AMD 2600+ REVIEW Still not #1!

[alexruiz]

To all those die-hard no-reasoning fanboys, go and read JC's comments. He is stressing the point of benchmark selection, and how sometimes even in the same benchmark the results are different (good point about xbitlabs review. They were the only ones to perform 2 different tests using lightwave, the "rendering" and the "sunset") All the other sites focused only in the "rendering" parts and declared the P4 winner..... however, the Athlon beats the P4 in sunset very badly.

Go read JC's comments! JC's site

 

[CrazySaint]

Originally posted by: SSXeon5

Originally posted by: CrazySaint

Originally posted by: SSXeon5

Originally posted by: Boonesmi
my point is either you deal with officially supported for both or you dont for both

whining about the anantech article not using pc 1066 with the P4 is just silly since its not officially supported

You know your just saying that because wile anandtech handicaped (CrazySaint this time i can say it) the 533Mhz fsb with 800Mhz Ram, and when tom didnt handicap it, the P4 2.53GHz wooped up on the 2600+

SSXeon

No you can't Near as I can tell, PC800 and PC2700DDR are nearly dead even, with the PC2700 having a slight lead. That's not handicapping. Second, running an XP with 3:4 mem ratio at 133FSB gives it a whopping 2% performance increase over PC2100.

I've already stated my opinion of Tom's review. Is it possible that the reason you enjoyed Tom's review and not Anand's is because Tom's review showed the P4 2.53GHz slightly (no whooping going on) more favorably than the 2600+ whereas Anand's review didn't? As near as I can tell, the 2600+ and the 2.53GHz are pretty much dead even.

No I liked toms because he used the top of the line RDRAM (PC1066) and PC800 to compare to the new XP. Anandtech did not, and because it is not offically supported is bullsh*t. So what if they used PC1600 with the XP? And the P4 had PC1066? Would that be fair? No sir, and the way you all act that anandtechs was better because they used PC800 instead of PC1066 or even DDR333 on the BG7/P4B533-V.

SSXeon

PC2700 on the XP and PC800 on the P4 is hardly comparable to PC1600 on the XP and PC1066 on the P4. Would you have been happy if they had used PC2100 on the XP instead of PC2700? If they had, it wouldn't have changed the results as PC2700 on an XP is negligibly faster than PC2100, and Anand speculates that it has more to do with the KT333's chipset being optimized for PC2700 instead of PC2100 than the XP actually using any of the extra bandwidth. Even PC1066 wouldn't have made much difference as, IIRC, its only about 2-8% (averaging around 3-5%) faster than PC800. DDR333 on a BG7/P4B533-V certainly wouldn't have made any difference.

 

[jiffylube1024]

Why am I up this late? Well, it's a long story that I don't care to tell.

Here's what I have gotten out of this thread so far: a lot of arguing over not too much.

Anyways, here's my take on this: there's a lot of complaining over benchmarks and one platform being handicapped versus another, and on an on...

SSXeon5 earlier wanted to know why everyone was saying how this was so good for AMD, seeing as they released a new revision of CPUs which are only now as fast as the current best offering from Intel, or perhaps not quite as good if "not held back," or perhaps slightly better. The point is this: the reason AMD is getting so much props and it *seems* like everyone is jumping on their bandwagon is because what seemed like a dead revision (the Thoroughbred Athlon) is actually quite vital with this new revision, and most importantly AMD is still in the top-of-the-line ballgame. Maybe they aren't the "absolute" fastest, but they're right back up there again. This is great for competition (as we all know), and keeps our CPU prices low.

It's great to stuff like "I have a P4. It is the FASTEST. You have no argument that can beat me, especially with XXXXX on the horizon to really seal the deal", but with AMD keeping pace it's better for all of us in the long run (heck even the short term).

