Yonah 1.83 GHz vs. NetBurst 3 GHz

[Dadofamunky]

Wondering if anyone can give me some quick insight. I just received a fresh new Lenovo laptop with a T2400 Yonah chip at 1.83 GHz. (ATI X1300, 1GB RAM, etc.) I'm considering swapping out my desktop system for this unit; my question is whether I will take any hits in single-threaded performance going to the Yonah. My impression is that I won't. As I understand it, and based on my limited usage thus far, it really seems to rip. Also, I hate my workplace desktop machine. (My home desktop machine is another story.)

So: 1,83 GHz Yonah vs. 3.00 NetBurst. Which is faster? :beer:

 

[JAG87]

run super pi and see for yourself.

yonah will obliterate the pentium. my 1.66 core duo does super pi in 35 seconds and my 3.2 ghz prescott does it in 41.

 

[Furen]

SuperPI is hardly an accurate depiction of overall system performance. A Yonah at 1.83 will be a bit slower than a 3.0GHz Prescott and something like 10-15% slower than a 3.0GHz Northwood, but only in single-threaded applications (and overall, there are some applications that will be significantly faster on a P4 and some that will run better on Yonah). Once you hit anything that is even slightly multi-threaded then the Yonah will completely destroy the P4s, the same if you multi-task to a significant degree.

 

[gOJDO]

Yonah T2400 outperforms a 3GHz dualcore Netburst CPU, but it wipes the floor with all singlecores

 

[Dadofamunky]

Originally posted by: Furen
SuperPI is hardly an accurate depiction of overall system performance. A Yonah at 1.83 will be a bit slower than a 3.0GHz Prescott and something like 10-15% slower than a 3.0GHz Northwood, but only in single-threaded applications (and overall, there are some applications that will be significantly faster on a P4 and some that will run better on Yonah). Once you hit anything that is even slightly multi-threaded then the Yonah will completely destroy the P4s, the same if you multi-task to a significant degree.

Yeah, and I'm definitely a sick multitasker.

 

[JAG87]

Originally posted by: Furen
SuperPI is hardly an accurate depiction of overall system performance.

riiiiiight

lets see, it depicts pure cpu performance, fsb performance, and ram performance. I think thats pretty much the whole system. super pi is only biased towards large cache, and this is not the case, since both have 2MB of cache

have a

 

[SickBeast]

Originally posted by: JAG87

Originally posted by: Furen
SuperPI is hardly an accurate depiction of overall system performance.

riiiiiight

lets see, it depicts pure cpu performance, fsb performance, and ram performance. I think thats pretty much the whole system. super pi is only biased towards large cache, and this is not the case, since both have 2MB of cache

have a

He's right. It's just doing one specific type of arithmetic. You should eat your yourself and say you're sorry!



Have another one!

:beer:

 

[SickBeast]

On a laptop you want *anything* other than a netburst/P4. I think I would even rather a P3! So...Yonah all the way.

 

[Furen]

Originally posted by: SickBeast
On a laptop you want *anything* other than a netburst/P4. I think I would even rather a P3! So...Yonah all the way.

Haha, I think he has a desktop at work that is a P4 and is looking to replace it with a laptop... that's what I got from his post, anyway.

 

[dguy6789]

Originally posted by: Furen
SuperPI is hardly an accurate depiction of overall system performance. A Yonah at 1.83 will be a bit slower than a 3.0GHz Prescott and something like 10-15% slower than a 3.0GHz Northwood, but only in single-threaded applications (and overall, there are some applications that will be significantly faster on a P4 and some that will run better on Yonah). Once you hit anything that is even slightly multi-threaded then the Yonah will completely destroy the P4s, the same if you multi-task to a significant degree.

I don't see how this is very possible. An Athlon 64 2800+(1.8Ghz) is on par if not slightly faster than a P4 3Ghz. Yonah is faster clock for clock than the K8.

 

[JAG87]

yonah is faster clock for clock than both K8 and Netburst. so stop making up stories.

and super pi is a very good indicator of single threaded performance. it uses floating point unit calculations which is the most important part of a cpu nowadays. most applications rely heavily on floating point, not integer calculations. K8 was much better than netburst because its FPU beat the crap out of the P4. the only thing that biases super pi results is cache, and like i already said this is NOT the case. go learn something before you try to help someone.

 

[Dadofamunky]

This lappie runs the 2M SuperPi test at 1:23, which is right there with comparable systems. The next thing is to run it on the ol' Netburst desktop machine and see what its results are - when I have time later. I think that's gonna be the big arbiter here. Good idea all; I'd forgotten about that program.

 

[StopSign]

Originally posted by: JAG87
yonah is faster clock for clock than both K8 and Netburst. so stop making up stories.

Even Dothan was faster clock for clock than K8 and NetBurst in most applications.

