thought you were about to blow a blood vessel, or cry. No worries, man, we'll see soon enough. Intel should be ready with at least one of its proposed offerings by the time Tbred hits the market. That's when we'll see what happens.

-kermit4karate


---yous stuff---

It's all speculation. Right now it takes an XP about 2000MHz to beet a P4 2400 which has a 100MHz FSB. We know that going to a 133MHz FSB is worth a good increase on the P4, probably on the order of 10%, I'm sure someone will point to the benchmarks. The P4 2.8 will have that advantage, as everything from the 2.5 up is supposed to. We also know that going from dual channel RDRAM to dual channel DDR SDRAM will improve performance by reducing latency. I'm guessing another 10%, but that's only an educated guess. So we're talking about an estimated 20% increase in performance. OK, so now compare currently available speeds: Imagine the P4 2.4 had these features, that would mean the Athlon would need to be 20% faster than 2000MHz, which means it would have to be 2400MHz. That's PAR! But let's say those estimates are too generous, and that each enhancement will only add 5% performance to the P4, for a total 10% increase. That's fairly conservative, don't you think? That would mean the Athlon would have to run 10% faster to match the P4 2.4 with both enhancements, as in 2200MHz.

Now fast forward to the real deal, the P4 at 2.8GHz would be more than 15% faster than the P4 2.4 just by clock speed. Plus it will be available with Dual Channel DDR and at 133MHz FSB, adding at least 10%, probably 20%. Compare that to the Athlon 2000MHz. Make it 15% faster to match the clock increase, now it's 2300MHz. It takes 2300MHz on the Athlon to match the P4 2.8 with a 100MHz bus and RDRAM. Ouch. Now add the 10%-20% performance gain I mentioned for using the 133MHz bus speed and dual channel DDR. 2300+10%~2533. 2300+20%~2666.

So it will take a stock Athlon of at least 2533MHz to 2666MHz to match a P4 2800 with Dual Channel DDR SDRAM. Funny thing about the XP+ rating, those processors would be rated at what, 3400+? LMAO@XP ratings!

OK, I'll give you 10% back for overclocking the AMD, giving it faster bus speed. That's still ~2300MHz to match an Atlon with an overclocked bus to a P4 with a stock bus. Nice.

Look I like AMD a lot. I had my board all picked out (the Leadtek Winfast 7350KDA) and everything for my new system. But AMD didn't release the Thoroughbred on time, so it's their fault. Intel's biggest mistakes on releasing the P4 were it's reliance on Dual Channel RDRAM instead of Dual Channel DDR SDRAM, and releasing the Willy in a crippled state. But those mistakes are now being addressed. Sometimes you have to step into the dark to see the light!

Before you go blabbing garbage, think about just HOW LEAN that 10% performance improvement on the Northwood "B" mounted to a Dual Channel DDR board really is. Were talking a VERY conservative estimate. I fully expect the gains to be on the order of 20% or more.

Now I can only fault Intel for jerking customers around with the Willy and Rambus, and for using an abusive pricing scheme. AMD will probably not overtake them again until 64-bit programs for the Hammer are popular.

What's the frequency, Kenneth?


The number of paychecks I'm owed by my .dot com employer: 6 (paid twice/month).
a b à CPUs
a b K Overclocking
April 11, 2002 3:24:31 PM

Excuse the "blabbing garbage" statement, but I have not debated with you before, and it's often the case that certain members, confounded by technical data, resort to such devices as name calling.

We see the P4 1.8 at 2.4/PC1066 significantly outperforming the stock 2.4/PC800. So that's one performance improvement that will be implemented unchanged on production processors.

We see the P4 with DDR333 matching the performance of PC800, even though its bandwidth is lower, most likely do to lower latency of SDRAM. So Dual Channel DDR is almost garunteed to offer some improvement.

Those things are facts, the rest is speculation. I honestly do not believe AMD has the nerve to maintain it's XP+ rating system in its current form, in light of these facts. What I see is the clock-for-clock performance margin between these two processors closing, and AMD is looking pretty sad for not releasing the Tbred sooner, when it would have been more competitive. Two weeks ago would have done the trick, things move fast in this industry.

I never cared much for Intel as a company, mainly because their products were so overpriced as to be a true rip-off. Look at the P4-XEON as a perfect example, or the current P4 2400 pricing as further proof.

I used to tell everybody and his brother to go AMD, just because I liked the competition between the two companies. At that time AMD was doing everything they could to survive, instead of resting on their laurels. I'm very disappointed in them taking their latest position, which is that they are waiting for Intel to take their next step before AMD takes theirs. AMD stongly implied that it is their position to stay one step behind Intel for the forseable future, so as to avoid the price war which lead to the current market glut of PC's. I look at this as taking the "followers stance", which to me means "looser mentality" and "screw the faithfull, lets milk the cow". Did I say I was very disappointed yet?


Crashman = Disillusioned AMD fan.
Anonymous
a b à CPUs
a b K Overclocking
April 11, 2002 4:25:28 PM

So the dual-channel DDR nForce chipset didn't do a whole lot for performance because it didn't give the CPU dual-channel's (just made more memory availible to everything else). This is a chipset issue, yes? So perhaps if we get an AMD-based dual-channel DDR chipset with DDR333 or DDR400 the t-bred could catch back up?
a b à CPUs
a b K Overclocking
April 11, 2002 10:40:24 PM

The Athlon derivatives use a DDR266/DDR200 CPU bus, which exactly matches DDR266/DDR200 SDRAM, aka PC2100 and PC1600. So in order for a dual channel DDR chipset to do it much good, you would have to double the CPU bus frequency to DDR533/400. So it's a processor limmitation.

The P4 uses a QDR400 bus, which is twice the bandwidth of DDR200 (PC1600) memory. So the best option would be to double the bandwidth of the memory by doubling the buswidth, using Dual Channel. RDRAM motherboards use Dual Channels for similar reasons. Future P4's and overclocked current models at 133MHz CPU bus (QDR533) will evenly match Dual Channel PC2100.

What's the frequency, Kenneth?
Anonymous
a b à CPUs
a b K Overclocking
April 11, 2002 11:14:33 PM

Ah... it all makes alot more sense now.
!