So basically to close this off, this thread began with the petty claim that AMD isn't #1 (no disrespect intended SSXeon), and the point I'm trying to drive home is that it doesn't matter if they're not "truly" #1. AMD may not have the fastest CPU in all of the benchmarks, or even more than 50% of them, but they're right at the top again. From an objective viewpoint, the 2600+ and 2.53 GHz are very close in performance; neck and neck even. Intel has a 2.8 GHz CPU coming out very soon; so what. AMD is still breathing, their future line of CPU's are still coming (as are Intel's) and there's still competition left in the CPU market.

 

[SSXeon5]

Originally posted by: CrazySaint

PC2700 on the XP and PC800 on the P4 is hardly comparable to PC1600 on the XP and PC1066 on the P4. Would you have been happy if they had used PC2100 on the XP instead of PC2700? If they had, it wouldn't have changed the results as PC2700 on an XP is negligibly faster than PC2100, and Anand speculates that it has more to do with the KT333's chipset being optimized for PC2700 instead of PC2100 than the XP actually using any of the extra bandwidth. Even PC1066 wouldn't have made much difference as, IIRC, its only about 2-8% (averaging around 3-5%) faster than PC800. DDR333 on a BG7/P4B533-V certainly wouldn't have made any difference.

Dude the pentium 4 with 533Mhz fsb has 4.2GB/s, if you uses PC800 it does Not get its full 4.2gb/s bandwidth right? So giving it 3.2GB/s PC-800 is handicaping it from its performance. As for you assuption of PC2100 instead of PC2700 for the XP ... do you really think im that stupid? Did i say PC2100? No i said PC1600 (DDR200 if you didnt know) given that it would be the same comparing a P4 w/ 533 FSB with PC800 as to a XP w/ 266Mhz fsb and PC1600. As for you saying the I845Gs wouldnt have made any difference? The Asus BG7/P4B533-V are averaged 2-4% slower then PC1066 in most applications as for PC800, it is averaged 4-6% slower then PC1066 in most applications, basing this off toms reviews. And the i845G's would make a huge difference matching the 2600+ in performance and being lower priced then the AMD setup. But like I said DDR does make a difference, in performance and in price. There is two setups for the P4, proformance/ lowpriced (i845G/PC2700), max proformance/ highpriced (i850E/PC1066). Just wait, in 3 days all hell will break lose

Originally posted by: jiffylube1024
Why am I up this late? Well, it's a long story that I don't care to tell.

Here's what I have gotten out of this thread so far: a lot of arguing over not too much.

Anyways, here's my take on this: there's a lot of complaining over benchmarks and one platform being handicapped versus another, and on an on...

SSXeon5 earlier wanted to know why everyone was saying how this was so good for AMD, seeing as they released a new revision of CPUs which are only now as fast as the current best offering from Intel, or perhaps not quite as good if "not held back," or perhaps slightly better. The point is this: the reason AMD is getting so much props and it *seems* like everyone is jumping on their bandwagon is because what seemed like a dead revision (the Thoroughbred Athlon) is actually quite vital with this new revision, and most importantly AMD is still in the top-of-the-line ballgame. Maybe they aren't the "absolute" fastest, but they're right back up there again. This is great for competition (as we all know), and keeps our CPU prices low.

It's great to stuff like "I have a P4. It is the FASTEST. You have no argument that can beat me, especially with XXXXX on the horizon to really seal the deal", but with AMD keeping pace it's better for all of us in the long run (heck even the short term).

So basically to close this off, this thread began with the petty claim that AMD isn't #1 (no disrespect intended SSXeon), and the point I'm trying to drive home is that it doesn't matter if they're not "truly" #1. AMD may not have the fastest CPU in all of the benchmarks, or even more than 50% of them, but they're right at the top again. From an objective viewpoint, the 2600+ and 2.53 GHz are very close in performance; neck and neck even. Intel has a 2.8 GHz CPU coming out very soon; so what. AMD is still breathing, their future line of CPU's are still coming (as are Intel's) and there's still competition left in the CPU market.

Hmm I can see that and you brought up a great point. I should give amd props for getting there CPU up to speeds to compete with intel:

AMD Props for getting to 2.13GHz on the TBred core!

