Is the only advantage of Nehalem over C2Q, Hyperthreading?

[VirtualLarry]

Just curious. Since it seems like, at the same clock speed, without HT enabled, Nehalem performs like C2Q in games, right? Does the IMC help much, as compared to the C2Q's mega-prefetchers?

 

[Sureshot324]

No the i7 is considerably faster clock for clock even if you disable hyperthreading. Most games are video card limited though so it's a moot point.

 

[dguy6789]

Hyperthreading is definitly the largest benefit to using an i7 over a Core 2 Quad, but it's not the only one. i7s are a bit faster clock per clock in some kinds of applications, a lot faster in others.

 

[Idontcare]

Just curious. Since it seems like, at the same clock speed, without HT enabled, Nehalem performs like C2Q in games, right? Does the IMC help much, as compared to the C2Q's mega-prefetchers?

Power-consumption goes down while delivering the same fps in games. So fps/watt is better for nehalem versus yorkfield.

from: http://forums.anandtech.com/showpost.php?p=26801811&postcount=145

Thanks, I appreciate the sanity check, so I'm not entirely off-base on this I guess.

Still though I like the new power numbers that Anand published. It's actually quite a nice showing for Nehalem.

Not sure why nobody actually crunches the data into performance/watt metrics anymore, guess its not sexy enough anymore. It's so 2007.

I went ahead and crunched Anand's data to convert it to performance/watt:

CPU...................................QX9770 (3.2GHz)..........Core i7-965 (3.2GHz).............Improvement
POV-Ray..............................11.4 PPS/Watt..............17.5 PPS/Watt......................53%
Cinebench (1 thread)............20.3 CBMarks/Watt.......26.6 CBMarks/Watt...............31%
Cinebench (max threads)......61.8 CBMarks/Watt.......81.5 CBMarks/Watt...............32%
3dsmax 9 SPECapc CPU........0.060 /Watt..................0.084 /Watt..........................41%
x264 HD Encode Test............0.32 fps/Watt................0.44 fps/Watt.......................38%
DivX 6.8.3............................2.61 Watts...................1.84 Watts............................29%
Windows Media Encoder........2.01 Watts....................1.34 Watts............................33%
Age of Conan.......................0.35 fps/Watt................0.46 fps/Watt........................31%
Race Driver GRID.................0.30 fps/Watt...............0.34 fps/Watt........................15%
Crysis..................................0.14 fps/Watt...............0.16 fps/Watt........................15%
FarCry 2..............................0.32 fps/Watt................0.42 fps/Watt........................34%
Fallout 3...............................0.25 fps/Watt...............0.37 fps/Watt........................45%

Unless I made a mistake in the math the i7 beat the QX9770 in every test. The average percent power consumption reduction per unit of work being done is 33% for the i7 over yorkfield.

Now I am finally seeing the 30-40% power consumption reduction numbers I was expecting once performance is normalized Me much happier now!

See Anand's writeup about Nehalem advantages here: http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=3453&p=3

 

[heyheybooboo]

Integrated Memory Controller ...

faster RAMs and more through-put

 

[Ben90]

Lol IDC, I really want to see some xxxx/Watt measurements between lynnfield and prescott now

Also curious, doesn't X58 draw less power than its x48 counterpart due to the IMC? Or is it not a big difference

 

[VirtualLarry]

Just a note - the central thesis for this thread was based around comments such as this one, which I took at face value:

Where are you getting the 20% IPC improvement from? Core 2's often outperform i7s clock per clock in games, and in general are at least comparable, except in heavily multithreaded titles where the lack of L3 cache hurts the core 2 quads.

So which is it? Do C2D/C2Q outperform i7 in games, or does the i7 outperform the C2Q?

 

[Fox5]

Just a note - the central thesis for this thread was based around comments such as this one, which I took at face value:


So which is it? Do C2D/C2Q outperform i7 in games, or does the i7 outperform the C2Q?

It varies. That comment of mine really only applies to the core 2 quads with 12MB of cache btw, and the faster core 2 quads will usually outperform the lower i7s just on the mhz advantage.

The IMC addition seems to have had the same effect it did going from athlon xp to athlon 64:
Sometimes there's no improvement, sometimes you get a 10%-20% boost per clock. However, the Athlon 64 had the same memory or faster, and the same cache or more. The core 2 quad's have more L2 (slower than i7's L2 but faster than its L3), so some apps prefer that.

I wouldn't say i7s vastly outperform the Core 2 line though. Not yet, anyway.

 

[Thor86]

Integrated memory controller and triple channel memory bandwidth = way better performance. Does this matter in games? Only time will tell when GPUs aren't the bottleneck.

 

[Munky]

What happens though when gpu is not the bottleneck, for example with something like a radeon 5970? I'm pretty sure an i7 will beat an c2q in that case.

 

[evolucion8]

http://www.legionhardware.com/document.php?id=869&p=23

The Phenom II X4 results were quite different to those recorded when testing with the Core i7 processors, though this was not necessarily a bad thing. When operating at lower clock speeds, the Phenom II X4 did not fair all that well, as we saw a sharp decline in performance. However when clocked at 3.0GHz and beyond, the Phenom II X4 really picked up the pace, and in many cases was able to outclass the Core i7.
In games such as Wolfenstein, Call of Duty: Modern Warfare 2, Tom Clancy’s H.A.W.X, BattleForge and Far Cry 2 the Phenom II X4 processors were actually faster when clocked up near 4GHz! This is quite amazing as out of the 9 games tested, the Phenom II X4 series was faster than the Core i7’s in 5 of them. Although the margins were very limited, the Phenom II X4 was found to be faster, and had it just managed to match the Core i7 series with the Radeon HD 5970, we would have been impressed.

While the Phenom II X4 matched the Core i7 in Crysis Warhead, the only games where it failed was Company of Heroes Opposing Fronts, Left 4 Dead 2 and Batman Arkham Asylum. The Phenom II X4 was noticeably slower in these games, making the Core i7 the better choice here. Still, for the most part we found the Phenom II X4 to be every bit as good as the Core i7 processors when gaming with the new Radeon HD 5970.

Having said that, we recommend that AMD users looking at buying this powerful graphics card make sure that they have a Phenom II X4 processor that is clocked at 3.0GHz or greater. Most Phenom II X4 processors are capable of overclocking to 3.0GHz and beyond, while the more high-end options, such as the Phenom II X4 955 and 965 processors, come clocked at 3.2GHz and 3.4GHz respectively.

While we hardly expect there will be many users trying to pair a $600 US graphics card, such as the Radeon HD 5970, with a budget processor, it is nice to see that the sub-$200 US processors are up to the task. The Intel Core i7 920 proved to be more than powerful enough at $280 US, while the AMD Phenom II X4 955 will work just as well at $165 US, giving users plenty of great options.

In my opinion, considering that the Phenom X4 is as fast as an equivalent Yorkfield Quad Core, I don't see the benefits of going i7 unless if you do Quad Fire with the HD 5970.