 

[Furen]

Originally posted by: JAG87
yonah is faster clock for clock than both K8 and Netburst. so stop making up stories.

and super pi is a very good indicator of single threaded performance. it uses floating point unit calculations which is the most important part of a cpu nowadays. most applications rely heavily on floating point, not integer calculations. K8 was much better than netburst because its FPU beat the crap out of the P4. the only thing that biases super pi results is cache, and like i already said this is NOT the case. go learn something before you try to help someone.

Hmm, I tried to ignore you but obviously you can't just let things go...

Here you can see benchmarks comparing Yonah and the K8 in a clock-for-clock basis. You'll notice that they perform pretty much the same, sometimes one beats the other but overall they're pretty equal. Now, you might remember that the socket 754 3000+ was a pretty good competitor against the 3.0GHz Prescott, in fact, it beat it more often than not, but that was a 2.0GHz part. Once you go down to 1.8GHz (which is 10% slower), the two chips are pretty evenly matched, though I'd say the Prescott has a bit of an edge. This is a Prescott, though, Northwood has significantly better performance.

Now, notice that I'm not comparing Yonah to any of the Pentium Ds, I'm mostly trying to extrapolate available data to make a comparison between Yonah at 1.86 and a P4 at 3.0 and only in a purely single-threaded basis. Once you multithread then the little advantage these single-core CPUs have evaporates.

Also, SuperPI is a good benchmark? Please. SuperPI is not a good benchmark at all, it massively overstates the importance of the cache sizes, the cache latencies, the memory subsystem and, yes, even the FPU. FPU performance is not the most important thing to measure at all, most applications are heavily reliant in ALU and SSE performance. The fact that games rely on the FPU heavily doesn't make it " the most important part of a cpu nowadays." Also remember that the K8 still beats the crap out of Core 2 in pure x87 performance yet it not even close in real-world benchmarks, or in SuperPI, for that matter. Hell, running SYSmark is probably a better indicator of relative performance than SuperPI and I've never been a big fan of synthetics.

 

[JAG87]

Originally posted by: Furen

Also remember that the K8 still beats the crap out of Core 2 in pure x87 performance yet it not even close in real-world benchmarks, or in SuperPI, for that matter.

hmm, you say K8 beats the crap out of Core 2 in x87, yet superpi is solely based on x87 floating point calculations. so show me a single super pi run where a K8 beats Core 2.

second of all, the link you posted shows a 2.0 GHZ yonah coming close and sometimes beating a 2.0 GHZ Athlon X2 in both single and multithreaded applications. are you saying that a 3.0 GHZ prescott is better than a 2.0 GHZ K8?

good grief stop talking while its still early.

 

[Dadofamunky]

It definitely looks like the Yonah's gonna kick out my desktop machine.
To wit:

"Conclusions
Based on what we've seen, one can't help but conclude the Core Duo's performance per watt is unmatched in the world of PC processors. The Core Duo is obviously the best mobile CPU on the market, more than doubling the peak performance of the Pentium M while operating in the same power envelope. What's more shocking is the fact that the Core Duo T2600's outright performance is easily superior to Intel's supposed flagship desktop processor, the Pentium Extreme Edition 965. Given its performance, the Core Duo is clearly well-suited for desktop use, where performance is king but quiet computing is still a blessing. Not only that, but at under $700, the Core Duo T2600 costs less than the Extreme Edition 965. The lesser models are more affordable and a better value, such as the T2400 at 1.83GHz that might tempt us away from one of our favorites in that price range, the Athlon 64 X2 3800+. Even the Asus N4L-VM DH's rough edges are blunted somewhat by the fact that this board isn't your typical mobile-on-desktop prima donna; its suggested retail price is only $159.
This combination makes the Core Duo Intel's most attractive processor for PC enthusiasts, and that proposition could become downright irresistible if Asus or somebody else can deliver a mobo and BIOS with the kind of tweaking options PC enthusiasts have come to expect. The T2600 can't quite take the overall performance crown from the likes of the Athlon 64 FX-60 or the X2 4800+, but jeez, it's startlingly close. If we could get the Core Duo overclocked reasonably well, it might just be able to make a run at the title of the fastest x86-compatible CPU?or at least grab a share of that title.

As it stands, the Core Duo is an excellent choice for a quiet desktop PC or a silent gaming rig, and it's perfect for a home theater PC, where the 64-bit memory space issue isn't likely to rear its ugly head for at least several years. Were it not for the fact that Core Duo can't handle 64-bit addressing, I'd say Intel should transition its desktop and server product lines to this microarchitecture right now rather than waiting for Conroe, Merom, and Woodcrest. "

http://techreport.com/reviews/2006q2/core-duo/index.x?pg=1

And this new unit actually has a GPU with its own buffer. Strangely, the Lenovo T60 I now have appears to have 512 MB of video ram! I find that hard to believe.