But the prob is still that people will have to wait about 2-3weeks to get a 2600+ when the 2.8GHz/2.66GHz will be here in 3 days. But I see amd had to do the paperlaunch to stay in the race and I would have done the same too

SSXeon

 

[DivideBYZero]

Originally posted by: jiffylube1024
Why am I up this late? Well, it's a long story that I don't care to tell.

Here's what I have gotten out of this thread so far: a lot of arguing over not too much.

Anyways, here's my take on this: there's a lot of complaining over benchmarks and one platform being handicapped versus another, and on an on...

SSXeon5 earlier wanted to know why everyone was saying how this was so good for AMD, seeing as they released a new revision of CPUs which are only now as fast as the current best offering from Intel, or perhaps not quite as good if "not held back," or perhaps slightly better. The point is this: the reason AMD is getting so much props and it *seems* like everyone is jumping on their bandwagon is because what seemed like a dead revision (the Thoroughbred Athlon) is actually quite vital with this new revision, and most importantly AMD is still in the top-of-the-line ballgame. Maybe they aren't the "absolute" fastest, but they're right back up there again. This is great for competition (as we all know), and keeps our CPU prices low.

It's great to stuff like "I have a P4. It is the FASTEST. You have no argument that can beat me, especially with XXXXX on the horizon to really seal the deal", but with AMD keeping pace it's better for all of us in the long run (heck even the short term).

So basically to close this off, this thread began with the petty claim that AMD isn't #1 (no disrespect intended SSXeon), and the point I'm trying to drive home is that it doesn't matter if they're not "truly" #1. AMD may not have the fastest CPU in all of the benchmarks, or even more than 50% of them, but they're right at the top again. From an objective viewpoint, the 2600+ and 2.53 GHz are very close in performance; neck and neck even. Intel has a 2.8 GHz CPU coming out very soon; so what. AMD is still breathing, their future line of CPU's are still coming (as are Intel's) and there's still competition left in the CPU market.

I'll second that.

 

[alexruiz]

Amen
R.I.P.

 

[spanky]

it won some benchies... it lost some. in that case... u should also change ur thread title to "neither amd nor intel is #1".

 

[SupermanCK]

Originally posted by: spankyOO7
it won some benchies... it lost some. in that case... u should also change ur thread title to "neither amd nor intel is #1".

i agree....or both INTEL AND AMD ARE #1

 

[rIpTOr]

I think Anandtech did a good review even though they used PC800. RDRAM is going to die off anyway so you'd better get used to DDR benchmarks in the future.SSXeon looks like you might want to wait a bit longer for your P4. The P4 3GHz will require a new motherboard. (Unless you want to buy a 2.53~2.8 and have to upgrade again)
http://zdnet.com.com/2100-1103-954842.html

As far as I'm concerned a P4 and Athlon XP are tied. The real test will be the Hammer vs the Prescott. Then the K9 vs the ???

 

[CrazySaint]

Originally posted by: SSXeon5

Originally posted by: CrazySaint

PC2700 on the XP and PC800 on the P4 is hardly comparable to PC1600 on the XP and PC1066 on the P4. Would you have been happy if they had used PC2100 on the XP instead of PC2700? If they had, it wouldn't have changed the results as PC2700 on an XP is negligibly faster than PC2100, and Anand speculates that it has more to do with the KT333's chipset being optimized for PC2700 instead of PC2100 than the XP actually using any of the extra bandwidth. Even PC1066 wouldn't have made much difference as, IIRC, its only about 2-8% (averaging around 3-5%) faster than PC800. DDR333 on a BG7/P4B533-V certainly wouldn't have made any difference.

Dude the pentium 4 with 533Mhz fsb has 4.2GB/s, if you uses PC800 it does Not get its full 4.2gb/s bandwidth right? So giving it 3.2GB/s PC-800 is handicaping it from its performance. As for you assuption of PC2100 instead of PC2700 for the XP ... do you really think im that stupid? Did i say PC2100? No i said PC1600 (DDR200 if you didnt know) given that it would be the same comparing a P4 w/ 533 FSB with PC800 as to a XP w/ 266Mhz fsb and PC1600 As for you saying the I845Gs wouldnt have made any difference? The Asus BG7/P4B533-V are averaged 2-4% slower then PC1066 in most applications as for PC800, it is averaged 4-6% slower then PC1066 in most applications, basing this off toms reviews. And the i845G's would make a huge difference matching the 2600+ in performance and being lower priced then the AMD setup. But like I said DDR does make a difference, in performance and in price. There is two setups for the P4, proformance/ lowpriced (i845G/PC2700), max proformance/ highpriced (i850E/PC1066). Just wait, in 3 days all hell will break lose

Eh, so you're complaining about Anand's article because it shows the P4 being 2% slower than it could have been?! 2% only matters for people who put a great deal of stock in benchmarks, 'cuz no way in hell is anybody going to notice the difference between 2%. And if the 2600+ and 2.53GHz are close enough that 2% would make a difference in which one is faster, then they are very close.

 

[statix]

Originally posted by: SupermanCK

Originally posted by: spankyOO7
it won some benchies... it lost some. in that case... u should also change ur thread title to "neither amd nor intel is #1".

i agree....or both INTEL AND AMD ARE #1

Amen. Guys, you should be proud of AMD and Intel, they are America's two best technology leader companies in the world.

Long live AMD and Intel!

 

[mtgcollector01]

Blah blah blah...intel is better than AMD, AMD is better than Intel, give it a rest. Only about .0000001% of the world's population "NEEDS" this difference in peformance. and just a few more than that care enough or are fanboys enough to argue about it.

Anyways, it's just gonna be a seesaw battle until hammer comes out or whatnot...AMD releases a new processor which is faster, then Intel releases one that is faster...and we're just gonna be wasting 8 hours a day doing flame wars.

BTW, I have an XP 1800+ but i wouldn't mind a p4 1.8GHz...all in all it's not different enough to matter!

 

[Wingznut]

Originally posted by: mtgcollector01
Blah blah blah...intel is better than AMD, AMD is better than Intel, give it a rest.

Actually, it looks like this thread was given "a rest" two days ago.

So, why'd you bring it back up?

 

[mtgcollector01]

Don't shoot me! It was in the "hot threads" sectioN!

 

[pRoNe]

Originally posted by: rIpTOr
I think Anandtech did a good review even though they used PC800. RDRAM is going to die off anyway so you'd better get used to DDR benchmarks in the future.SSXeon looks like you might want to wait a bit longer for your P4. The P4 3GHz will require a new motherboard. (Unless you want to buy a 2.53~2.8 and have to upgrade again)
http://zdnet.com.com/2100-1103-954842.html

As far as I'm concerned a P4 and Athlon XP are tied. The real test will be the Hammer vs the Prescott. Then the K9 vs the ???

I read this and must mention only what is currently being reviewed on many sites. RD RAM is being revised to a more cost effective consumer ram called rimm, in the form of rimm4200 (800mhz 3.2gb/s throughput) and rimm3200 (1066mhz 4.2gb/s throughput).
This revision cancels the need to have two identical 16bit rdram modules, insted using a single 32bit rimm module.


Also must reply to my earlier statement, with the launch of the new p4 2.gghz with 533fsb pentium has most definatively taken #1 pushing AMD back to second without even a contest.
Even if the p4 beats the xp2600 by only a few frames/second or a few points in a benchmark, it still wins and thats that.
To be number one I feel as long as a cpu can win over 50% of the benchmarks it is #1.
I know I will be corrected if people think otherwise.

Curious thought please post back similar thoughts or criticism: AMD has 266 fsb, if a chipset maker made a rdram or rimm chipset that could work with an amd xp cpu would amd still be #2.
Imagine an xp2600 with 1066mhz(3200 rimm) ram that would be 4 x a 266fsb.
could this work???