 

[JAG87]

its probably shared video ram

 

[Furen]

Originally posted by: JAG87

Originally posted by: Furen

Also remember that the K8 still beats the crap out of Core 2 in pure x87 performance yet it not even close in real-world benchmarks, or in SuperPI, for that matter.

hmm, you say K8 beats the crap out of Core 2 in x87, yet superpi is solely based on x87 floating point calculations. so show me a single super pi run where a K8 beats Core 2.

second of all, the link you posted shows a 2.0 GHZ yonah coming close and sometimes beating a 2.0 GHZ Athlon X2 in both single and multithreaded applications. are you saying that a 3.0 GHZ prescott is better than a 2.0 GHZ K8?

good grief stop talking while its still early.

No, I'm saying that a 3.0GHz Prescott is better than a 1.86GHz K8 (which is about the same as a 1.86GHz Yonah), maybe, but this is ceirtainly true for a 3.0GHz Northwood. Since he didn't say exactly which CPU he had, it could be a Northwood for all I know.

FP performance is the ONLY advantage K8 has over Conroe, look it up and you'll see for yourself. Notice I said that Core 2 beats K8 eventhough the K8 has better FP performance. My point was that SuperPI relies too heavily in cache sizes, cache latencies, memory latencies, SSE performance, etc. Most of these things do yield benefits in some applications but SuperPI benefits from them greatly, which is ceirtainly not typical. Core 2 is something like 30-40% faster than K8 in SuperPI but you'd be extremely hard-pressed to find applications that exhibit this same performance growth...

 

[gOJDO]

I am not sure if dualcore Northwood at 3.0GHz can beat a Yonah at 1.86GHz in real-life apps. Northwood will show some good sandra score, but in real-life apps, Yonah dominates. I was making a comparison between the performance of one of my notebooks with Pentium M 760(Dothan 2GHz 2MB L2) and a P4(Presscott 1MB L2) at 3.2GHz. My notebook wiped the floor with the P4 for: Matlab, Simulink, VS 6 & VS.Net compiling and application performance, MS Fortran(mostly complex matrixes math algorythms), Photoshop, encoding and ofcourse gaming. I have a notebook with a Turion64 ML-34(Lancaster 1.8GHz 1MB L2) also. I've compared it to my Dothan and concluded that clock for clock the Dothan is marginaly faster for the purposes I've mentioned before.

About SuperPi, it is not a pure FPU, but it is dependend of the ALU also. The size of the cache is only little affecting its performance, and the difference is almost unnoticable on the K8. Core2 has great SuperPi results because SuperPi uses a lot of branchnings in the calculations and here the macroOP fussion comes into play. SuperPi is not a relevant factor for the performance of the CPU or the system, but I have noticed that CPUs that perform good in SuperPi, perform good in games too.

 

[A554SS1N]

Considering the close single threaded performance, one has to look at the extreme benefit of much lower power consumption by the yonah chip, as afterall, this is a laptop we're talking about, and it will be able to provide good performance for longer on the same battery.

 

[JAG87]

Furen, if he has a northwood it might be a different story. The only reason I assumed that he has a prescott is because 3.0 ghz northwood were super expensive, so people mostly bought 2.4 chips and overclocked them to 3.0

but in this situation we have 2 chips both with 2MB of cache. the prescott is running at 800 FSB while yonah runs at 533, so it already has a nice advantage. still the yonah gets better times in super pi. clearly its the chip making the difference. so IMHO 1.86 ghz yonah > 3 ghz prescott. if you count that its dual core too, than it becomes a no brainer.

 

[Furen]

Originally posted by: JAG87
Furen, if he has a northwood it might be a different story. The only reason I assumed that he has a prescott is because 3.0 ghz northwood were super expensive, so people mostly bought 2.4 chips and overclocked them to 3.0

but in this situation we have 2 chips both with 2MB of cache. the prescott is running at 800 FSB while yonah runs at 533, so it already has a nice advantage. still the yonah gets better times in super pi. clearly its the chip making the difference. so IMHO 1.86 ghz yonah > 3 ghz prescott. if you count that its dual core too, than it becomes a no brainer.

Well, since I won't convince you and you won't convince me, we'll just leave it at that. One thing I do want to mention is that the cache is significantly different between Yonah and Prescott, as are the prefetchers, the branch predictors, etc. Just because they both have 2MB of L2 doesn't mean that they both benefit from it the same way. That said, I do agree that Yonah is the better part, I'm just saying that I don't believe he'll get any single-threaded performance benefits from it and even if he does (this depends on the application mix, after all) they'll be relatively insignificant.

 

[Dadofamunky]

Nope, I don't think so. There's no way an IGP in a laptop would eat up half the system RAM. That would be a horrible design, in any case (My Dell lappie has an Intel IGP, and it uses a max of 128 MB). The ATI X1300 in the ThinkPad is a separate video chip and not an IGP, is NOTICEABLY faster than the Dell video, and I just don't think it would get that kind of performance using system RAM as its buffer. Man, paying $1500 for this thing, it'd BETTER not.