 

[CrazySaint]

Originally posted by: pRoNe

Originally posted by: rIpTOr
Curious thought please post back similar thoughts or criticism: AMD has 266 fsb, if a chipset maker made a rdram or rimm chipset that could work with an amd xp cpu would amd still be #2.
Imagine an xp2600 with 1066mhz(3200 rimm) ram that would be 4 x a 266fsb.
could this work???

The XP wouldn't be able to use the bandwidth. The most it can use is 2.1GB/s memory bandwidth.

 

[Duvie]

crazysaint has it right....What would be the point!!! even at 333fsb and theoretical 2.7gb/s of bandwidth it almost wouldn't be enough. Then remember the premium cost many pay for rdram mobos now and the actually memory chips...AMD is value and budget leader for the opposite of those things...

 

[rIpTOr]

Your right... thats one of the main reasons Intel is beating AMD now... Pure bandwidth.

 

[Acanthus]

RDRAM will be in trouble when DCDDR comes out, unless they ramp up to PC1200 and PC1333 then or dual channel 32-bit, but i dont think will happen until the eventual move to 667 and 800fsb. Which is still a long ways off.

Granite Bay, SiS655, and P4X600 are going to be very popular (especially if one of them decides to go with 333/400 support instead of 266).

Ill have to say because im an overclocker ill wait to see which of the 3 chipset is the most overclockable before i make my decision, as FSB will be a big factor in DCDDR mobos. I believe GB is a workstation mobo, it may not have many overclocking options. I bet SiS and VIA will be though.

 

[Ninaholic]

An effective throughput of 2.1GB/s is still nothing to sneeze at.

Certainly, there is agreement on the fact that the XP is not a workstation platform, simply because its lower memory bandwidth can become saturated while undertaking tasks that require high levels of memory access.

For the rest of us, however, the issue is moot. Unless you constantly use programs that require high memory bandwidth, the real-world differences in performance are fairly negligable.

The first principle of benchmarking or reviewing components is to run them at maximum workload to see what their limits are. The same goes for performance car reviews - the magazine takes the car on a closed road and tries to push it hard to see how well it caters to the 5% of people who actually care. Realistically, you're still going to have to drive at the speed limit (well, I hope you do!) and the performance will only rear its head when you undertake the occasional maximium-workload manouver like quick overtaking or burn-outs at the traffic lights

What does this mean?

That 2600+ you might buy might not be as super-funky as the equivalent P4, but it doesn't really matter. Your life is not going to fall apart when your friend with the new P4 beats you by 1-3% in real use.

AMD still provide the best bang-for-buck processor on the market. They have the largest number of supporting motherboards and chipsets, and are custom-designed for the person who doesn't want to buy a Compaq. If you need the extra memory bandwidth, then you should realistically consider what your budget is. You'd really want 1066 RDRAM because otherwise the increase in cost isn't worthwhile.

I'm not an AMD fanboy. I don't hate intel. They seem to have been caught at a little sabotage with the whole bapco thing, but that's normal for any competitive industry (you'd be surprised, but it does happen).

Shock horror: Most people don't spend the price of a small car on computer parts. Most people don't overclock either. And 40% of people still use a GF2MX. Do these people care about theoretical bandwidth they'll never use? Hell no.

 

[rogue1979]

Intel's price drop is a great thing for both P4 and Athlon. $193 for the 2400+ is the predicted release price, but since it will not be the fastest Athlon chip I predict within a month or two at most it will go on pricewatch for $150. I am sure after a short time the overclocking potential will get better just like it did for the P4 1.6a. I also think the 1.6a will hold steady at $130 shipped. $150 for a 2400+ maxed out at a 166MHz fsb will beat the performance of the 1.6a running 2.6GHz. Performance/price advantage goes to AMD. I really don't care if the absolute fastest cpu might be the P4 2.8GHz. I don't plan on paying the premuim price when I can overclock to close to that level for a big savings.